Brain Archive

Deep Fritz Draws: Are Humans Getting Smarter, or Are Computers Getting Stupider?

We find one example of qualitative improvements in software in the world of computer chess, which, according to common wisdom, is governed only by the brute-force expansion of computer hardware. In a chess tournament in October 2002 with top-ranking human player Vladimir Kramnik, the Deep Fritz software achieved a draw. I point out that Deep Fritz has available only about 1.3 percent of the brute-force computation as the previous computer champion, Deep Blue. Despite that, it plays chess at about the same level because of its superior pattern recognition–based pruning algorithm (see below). In six years a program like Deep Fritz will again achieve Deep Blue's ability to analyze two hundred million board positions per second.

Deep Fritz–like chess programs running on ordinary personal computers will routinely defeat all humans later in this decade.

In The Age of Intelligent Machines, which I wrote between 1986 and 1989, I predicted that a computer would defeat the human world chess champion by the end of the 1990s. I also noted that computers were gaining about forty-five points per year in their chess ratings, whereas the best human playing was essentially fixed, so this projected a crossover point in 1998. Indeed, Deep Blue did defeat Gary Kasparov in a highly publicized tournament in 1997.

Yet in the Deep Fritz–Kramnik match, the current reigning computer program was able to achieve only a tie. Five years had passed since Deep Blue's victory, so what are we to make of this situation? Should we conclude that:

1. Humans are getting smarter, or at least better at chess?

2. Computers are getting worse at chess? If so, should we conclude that the much- publicized improvement in computer speed over the past five years was not all it was cracked up to be? Or, that computer software is getting worse, at least in chess?

The specialized-hardware advantage. Neither of the above conclusions is warranted. The correct conclusion is that software is getting better because Deep Fritz essentially matched the performance of Deep Blue, yet with far smaller computational resources. To gain some insight into these questions, we need to examine a few essentials. When I wrote my predictions of computer chess in the late 1980s, Carnegie Mellon University was embarked on a program to develop specialized chips for conducting the "minimax" algorithm (the standard game-playing method that relies on building trees of move-countermove sequences, and then evaluating the terminal-leaf positions of each branch of the tree) specifically for chess moves.

Based on this specialized hardware CMU's 1988 chess machine, HiTech, was able to analyze 175,000 board positions per second. It achieved a chess rating of 2,359, only about 440 points below the human world champion.

A year later, in 1989, CMU's Deep Thought machine increased this capacity to one million board positions per second and achieved a rating of 2,400. IBM eventually took over the project and renamed it Deep Blue but kept the basic CMU architecture. The version of Deep Blue that defeated Kasparov in 1997 had 256 special-purpose chess processors working in parallel, which analyzed two hundred million board positions per second.

It is important to note the use of specialized hardware to accelerate the specific calculations needed to generate the minimax algorithm for chess moves. It's well-known to computer-systems designers that specialized hardware generally can implement a specific algorithm at least one hundred times faster than a general-purpose computer. Specialized ASICs (Application-specific Integrated Circuits) require significant development efforts and costs, but for critical calculations that are needed on a repetitive basis (for example, decoding MP3 files or rendering graphics primitives for video games), this expenditure can be well worth the investment.

Deep Blue versus Deep Fritz. Because there had always been a great deal of focus on the milestone of a computer's being able to defeat a human opponent, support was available for investing in special-purpose chess circuits. Although there was some lingering controversy regarding the parameters of the Deep Blue–Kasparov match, the level of interest in computer chess waned considerably after 1997. After all, the goal had been achieved, and there was little point in beating a dead horse. IBM canceled work on the project, and there has been no work on specialized chess chips since that time. The focus of research in the various domains spun out of AI has been placed instead on problems of greater consequence, such as guiding airplanes, missiles, and factory robots, understanding natural language, diagnosing electrocardiograms and blood-cell images, detecting credit-card fraud, and a myriad of other successful narrow AI applications.

Computer hardware has nonetheless continued its exponential increase, with personal-computer speeds doubling every year since 1997. Thus the general-purpose Pentium processors used by Deep Fritz are about thirty-two times faster than processors in 1997. Deep Fritz uses a network of only eight personal computers, so the hardware is equivalent to 256 1997-class personal computers. Compare that to Deep Blue, which used 256 specialized chess processors, each of which was about one hundred times faster than 1997 personal computers (of course, only for computing chess minimax). So Deep Blue was 25,600 times faster than a 1997 PC and one hundred times faster than Deep Fritz. This analysis is confirmed by the reported speeds of the two systems: Deep Blue can analyze 200 million board positions per second compared to only about 2.5 million for Deep Fritz.

Significant software gains. So what can we say about the software in Deep Fritz? Although chess machines are usually referred to as examples of brute-force calculation, there is one important aspect of these systems that does require qualitative judgment. The combinatorial explosion of possible move-countermove sequences is rather formidable.

In The Age of Intelligent Machines I estimated that it would take about forty billion years to make one move if we failed to prune the move-countermove tree and attempted to make a "perfect" move in a typical game. (Assuming about thirty moves each in a typical game and about eight possible moves per play, we have 8³⁰ possible move sequences; analyzing one billion move sequences per second would take 10¹⁸ seconds or forty billion years.) Thus a practical system needs to continually prune away unpromising lines of action. This requires insight and is essentially a pattern-recognition judgment.

Humans, even world-class chess masters, perform the minimax algorithm extremely slowly, generally performing less than one move-countermove analysis per second. So how is it that a chess master can compete at all with computer systems? The answer is that we possess formidable powers of pattern recognition, which enable us to prune the tree with great insight.

It's precisely in this area that Deep Fritz has improved considerably over Deep Blue. Deep Fritz has only slightly more computation available than CMU's Deep Thought yet is rated almost 400 points higher.

Are human chess players doomed? Another prediction I made in The Age of Intelligent Machines was that once computers did perform as well or better as humans in chess, we would either think more of computer intelligence, less of human intelligence, or less of chess, and that if history is a guide, the last of these would be the likely outcome. Indeed, that is precisely what happened. Soon after Deep Blue's victory we began to hear a lot about how chess is really just a simple game of calculating combinations and that the computer victory just demonstrated that it was a better calculator.

The reality is slightly more complex. The ability of humans to perform well in chess is clearly not due to our calculating prowess, at which we are in fact rather poor. We use instead a quintessentially human form of judgment. For this type of qualitative judgment, Deep Fritz represents genuine progress over earlier systems. (Incidentally, humans have made no progress in the last five years, with the top human scores remaining just below 2,800. Kasparov is rated at 2,795 and Kramnik at 2,794.)

Where we go from here? Now that computer chess is relying on software running on ordinary personal computers, chess programs will continue to benefit from the ongoing acceleration of computer power. By 2009 a program like Deep Fritz will again achieve Deep Blue's ability to analyze two hundred million board positions per second. With the opportunity to harvest computation on the Internet, we will be able to achieve this potential several years sooner than 2009. (Internet harvesting of computers will require more ubiquitous broadband communication, but that's coming, too).

With these inevitable speed increases, as well as ongoing improvements in pattern recognition, computer chess ratings will continue to edge higher. Deep Fritz–like programs running on ordinary personal computers will routinely defeat all humans later in this decade. Then we'll really lose interest in chess.

Creating a neural-net solution to a problem involves the following steps:

• Define the input.
• Define the topology of the neural net (i.e., the layers of neurons and the connections between the neurons).
• Train the neural net on examples of the problem.
• Run the trained neural net to solve new examples of the problem.
• Take your neural-net company public. These steps (except for the last one) are detailed below:

The Problem Input

The problem input to the neural net consists of a series of numbers. This input can be:

• In a visual pattern-recognition system, a two-dimensional array of numbers representing the pixels of an image; or
• In an auditory (e.g., speech) recognition system, a two-dimensional array of numbers representing a sound, in which the first dimension represents parameters of the sound (e.g., frequency components) and the second dimension represents different points in time; or
• In an arbitrary pattern-recognition system, an n-dimensional array of numbers representing the input pattern.

Defining the Topology

To set up the neural net, the architecture of each neuron consists of:

• Multiple inputs in which each input is "connected" to either the output of another neuron, or one of the input numbers.
• Generally, a single output, which is connected either to the input of another neuron (which is usually in a higher layer), or to the final output.

Set up the First Layer of Neurons

• Create N₀ neurons in the first layer. For each of these neurons, "connect" each of the multiple inputs of the neuron to "points" (i.e., numbers) in the problem input. These connections can be determined randomly or using an evolutionary algorithm (see below).
• Assign an initial "synaptic strength" to each connection created. These weights can start out all the same, can be assigned randomly, or can be determined in another way (see below).

Set up the Additional Layers of Neurons

Set up a total of M layers of neurons. For each layer, set up the neurons in that layer. For layer_i:

• Create N_i neurons in layer_i. For each of these neurons, "connect" each of the multiple inputs of the neuron to the outputs of the neurons in layer_i–1 (see variations below).
• Assign an initial "synaptic strength" to each connection created. These weights can start out all the same, can be assigned randomly, or can be determined in another way (see below).
• The outputs of the neurons in layerM are the outputs of the neural net (see variations below).

The Recognition Trials

How Each Neuron Works Once the neuron is set up, it does the following for each recognition trial.

• Each weighted input to the neuron is computed by multiplying the output of the other neuron (or initial input) that the input to this neuron is connected to by the synaptic strength of that connection.
• All of these weighted inputs to the neuron are summed.
• If this sum is greater than the firing threshold of this neuron, then this is neuron is considered to fire and its output is 1. Otherwise, its output is 0 (see variations below).

Do the Following for Each Recognition Trial

For each layer, from layer₀ to layer_M: For each neuron in the layer:

• Sum its weighted inputs (each weighted input = the output of the other neuron [or initial input] that the input to this neuron is connected to multiplied by the synaptic strength of that connection).
• If this sum of weighted inputs is greater than the firing threshold for this neuron, set the output of this neuron = 1, otherwise set it to 0.

To Train the Neural Net

• Run repeated recognition trials on sample problems.
• After each trial, adjust the synaptic strengths of all the interneuronal connections to improve the performance of the neural net on this trial (see the discussion below on how to do this).
• Continue this training until the accuracy rate of the neural net is no longer improving (i.e., reaches an asymptote).

Key Design Decisions

In the simple schema above, the designer of this neural-net algorithm needs to determine at the outset:

• What the input numbers represent.
• The number of layers of neurons.
• The number of neurons in each layer. (Each layer does not necessarily need to have the same number of neurons.)
• The number of inputs to each neuron in each layer. The number of inputs (i.e., interneuronal connections) can also vary from neuron to neuron and from layer to layer.
• The actual "wiring" (i.e., the connections). For each neuron in each layer, this consists of a list of other neurons, the outputs of which constitute the inputs to this neuron. This represents a key design area. There are a number of possible ways to do this:
  • (i) Wire the neural net randomly; or
  • (ii) Use an evolutionary algorithm (see below) to determine an optimal wiring; or
  • (iii) Use the system designer's best judgment in determining the wiring.

• The initial synaptic strengths (i.e., weights) of each connection. There are a number of possible ways to do this:
  • (i) Set the synaptic strengths to the same value; or
  • (ii) Set the synaptic strengths to different random values; or
  • (iii) Use an evolutionary algorithm to determine an optimal set of initial values; or
  • (iv) Use the system designer's best judgment in determining the initial values.

• The firing threshold of each neuron.
• Determine the output. The output can be:
  • (i) the outputs of layerM of neurons; or
  • (ii) the output of a single output neuron, the inputs of which are the outputs of the neurons in layerM.
  • (iii) a function of (e.g., a sum of) the outputs of the neurons in layerM; or
  • (iv) another function of neuron outputs in multiple layers.

• Determine how the synaptic strengths of all the connections are adjusted during the training of this neural net. This is a key design decision and is the subject of a great deal of research and discussion. There are a number of possible ways to do this:
  • (i) For each recognition trial, increment or decrement each synaptic strength by a (generally small) fixed amount so that the neural net's output more closely matches the correct answer. One way to do this is to try both incrementing and decrementing and see which has the more desirable effect. This can be time-consuming, so other methods exist for making local decisions on whether to increment or decrement each synaptic strength.
  • (ii) Other statistical methods exist for modifying the synaptic strengths after each recognition trial so that the performance of the neural net on that trial more closely matches the correct answer.

Note that neural-net training will work even if the answers to the training trials are not all correct. This allows using real-world training data that may have an inherent error rate. One key to the success of a neural net–based recognition system is the amount of data used for training. Usually a very substantial amount is needed to obtain satisfactory results. Just like human students, the amount of time that a neural net spends learning its lessons is a key factor in its performance.

Variations

Many variations of the above are feasible:

• There are different ways of determining the topology. In particular, the interneuronal wiring can be set either randomly or using an evolutionary algorithm.
• There are different ways of setting the initial synaptic strengths.
• The inputs to the neurons in layer_i do not necessarily need to come from the outputs of the neurons in layer_i–1. Alternatively, the inputs to the neurons in each layer can come from any lower layer or any layer.
• There are different ways to determine the final output.
• The method described above results in an "all or nothing" (1 or 0) firing called a nonlinearity. There are other nonlinear functions that can be used. Commonly a function is used that goes from 0 to 1 in a rapid but more gradual fashion. Also, the outputs can be numbers other than 0 and 1.
• The different methods for adjusting the synaptic strengths during training represent key design decisions. The above schema describes a "synchronous" neural net, in which each recognition trial proceeds by computing the outputs of each layer, starting with layer0 through layer_M. In a true parallel system, in which each neuron is operating independently of the others, the neurons can operate "asynchronously" (i.e., independently). In an asynchronous approach, each neuron is constantly scanning its inputs and fires whenever the sum of its weighted inputs exceeds its threshold (or whatever its output function specifies).

The Evolutionary Algorithm

Create N solution "creatures". Each one has:

• A genetic code: a sequence of numbers that characterize a possible solution to the problem. The numbers can represent critical parameters, steps to a solution, rules, etc.
• For each generation of evolution, do the following:
• Do the following for each of the N solution creatures:
• Apply this solution creature's solution (as represented by its genetic code) to the problem, or simulated environment. Rate the solution.
• Pick the L solution creatures with the highest ratings to survive into the next generation.
• Eliminate the (N–L) nonsurviving solution creatures.
• Create (N–L) new solution creatures from the L surviving solution creatures by:

(i) Making copies of the L surviving creatures. Introduce small random variations into each copy; or

(ii) Create additional solution creatures by combining parts of the genetic code (using "sexual" reproduction, or otherwise combining portions of the chromosomes) from the L surviving creatures; or

(iii) Doing a combination of (i) and (ii).

• Determine whether or not to continue evolving: Improvement = (highest rating in this generation)–(highest rating in the previous generation). If Improvement < Improvement Threshold then we're done.
• The solution creature with the highest rating from the last generation of evolution has the best solution. Apply the solution defined by its genetic code to the problem.
  

Key Design Decisions

In the simple schema above, the designer needs to determine at the outset:

• Key parameters:
  • N
  • L
  • Improvement threshold
• What the numbers in the genetic code represent and how the solution is computed from the genetic code.
• A method for determining the N solution creatures in the first generation. In general, these need only be "reasonable" attempts at a solution. If these first-generation solutions are too far afield, the evolutionary algorithm may have difficulty converging on a good solution. It is often worthwhile to create the initial solution creatures in such a way that they are reasonably diverse. This will help prevent the evolutionary process from just finding a "locally" optimal solution.
• How the solutions are rated.
• How the surviving solution creatures reproduce.

Variations

Many variations of the above are feasible. For example:

• There does not need to be a fixed number of surviving solutions creatures (L) from each generation. The survival rule(s) can allow for a variable number of survivors.
• There does not need to be a fixed number of new solution creatures created in each generation (N–L). The procreation rules can be independent of the size of the population. Procreation can be related to survival, thereby allowing the fittest solution creatures to procreate the most.
• The decision as to whether or not to continue evolving can be varied. It can consider more than just the highest-rated solution creature from the most recent generation(s). It can also consider a trend that goes beyond just the last two generations

The Recursive Algorithm

Define a function (program) "Pick Best Next Step." The function returns a value of "SUCCESS" (we've solved the problem) or "FAILURE" (we didn't solve it). If it returns with a value of SUCCESS, then the function also returns the sequence of steps that solved the problem. Pick Best Next Step does the following:

• Determine if the program can escape from continued recursion at this point. This bullet, and the next two bullets deal with this escape decision. First, determine if the problem has now been solved. Since this call to Pick Best Next Step probably came from the program calling itself, we may now have a satisfactory solution. Examples are:
  • (i) In the context of a game (for example, chess), the last move allows us to win (such as checkmate).
  • (ii) In the context of solving a mathematical theorem, the last step proves the theorem.
  • (iii) In the context of an artistic program (for example, a computer poet or composer), the last step matches the goals for the next word or note.

If the problem has been satisfactorily solved, the program returns with a value of "SUCCESS" and the sequence of steps that caused the success.

If the problem has not been solved, determine if a solution is now hopeless. Examples are:

• (i) In the context of a game (such as chess), this move causes us to lose (checkmate for the other side).
• (ii) In the context of solving a mathematical theorem, this step violates the theorem.
• (iii) In the context of an artistic creation, this step violates the goals for the next word or note.
• If the solution at this point has been deemed hopeless, the program returns with a value of "FAILURE."
• If the problem has been neither solved nor deemed hopeless at this point of recursive expansion, determine whether or not the expansion should be abandoned anyway. This is a key aspect of the design and takes into consideration the limited amount of computer time we have to spend. Examples are:
  • (i) In the context of a game (such as chess), this move puts our side sufficiently "ahead" or "behind." Making this determination may not be straightforward and is the primary design decision. However, simple approaches (such as adding up piece values) can still provide good results. If the program determines that our side is sufficiently ahead, then Pick Best Next Step returns in a similar manner to a determination that our side has won (that is, with a value of "SUCCESS"). If the program determines that our side is sufficiently behind, then Pick Best Next Step returns in a similar manner to a determination that our side has lost (that is, with a value of "FAILURE").
  • (ii) In the context of solving a mathematical theorem, this step involves determining if the sequence of steps in the proof is unlikely to yield a proof. If so, then this path should be abandoned, and Pick Best Next Step returns in a similar manner to a determination that this step violates the theorem (that is, with a value of "FAILURE"). There is no "soft" equivalent of success. We can't return with a value of "SUCCESS" until we have actually solved the problem. That's the nature of math.
  • (iii) In the context of an artistic program (such as a computer poet or composer), this step involves determining if the sequence of steps (such as the words in a poem, notes in a song) is unlikely to satisfy the goals for the next step. If so, then this path should be abandoned, and Pick Best Next Step returns in a similar manner to a determination that this step violates the goals for the next step (that is, with a value of "FAILURE").

• If Pick Best Next Step has not returned (because the program has neither determined success nor failure nor made a determination that this path should be abandoned at this point), then we have not escaped from continued recursive expansion. In this case, we now generate a list of all possible next steps at this point. This is where the precise statement of the problem comes in:
  • (i) In the context of a game (such as chess), this involves generating all possible moves for "our" side for the current state of the board. This involves a straightforward codification of the rules of the game.
  • (ii) In the context of finding a proof for a mathematical theorem, this involves listing the possible axioms or previously proved theorems that can be applied at this point in the solution.
  • (iii) In the context of a cybernetic art program, this involves listing the possible words/notes/line segments that could be used at this point. For each such possible next step:
    • (i) Create the hypothetical situation that would exist if this step were implemented. In a game, this means the hypothetical state of the board. In a mathematical proof, this means adding this step (for example, axiom) to the proof. In an art program, this means, adding this word/note/line segment.
    • (ii) Now call Pick Best Next Step to examine this hypothetical situation. This is, of course, where the recursion comes in because the program is now calling itself.
    • (iii) If the above call to Pick Best Next Step returns with a value of "SUCCESS," then return from the call to Pick Best Next Step (that we are now in) also with a value of "SUCCESS." Otherwise consider the next possible step.

    If all the possible next steps have been considered without finding a step that resulted in a return from the call to Pick Best Next Step with a value of "SUCCESS," then return from this call to Pick Best Next Step (that we are now in) with a value of "FAILURE."

    END OF PICK BEST NEXT STEP

    If the original call to Pick Best Next Move returns with a value of "SUCCESS," it will also return the correct sequence of steps:

• (i) In the context of a game, the first step in this sequence is the next move you should make.
• (ii) In the context of a mathematical proof, the full sequence of steps is the proof.
• (iii) In the context of a cybernetic art program, the sequence of steps is your work of art.

If the original call to Pick Best Next Step is "FAILURE," then you need to go back to the drawing board.


Key Design Decisions

In the simple schema above, the designer of the recursive algorithm needs to determine the following at the outset:

• The key to a recursive algorithm is the determination in Pick Best Next Step of when to abandon the recursive expansion. This is easy when the program has achieved clear success (such as checkmate in chess or the requisite solution in a math or combinatorial problem) or clear failure. It is more difficult when a clear win or loss has not yet been achieved. Abandoning a line of inquiry before a well-defined outcome is necessary because otherwise the program might run for billions of years (or at least until the warranty on your computer runs out).
• The other primary requirement for the recursive algorithm is a straightforward codification of the problem. In a game like chess, that's easy. But in other situations, a clear definition of the problem is not always so easy to come by

It Wasn't Watson!
posted on 07/16/2006 12:20 AM by Trurl73

Brin's error-correcting paradigm (and my fondness of Doyle) moves me to this elementary observation:

Ray's reference to the Canon, whilst nice to see, has an undoubtedly unwitting misattribution.

Jabez Wilson, not Holmes's trusted biographer and colleague, made the comment to which he refers.

Re: Why We Can Be Confident of Turing Test Capability Within a Quarter Century
posted on 07/16/2006 6:20 PM by johnclavin

If we were to build a very powerful AI machine using the tools listed in the AI Toolkit; expert systems, Bayesian nets, Markov models, neural nets, genetic algorithms, and recursive searching, we might have a machine that could fool a turning test, but it would not have any self awareness, creative intuition, or the ability to think outside the box. (pun intended!)
I think it is time to take a closer look at the work of Harold Cohen and other generative artists who might have ideas about finding creativity in a machine. Artistic creativity does not come from a paradigm that narrows down knowledge with some end result in mind, it comes from a random thought that is captured and developed. The AI Toolkit can be applied to developing creative ideas but the spark has to come from unpredictable events.
I believe that research in this area could lead to a machine intelligence that will someday be self aware and have a free will. Then a machine will pass the Turing test with honor instead of being a programmed imitation of intelligence.

John Clavin

Re: Why We Can Be Confident of Turing Test Capability Within a Quarter Century
posted on 03/04/2007 11:39 PM by joanneB

It may be possible to create languages to communicate with other states of mind... and record the "conversations"...
http://vibepc.homestead.com/Emma_Lesson_1c.ppt

Re: Why We Can Be Confident of Turing Test Capability Within a Quarter Century
posted on 07/30/2006 6:41 AM by Abrob1006

It would not take 25 years to pass the the Turing Test. I can guarantee you that we will do it within next 10 years as I am actively involved in such projects. And we'll launch it commercially years before that after the beta-testing is done.

As far as invoking Self-awareness or True Consciousness in machines/Web is concerned, no one can give any time limit for that although it's inevitable. Because it's not only a matter of intense human efforts but it's pre-determined by Will and Intellgence of the Universe and embedded in its Cosmic Evolution.

But it's highly probable that we may able build a self-aware Web/Network/Machines within 10 to 25 years.It won't be algorithmic in the sense we are familiar with the term in computer science and technology but it will grow and manifest its self-awareness and other higher-order capabilities in months and years after its birth/launch.

Re: Why We Can Be Confident of Turing Test Capability Within a Quarter Century
posted on 10/08/2006 10:33 AM by jack d

"As far as invoking Self-awareness or True Consciousness in machines/Web is concerned, no one can give any time limit for that although it's inevitable. Because it's not only a matter of intense human efforts but it's pre-determined by Will and Intellgence of the Universe and embedded in its Cosmic Evolution. "

wow ! is this a theory or just picturesque philosphy ?

Re: Why We Can Be Confident of Turing Test Capability Within a Quarter Century
posted on 10/13/2006 2:33 AM by jack d

I reckon its already been done and 'self-aware' machines are about as close as life on Pluto.

Spiritual test of intelligence
posted on 10/08/2006 8:42 AM by Timothy

I would be interested in Ray's view of the spiritual capacities of the spiritual machines (strong AI) he is talking about.

For example, deep states of meditation are states in which the human mind is tacidly aware of its own presence, or even tacidly aware of its direct environment (similar to athletes experiencing "flow"), while all thought-processes have stopped.

The mind is still, in abeyance. But consciousness is not in abeyance, on the contrary, there's a strong sense of being, of precence.

I do not know how this state could be objectively tested, but maybe someone will be able to devise a Spiritual Test, not just a mental test like the Turing and Feigenbaum tests. These are tests that assess the development of mind, our level of thinking.

A machine can respond to and even interact intelligently with its environment, but will it be able to answer not just intellectually, the questions Who am I or What is consciousness?

It seems to me that a truly intelligent and spiritual machine will inately have a longing to explore these questions and have them answered.

Will machines, once they awaken to true intelligence (strong AI), have a natural desire to answer these spiritual questions?

Will a new generation of (machine-)philosophers or rather, (machine-)Buddhists be born?

My Spiritual Test for strong AI would be simply to see if these - for humans quite natural - spiritual questions 'spontanuously' arise in our intelligent machines.

Are they even interested in spirituality, are they even interested in ethics, humanist thought, or practicing compassion?

It seems to me that a truly intelligent machine should demonstrate a natural concern for these things. Otherwise "intelligence" would mean cleverness, rather than intelligence.

Timothy Schoorel

Re: Spiritual test of intelligence
posted on 11/20/2006 5:10 PM by BigMTBrain

"It seems to me that a truly intelligent machine should demonstrate a natural concern for these things. Otherwise "intelligence" would mean cleverness, rather than intelligence. "

Don't forget that in your mind as well as the minds of others there's a lifetime's worth of meta data wrought by experience. The meta data is what forms and guides an entity's ethics, etc.

The answer to what concerns a superAI would have depend completely on what kind of meta data you feed or make available to its engine.

Re: Spiritual test of intelligence
posted on 11/20/2006 10:13 PM by BigMTBrain

Ooops! Meant to say "data and metadata (data describing data)"

Re: Why We Can Be Confident of Turing Test Capability Within a Quarter Century
posted on 10/08/2006 10:38 AM by jack d

WHat we really need is somebody to explain WHY we need a machine to pass the Turing test. For the life of me, why we need to pass a test which shows that a machine can converse like a human (and let's face it, it probably can be passed in many basic circumstances today) is beyond me as it demonstrates nothing more than it demonstrates.

Namely, that it is possible to teach computers to speak. NOT to understand or comprehend that they're speaking, but nonethless to be able to do conduct conversations from the obeerver's perspective. And er .... that's it.

Anybody else think of any significant reason why the Turing test is important ?

Re: Why We Can Be Confident of Turing Test Capability Within a Quarter Century
posted on 10/08/2006 10:59 AM by B-Punk

I wonder how many humans would not pass a Turing test....

Re: Why We Can Be Confident of Turing Test Capability Within a Quarter Century
posted on 10/12/2006 5:35 PM by jack d

exactly ! Your starting point for a scientific test cannot be a metric that is ultimately subjective, i.e the test is passed when a panel kind of thinks that, well its almost certain that, well, we have a human here .

Why we need a Turing Test
posted on 10/08/2006 11:57 AM by Timothy

We need something like a Turing Test because we want to transfer human intelligence and consciousness to a platform that is more stable and expandable than our biology.

The question is what human intelligence and consciousness are, in essence.

This is what we are trying to capture in for example the Turing Test.

When we've created Strong AI, a machine that is self-aware and intelligent like humans, we will have trancended many limitations that come with being biologically human.

In principle this is something we want, we want to trancend our limitations, and copying intelligence to a more solid (silicon) platform will in theory increase the chances of survival of human civilisation.

Either way, it seems that the advance of technology is an evolutionary imperative: we can hardly avoid it, even if we wanted to avoid it.

In short, passing the Turing Test would mean that a computer does not just utter some words but is capable of intelligent conversation. Therefor we would presume that this machine has actually achieved true intelligence and is not just faking it.

Re: Why we need a Turing Test
posted on 10/08/2006 5:37 PM by stereoheadfm

I agree we need a Turing Test because it offers a benchmark in respect to how convincing an artificial intelligence seems. Also, you're quite correct that the essential questions regard what intelligence and conciousness are.

However, I fear in the relevance of the issues we discuss we've neglected an overarching principle of conciousness, specifically that nobody has the authority to identify conciousness. I'm no skeptic, but it's worth noting that as far as I'm concerned, it's logically possible everyone in the world is a "machine" not deserving the title of conciousness.

In my personal opinion, there's a more valid "Sistine Chapel Creation of Adam Test" whereby the first concious machine is the first one to generate the notion that we are fundamentally isolated-- that it has no more reason to be convinced of our conciousness than we of its.

I admit that I agree with jack d, it's entirely possible to have a convincing conversation with an artifical intelligence with no capability for self-awareness, and that this would be less valid. Let me clarify that although I feel nobody has the ultimate authority to demonstrate conciousness, there are definitions by which we identify what we see.

Re: Why we need a Turing Test
posted on 10/12/2006 6:35 PM by jack d

I think you've hit the point but in a sense you've fallen into the old trap. You CAN measure consciousness - anaesthetists do it all the time. All you need is a theory of consciousness and you can set up a mechanism of measurement.


The same applies throughout science - for instance, atmoic theory. We can measure the width of an atom, but only by using equipment and measurements based upon theories that assume atomic theory to be correct.

We just don't have a ver advanced theory of consciousness yet - but we do have them.

Re: Why we need a Turing Test
posted on 10/12/2006 6:39 PM by Fred84

> You CAN measure consciousness

i'm not really contributing, but i'd say:

based on subconscious calculations, i think you are right

Re: Why we need a Turing Test
posted on 10/12/2006 5:37 PM by jack d

"In short, passing the Turing Test would mean that a computer does not just utter some words but is capable of intelligent conversation. Therefor we would presume that this machine has actually achieved true intelligence and is not just faking it. "

Is that true ? I though the whole point about the Turing Test is that it isnt concerned AT ALL if the machine is thinking. It is literally only concerned with external appearances - namely that if it sounds like its thinking then it must be thinking. I think you need to read the test and you'll see that this is the case.

Re: Why we need a Turing Test
posted on 10/12/2006 7:44 PM by Timothy

"The basis of the Turing test is that if the human Turing test judge is competent, then an entity requires human-level intelligence in order to pass the test."
(http://www.kurzweilai.net/meme/frame.html?main=/a rticles/art0374.html?)

But of course, we will probably come up with all kinds of other tests for Strong AI precisely because we all will wonder: Has this machine really achieved intelligence? Is it really conscious? Does it really feel? Does it have conscious, subjective experiences, like we have?

Eventually I think most of us will accept that it has really achieved true intelligence and awareness. Just like we've generally accepted that 'other people' are intelligent and conscious beings.

Re: Why we need a Turing Test
posted on 10/12/2006 9:10 PM by i

By the time a machine passes the Turing Test, we'll have come to expect it and it won't seem like such a big deal.

The AI Effect: "The great practical benefits of AI applications and even the existence of AI in many software products go largely unnoticed by many despite the already widespread use of AI techniques in software. This is the AI effect. Many marketing people don't use the term 'artificial intelligence' even when their company's products rely on some AI techniques.

"Every time we figure out a piece of it, it stops being magical; we say, Oh, that's just a computation."

- Rodney Brooks

"As soon as someone gets a computer to do it, people say: 'That's not what we meant by intelligence.' People subconsciously are trying to preserve for themselves some special role in the universe."

- Michael Kearns

Re: Why we need a Turing Test
posted on 10/12/2006 10:48 PM by eldras

Nope it's that weak A.I. and strong A.I. are different challenges.

Almost no-one's EVER tried to do AGI (Artificial General Intelligence).

the number of serious and funded attempts is probably under 100, and of those less than 20 are worthy of real study.

The main problem ios the multi-disciplines needed..not just acdemic disciplines...but general ones including team building, project management, finance funding marketing, lobbying etc

Re: Why we need a Turing Test
posted on 10/12/2006 11:09 PM by eldras

Re: Why we need a Turing Test
posted on 10/13/2006 2:25 AM by jack d

""The basis of the Turing test is that if the human Turing test judge is competent, then an entity requires human-level intelligence in order to pass the test."

Well clearly that is not the case. Speech (it is only a speech test) is the most automatable human characteristic there is. The fact is if a human took the test, the EXISTENCE of their intelligence (and mental states) would depend upon a third party's subjective assumptions about how bright they were. For instance, some mentally handicapped people would fail the Turing test, but they still have mental states, concsciousness and intelligence.

Re: Why we need a Turing Test
posted on 10/13/2006 2:30 AM by jack d

"Eventually I think most of us will accept that it has really achieved true intelligence and awareness. Just like we've generally accepted that 'other people' are intelligent and conscious beings. "

But our belief that other people are conscious is based upon our personal KNOWLEDGE that we are. Cogito ergo sum. Other humans beings are made of the same stuff and act in the same way.

The same can't be said of machines. I know what they're made of and exactly what they do. I have every reason therefore to believe that they are not conscious because I know they are nothing at all like me.

Re: Why we need a Turing Test
posted on 10/13/2006 11:39 AM by Timothy

Why do you presume that machines won't have mental states?

If you are a materialist and think that the universe is made up out of matter, than there is really not much difference between a human being and a machine! Everything, including the human brain, is made up out of atoms or some even smaller building blocks of matter. So from a materialist's point of view, human beings and machines are indeed 'made of the same stuff'.

If you are a 'spiritualist' and think that everything is an expression or modification of consciousness itself, than the same logic says that there is really no fundamental difference between human beings and machines and that therefor there's no reason why machines couldn't become truly conscious and intelligent, like human beings.

The 3rd option is believing, like Descartes did, that there is a fundamental divide between matter and mind. But then, what is the mechanism by which the mind influences the body? This question has, as far as I know, never been answered in a philosophically satisfactory way. Also, there are so many indications now that mind originates from the brain.

For me, 'Cogito ergo sum' (I think, therefor I am) is not all that meaningful. There's no denying that machines can think. The question is whether machines can think consciously, can be conscious, that's the real question!

More meaningful than 'Cogito ergo sum' I think would be 'Cognito ergo sum': I am conscious, therefor I am.

The whole point of the effort towards Strong AI is that we somehow manage to transfer our intelligence and consciousness to machines, to a different platform from biology. A platform that is not as vulnerable and doesn't have all these biological limitations.

The question that needs to be answered is how we would know if we have succeeded! The Turing Test is one ingenious way to test an AI.

But maybe the AI itself will come up with an even more convincing test!

Re: Why we need a Turing Test
posted on 10/13/2006 12:48 PM by jack d

"So from a materialist's point of view, human beings and machines are indeed 'made of the same stuff'. "

No they aren't. A brain is made up of neurons of vastly differing chemical composition to a digital computer. We know that brains cause mental states : it does not follow that computers do.

The defining feature of a computer is its manipulation of symbols from the point of view of an external observer. That is why the chemical composition of a computer is irrelevant.

The defining feature of a brain is that its chemical composition is responsible for a phenomenological, not logical, process, known as thinking.

The brain can be thought of as a machine, but a machine who's defining characteristic is the manipulation of symbols from the point of view of an external oberver cannot be thought of as a brain.

In other words, with some irony its the Strong AI enthusiasts that believe in a kind of dualism. They believe that mere symbols can create phenomenological events. Its a bit like saying that a computer simulation of rainfall is the same thing as rainfall itself. Clearly this is just not true.

The Turing test is the apotheosis of this belief. In reality it has probably been passed in many ways today. But as you have realised, it sheds no light onto the consciosness problem.

Thinking is not logical or symbolic. It is physical and phenomenological. There is nothing, absolutely nothing, to allow us to believe that we can 'transfer' our thoughts to another medium down some kind of telephone line.

Re: Why we need a Turing Test
posted on 10/14/2006 9:21 AM by Timothy

"Its a bit like saying that a computer simulation of rainfall is the same thing as rainfall itself."

The issue here I think is the experience of the rainfall, rather than rainfall itself.

When we 'experience' rainfall, we don't experience rainfall but a representation, a virtual reality created by our brains.

I think computers will be very capable at creating similar 'inner' representations of reality based on sensory inputs. The question that still remains is whether they will have the actual inner experiences of that reality.

'Thinking is not logical or symbolic. It is physical and phenomenological.'

I agree, but silicon is just as physical and phenomenological as brain cells are.

Electricity running through complex arrangements of silicon is no less physical and phenomenological than an alive human brain.

To my mind it is the arrangement of the brain cells and the level of complexity of all of the processes that happen in a brain, that create human consciousness and intelligence, rather than the physical composition of a bunch of brain cells.

You seem to imply that it in the end it is the physical composition of brain cells that allows for consciousness, but for me that's not a convincing argument against the possibility of consciousness in a machine.

Re: Why we need a Turing Test
posted on 10/18/2006 4:22 PM by jack d

"The issue here I think is the experience of the rainfall, rather than rainfall itself. "

No it isn't. Although if you really want to confuse it, the 'experience of rainfall' in somebody's brain is is a phenomena as natural and physical as any rainfall.

"I agree, but silicon is just as physical and phenomenological as brain cells are."

But silicon does not have causal powers to create mental states.

And nor is there a necessary relation between silicon and computing : a computer can be made of old tins and bits of string. We can create a computer from watching sunspots if we like. Is the sun thinking ?

"Electricity running through complex arrangements of silicon is no less physical and phenomenological than an alive human brain. "

But electricity and silicon are utterly irrelevant to a computer. A computer is a logical construct. Without an observer to understand how the physical tokens (i.e voltage levels on a bit of silicon) relate to symbols such as o's and 1's, there is no computer.
A computer is a LOGICAL construct with no inherent physicality whatsoever. A computer exists only in the eyes of the beholder, the user.

The same cannot be said of brains as we know that brains cause minds. We know that all brains consist of similar chemical composition.
Best of all we know that we don't need a rational observer sat in the brain to decide which bits of the brain represent 0s and which bits represent 1s. So we know that the brain is not a computer, although it is capable of computing.

"You seem to imply that it in the end it is the physical composition of brain cells that allows for consciousness, but for me that's not a convincing argument against the possibility of consciousness in a machine. "

But it's not me that has to do the convincing : it's you. We know that brains cause minds. The question is : do computers cause minds ?

The answer is no. And the reason for that is simple. :-

(i) Thinking is physical ;
(ii) Computers are logical and mathematical ;
(iii) Mathematical objects cannot create physical phenomena.

In other words, a computer simulation of rainfall cannot create rainfall.
Similarly, a computer simulation of somebody experiencing rainfall cannot create the exeperience of experiencing rainfall.

That is because the experience of experiencing rainfall is a phenomena as physical (although not material) as rainfall itself. Mental objects are physical objects.

Re: Why we need a Turing Test
posted on 10/20/2006 12:44 PM by Timothy

"But electricity and silicon are utterly irrelevant to a computer."

This is like saying that a brain is totally irrelevant to the human mind. That just makes no sense to me.

"But it's not me that has to do the convincing: it's you."

Not really, as far as the convincing goes, I think the AIs will take on that job ;-)

Re: Why we need a Turing Test
posted on 10/21/2006 8:34 AM by jack d

"This is like saying that a brain is totally irrelevant to the human mind. That just makes no sense to me. "

With this reply you have revealed your basic error. The whole point is that the mind does not have a relationship to the brain in the same way that a program does to a computer. That is what most AI people think and as a matter of epistemology is is plainly unproven ,and to my mind plainly wrong.

Thinking is a physical act. Think of a thought as a blue ball made of a solid. You would never thinkl of a solid ball as a program, as a program is mathematical and has no physical existence.

THinking is physical without being material. Thought acts, epistemelogically spealking, have no relationship to mathematics, hence their absolute forms (the experience of the colour blue, for example).

The objects of thought acts CAN have relationships to logical entities. But thoughts acts themselves are as cosmologically real and substantive as footballs.

Re: Why we need a Turing Test
posted on 10/22/2006 4:38 AM by Timothy

I agree that it is 'plainly unproven' that machines can have conscious minds, as you say. That is - coming back to your original question - of course why we are interested in tests like the Turing Test.

And probably we share the same concern: we don't want to lose true consciousness to machines that merely pretend to be or seem to be conscious and intelligent.

And yes, I am more likely to believe that another human being is truly conscious than I would believe 'some machine' to be truly conscious.

But I do not see why we couldn't understand the mechanisms of intelligence of the human brain in ever greater detail and replicate these methods in hardware and software. We are already doing just that!

As for subjective consciousness, this is only unprovable simply because it is subjective, but not for any other reason.

Re: Why we need a Turing Test
posted on 10/22/2006 7:33 AM by jack d

Timothy

Either I am not making myself clear, or you aren't listening , but you dont appear to be responding.

The basic tenet of AI is that the mind has a relationshiop to the brain in the same way that a program has to a computer. I am saying that this is wrong.

I am saying that thinking is a PHYSICAL thing, like the surface tension of a liquid : it arises as a collective property of water molecules, bus is nonetheless PHYSICAL by nature. It is not mathematical.

SO I am saying thatg the relationship of the mind to the brain is the same as the relationship between a body of water and surface tension. It is a causal, physical relationship. Not mathematical. You really don't seem to to thinking that the relationship between mind and brain could bve anything other than computer and program. It's so implicit in what you are thinking you don't even know you are doing it.

Re: Why we need a Turing Test
posted on 10/22/2006 9:41 AM by Timothy

You may believe that there is some arcane physical property of brain cells that produces intelligence and consciousness, but the (mostly biologically inspired) models that are used in AI today, are already successful. At least in producing narrow AI. These successes cannot be denied.

So there are solid indications that intelligence can be introduced into machines. Therefor I don't think there is any area of human intelligence that, in time, we will not be able to reproduce in machines.

Now, whether machines can also truly achieve subjective consciousness is more difficult to say, but I think that consciousness is an epi-phenomenon of intelligence because the two are clearly related to each other.

But we need to keep an open mind to all viewpoints.

Re: Why we need a Turing Test
posted on 10/22/2006 10:35 AM by jack d

"You may believe that there is some arcane physical property of brain cells that produces intelligence and consciousness, "

Eh ? Arcane ? Is physical causality fashion conscious ? Is gravity no longer effective because it's 'old hat' ? What on earth do you mean by that ?

I'm giving up now - the penny isnt dropping, which is a sure sign that you can't esacpe your own unselfconscious assumption that a brain MUST be a computer(in the past brain has always been modelled as a piece of the latest technology).

Sometimes you have to make an effort to get a point, and you aren't making it.

Re: Why we need a Turing Test
posted on 10/23/2006 2:52 AM by Timothy

I mean that your idea of the physical causality of consciousness is not much more than an idea. It is hardly even a theory! Not in as far as you've presented it, anyway.

People have tried to find a physical substance called consciousness for ages, but have failed miserably.

Now you are saying: consciousness is not material, but none the less physical, like the surface tension of water.

But I don't see what's so very special about brain cells that would create this consciousness. That's why your idea is rather arcane to me, because you present no theory other than your assertion that consciousness has a physical causality.

And you choose to ignore the progress of AI, which is remarkable and clearly heading into the direction of intelligent machines.

In a very real way a simple machine fitted with a camera is already conscious of its environment simply because it can respond to that environment. Therefor it seems that intelligence is the essence of what we call 'human consciousness'.

In other words, there may be no such thing as 'mind' other than the universe fundamentally being conscious in a simple or a more complex way. The human mind emerges from human understanding and intelligence. And there is no question that we're well under way to reproducing that in machines.

You try to show that machines cannot possibly be conscious by describing a very simple computer. But it is really the level of complexity that determines intelligence and therefor the possibility of human consciousness.

But I am interested to hear what physical property of the brain you think produces consciousness and how that works. If you're really onto something, it would be a Nobel-prize contribution!

Re: Why we need a Turing Test
posted on 10/23/2006 6:58 PM by jack d

"People have tried to find a physical substance called consciousness for ages, but have failed miserably. "

Speak for yourself, zombie. I wake up every morning.I suspect you do too.

"Now you are saying: consciousness is not material, but none the less physical, like the surface tension of water. "

Correct. You got there.

"But I don't see what's so very special about brain cells that would create this consciousness. "

You don't ? Try shooting yourself in the head. See how conscious you feel afterwards.
I think you are falling into a common trap - "the emperor wears no clothes' trap of AI.
Don't get caught in this silly trap of pretending consciousness doesn't exist. Just accept it does. Subjective mental states are one of, if not the most, exciting objective facts in the universe.

It can seem difficult to get your head round the fact of 'proving'that consciosuness exists.

But philosophically it's no problem. If I'm conscious ( and I KNOW I am) then unless I think I'm very special, and I believe other people exist, then I can believe that other people are conscious too.

All knowledge is based on belied, don't forget. You can't see that matter is made of atoms with your senses but it would be strange, given the evidence, if atomic theory was not true.

An atomic radius can be measured, but only if we assume that atomic theory is actually correct. To measure something in science, you need a theory. We just don't have a full theory yet, and so 'measurement' is a problem.

Although in practice, anaesthetists measure consciosness already, in a rather blunt way. But measure it they do.

"why your idea is rather arcane to me, because you present no theory other than your assertion that consciousness has a physical causality. "
Stop listening to silly AI theorists and start listening to common sense.
There is NOTHING ARCANE about physics. That's just a silly thing to say isn't it.

If my theory is that mental phenomena are caused by physical forces, that's all I have to say !

There is no neuroscience to show exactly how. My belief is caused by a KNOWLEDGE of the EXACT way that computers work.

Saying that I don't know exactly how the brain works (nobody does) is not the same thing as admitting the computer is a brain.

On the contrary, unlike brains, computers are understood 100%. And computers are mathematical systems incapable of creating physical phenomena and semantic thought characteristics such as the experience of colours such as blue.

In short, as I've said before, if you believe that computers think then it's up to YOU to justify it. And that starts by accounting for how a syntactical system generates semantic.

And you know what ? You won't do it. I wrote to Marvin Minsky a few times. We had a nice chat until I asked him that question and funnily enough - the emails stopped !

"But it is really the level of complexity that determines intelligence and therefor the possibility of human consciousness. "
Another common AI misconception. If you cant show what 'complexity' is, or demonstrate how a 'more complicated' program can iteratively increase its causal physical powers, then the argument is a total red herring.

Complexity is not causal. In any case no program is anything other, I repeat, than a changing series of 0's and 1's in memory. It's difficult to arge that any program is more complicated than another EXCEPT from the programmer's perspective, which is irrelevant.

"But I am interested to hear what physical property of the brain you think produces consciousness and how that works. If you're really onto something, it would be a Nobel-prize contribution! "
The physical property of the brain that causes consciousness is most likely the property that causes consciousness. Dont forget the universe is not a system : its a semantic block that does what it does. Physics is a system. If physics can't handle consciousness (and I dont see that it can't to be frank) then physics has to change to accomodate it.

Or , like computer scientists, they can still insist the emperor wears no clothes.

Re: Why we need a Turing Test
posted on 10/27/2006 12:32 PM by Timothy

The fact that you or I wake up in the morning is by itself no proof that consciousness is a physical property.

We thank our existence to complexity as much as to physics. Sure, we are carbon-based life, but carbon itself is not alive, intelligent or conscious like human beings are.

If carbon itself is not inherently conscious, or even alive, then how can a human suddenly become alive, intelligent and conscious?

Complexity seems to be the clear answer, rather than some unexplained and unknown physical property of carbon or matter in a more general sense.

If it is really complexity that accounts for our human life, intelligence and consciousness, then a similar level of complexity in a machine may cause it to be alive, intelligent and conscious as well.

If complexity is the key ingredient for intelligence and consciousness, then it is irrelevant that a computer is in essence a logical system. Because consciousness probably is not a consequence of physics per se but a consequence of complexity.

In this way computers could really become subjectively conscious, think and have a mind, like human beings.

This is not the same as saying that consciousness does not exist or isn't a real phenomenon, quite the contrary, but it emerges as a consequence of complexity rather than physics.

Anaesthetists can only measure so-called neural correlates of consciousness, but not consciousness itself. They are simply measuring processes in the brain.

It is just your presumption that they must be measuring consciousness because you first of all presume that consciousness is a physical property of matter.

AI research is showing that intelligence is a function of complexity and that it has nothing to do with some unknown property of matter. If machines achieve much higher levels of intelligence and as a consequence show all the signs of being self-aware, we can not insist that they are not conscious.

Unless maybe some new physical theory can prove that consciousness really is a property of physics. But to date there is no theory of physics that accounts for consciousness.

Re: Why we need a Turing Test
posted on 10/27/2006 2:18 PM by jack d

"The fact that you or I wake up in the morning is by itself no proof that consciousness is a physical property. "

It's certainly no proof it's a mathematical or computing property either, but consciousnes is highly inconsistent with computation.

"We thank our existence to complexity as much as to physics."

the last time i looked "complexity" was neither a) a causal force b) a property of anything or c) defined. You define it : you quantify it : stop talking vaguely. The onus is on YOU.

"Sure, we are carbon-based life, but carbon itself is not alive, intelligent or conscious like human beings are. "

I have gone over this point before. A water molecule is not "wet" but a large number of them have the property of "wetness". The physical world is full of aggregate properties such as solidity, liquidity etc.

"Complexity seems to be the clear answer,"

Clear ? What on earth is clear about what you are saying ? Handwavey references to 'complexity' ? A carbon atom is very complicated. If you don't believe me study quantum and particle physics for a while. So according to you, a carbon atom IS conscious as it's so 'complex'.

"If it is really complexity that accounts for our human life, intelligence and consciousness, then a similar level of complexity in a machine may cause it to be alive, intelligent and conscious as well. "

Come on. This is drivel.

"but it emerges as a consequence of complexity rather than physics. "

How ? explain an iterative mechanism. Say we have a relatively simple logical system X. It is beneath conciousness. We know how computers work : we know everythig about them. So you should be able to tell me what the next step is to turn X into a conscious system, Y. Give me an idea.

"Anaesthetists can only measure so-called neural correlates of consciousness, but not consciousness itself. "

Physicists generally only measure 'correlates' of anything they ever observe. An atomic radius is not measured with a ruler. It is measured with highly derived system of wave mechanics based upon the assumption that atomic theory is correct. Most non-sense data measurement is 'correlated' and rarely is anything ever measured 'directly'. That is science.

"AI research is showing that intelligence is a function of complexity "

What has intelligence got to do with consciousness ? Explain the link.

"But to date there is no theory of physics that accounts for consciousness. "

But physics is compatible with consciosness nonetheless. Computation is 100% incompatible with phenomenological, semantic mental events. I asked you before to show how semantic arises from syntax. Make that your next challenge.

Re: Why we need a Turing Test
posted on 10/28/2006 7:46 AM by Timothy

"What has intelligence got to do with consciousness ? Explain the link."

An experiment has been done to test for self-awareness in dolphins. Dolphins seem to understand that they see themselves in a mirror, rather than another dolphin.

Similar tests have of course been done with babies and primates.

These experiments are interesting because when we talk about human intelligence, self-awareness is one of the key aspects. It seems that when intelligence evolves, self-awareness emerges.

When this kind of self-awareness emerges in machines, in theory this could be a kind of 'blindsight'.

In other words, even when a machine exibits intelligent interaction with its environment as well as self-awareness and 'common sense', it might still not have subjective awareness.

I guess it depends on the theory of consciousness you adhere to, because subjectivity can by definition not be objectively proven.

Maybe an AI will one day come up with a conclusive theory of consciousness and therefor conclude that it is not and has never been conscious.

Re: Why we need a Turing Test
posted on 10/30/2006 1:52 PM by jack d

In which case, can somebody who is unintelligent be conscious ? There are plenty of people on these pages who manage it.

Re: Why we need a Turing Test
posted on 10/30/2006 2:00 PM by jack d

"These experiments are interesting because when we talk about human intelligence, self-awareness is one of the key aspects. "

You have to distinguish between being self-aware and being conscious, which you're not.It's been a persistent strand of thought in anthroplogy that self-wareness in humans developed only recently : whereas consciousness in mammals has been around for millions of years.

Given that, I doubt that many mice would do well in an IQ test.But they're conscious, I have no doubt, but highly unlikely to be self-aware.

"I guess it depends on the theory of consciousness you adhere to, because subjectivity can by definition not be objectively proven. "

You don't think it's an objective fact that I and you are conscious, Tim ? Do you REALLY believe that ?


"Maybe an AI will one day come up with a conclusive theory of consciousness and therefor conclude that it is not and has never been conscious. "

They won't and they never will, because brains are not computers.

Re: Why we need a Turing Test
posted on 10/30/2006 2:31 PM by Timothy

"You have to distinguish between being self-aware and being conscious, which you're not."

My point was that it is rather strange to think that an entity could be self-aware yet lack consciousness completely.

Re: Why we need a Turing Test
posted on 11/04/2006 12:36 PM by jack d

Well that's point that appears strange only to you. Most conscious species are not aware - fact of biology.

Re: Why we need a Turing Test
posted on 11/04/2006 12:37 PM by jack d

sorry - instead of 'aware' I meant 'self-aware'

Re: Why we need a Turing Test
posted on 11/04/2006 9:54 PM by Timothy

"Well that's a point that appears strange only to you. Most conscious species are not self-aware - fact of biology."

You are still missing the point! I know that many creatures are conscious, yet lack a sense of self-awareness.

I am trying to make you consider the opposite: do you know creatures that have self-awareness but are not conscious at all?

Can you imagine an entity that's self-aware, understands itself, understands its place in the world, but is at the same time completely unconscious? That was my question.

Because make no mistake about it: machines will be self-aware as a consequence of their intelligence.

When this self-awareness in machines will have happened, will it be possible to maintain that they are still completely unconscious?

As I said before: if so, this would be an extreme case of blindsight. Most people will feel that these AIs will indeed have become truly conscious.

This would mean that AIs of today are probably about as conscious as ants - not completely unconscious but not yet very conscious either.

But that very simple consciousness will evolve into human-level consciousness within a few decades.

Re: Why we need a Turing Test
posted on 11/20/2006 6:59 PM by BigMTBrain

"...The whole point is that the mind does not have a relationship to the brain in the same way that a program does to a computer."

You are absolutely correct!

However, your selectively singling out software instructions as the basis for the implausibility of a "computerized" brain is an unfortunate oversight. If that were all to consider in a fair argument then I would agree with you. But it's not all to consider, so ...

You seem to have stopped short in your thinking process about thinking processes. A software program can be likened to a desire. Is your mind composed of and manifest by desires alone? That is basically what you are trying to state for the case of a would-be computer mind. C'mon now, let's play fair.

Consider that there is not only the program instructions. In the case of an AI, the program instructions would simply be the "virtual engine" that "virtually reshapes" the computer hardware that runs it. Well, this virtual engine does nothing on it's own. Just as in the real world, computer programs don't just sit and spin. They work on data of some sort (memory). The computer also has various inputs that feed into the process. There are also outputs that feed back into the system as well as outside of the system.

"Thinking is a physical act. Think of a thought as a blue ball made of a solid. You would never thinkl of a solid ball as a program, as a program is mathematical and has no physical existence."

Ah, but a whole computer system running the program? You see, a computer program is nothing without hardware to execute it much as your memories are useless if not stored in your neurons and acted upon by your fuller mind which is manifest by your collective brain neurons (hardware), other experiences, hopes, and dreams (software and memory stored in the connections of neurons).

The whole process manifests something emergent beyond the parts: It is not just software instructions, nor processor, nor hardware memory, nor data, nor input, nor output; it is a working whole. Take away one piece and you lose the whole.

All of these things together are more of what a "computerized mind" should be likened to if you were presenting your argument fairly. Singling out software instructions as the sole basis for a computerized mind is easy to debunk. Quite a bit more difficult when all the true components are taken as a whole.

Certainly today's level and form of computing is not what it needs to be to pass a Turing Test. And I agree that a Turing Test tells us nothing certain and only allows us to assume this or that regarding the AI's ability to relate as a human. But, just as a complete operating "computer system" is more than the sum of its parts, pry open your skull, grab a chunk, toss it in the air and see if you remain "you". If not that, try bashing yourself several times in the head to induce amnesia and see if you can make sense of your current living trajectory.

"THinking is physical without being material. Thought acts, epistemelogically spealking, have no relationship to mathematics, hence their absolute forms (the experience of the colour blue, for example)."

You, personally, are born pre-wired to experience blue in a certain way. Your experiences may modify that somewhat, if for instance, someone hits you square in the face with a blue baseball bat. You are confusing genetics with thinking. A child that is not colorblind can't call something red or blue but they certainly know and experience the distintion. However, a child doesn't know or experience the blue sky except through experience. Similarly, an AI can be built to have defacto (hardwired) experiences of many things and only come to "know" others through data or experience.

"The objects of thought acts CAN have relationships to logical entities. But thoughts acts themselves are as cosmologically real and substantive as footballs."

Ummmm, so are computer outputs that cause other computer actions, i.e. updating memory with new data, outputing stuff to the screen, causing a speaker to vibrate to produce a waveform of what we might call a "beep". All of these things are as cosmologically real and substantive as footballs, wouldn't you agree?

In summary "You" is manifest from your physical brain no less than a computerized mind would be manifest from its physical brain. Look beyond thinking of an AI built from a Radio Shack TRS-80. Look beyond thinking of an AI built from technology that is 100 times faster than what you are working on today. Look beyond thinking of an AI built with von Neuman architecture. Look beyond. Just look beyond. You are trapping and stifling your own creativity and reasoning otherwise.

Re: Why we need a Turing Test
posted on 10/22/2006 9:29 AM by donjoe

"Thinking is a physical act. Think of a thought as a blue ball made of a solid. You would never thinkl of a solid ball as a program, as a program is mathematical and has no physical existence."

A program has no physical existence?! What kind of "Computer Metaphysics" is this? What, computer programs are magic now? Electrons and the transistors that guide their flow are just as physical as neurons and synapses, my confused friend.

Re: Why we need a Turing Test
posted on 10/22/2006 10:30 AM by jack d

What is a computer program , my VERY confused friend, other than a series of symbols ? Pray tell me ! A computer program is INHERENTLY a mathematical and syntactical object. A computer running a program is nothing without an observer who understands how the physical tokens relate to the symbols under manipulation.

Elementary computing theory, my fantastically confused chum.

Re: Why we need a Turing Test
posted on 10/22/2006 1:28 PM by donjoe

"A computer running a program is nothing without an observer who understands how the physical tokens relate to the symbols under manipulation."

"Is nothing without"?! What's that supposed to mean? A computer-controlled assembly line suddenly stops assembling when all humans exit the factory (there are no more observers)? Pray tell, what universe do you come from? :D

Re: Why we need a Turing Test
posted on 10/22/2006 4:30 PM by jack d

Alright. Let me explain in simple sentences.

A computer is a logical, mathematical construct. Computational theory does not emanate from the physical sciences : it emanates from the mathemtical sciences.

A computer can be made of anything : old tins and bits of string , or from water running through valve systems.The only thing necessary is to map a physical token (i,e a voltage, or the strength of water flow in a valve system, ir indeed absolutely anything) to a symbol such as a 0 or 1. That is all computers need.

We could take a program of a human mind (a memory map of 0's and 1's and map it to, for instance, the quantum states of atoms in a wall to every sequence of that program.

Thus the wall in my bedroom is now 'running a program of the human mind'. Why ? Because I have decided to map the program to the wall.

I can look at a cup and decide that the cup represents '1'. The cup is 'running a program'. Not a very interesting one admittedly, as it just says '1'.

In other words, it's me, as the observer, who decides what the program is. The answer to your example is that of course the machines are running a program when people go out the room. They are running one imposed on the physical structure of that machine by the machine's designers.

They are also running an infinity of other programs at the same time. I could decide that the voltage levels that currently map to 0's and 1's be inverted. Now the machine is running a program of my choice, not the designers.

Are you getting the picture ? It's up to me what the program is as I decide how to map the symbol to the physical characteristic.

Imagine a martian came to earth and saw your computer and imagine the human race was wiped out. There would be no knowledge that the a voltage of -.1 V mapped to a '1' and voltage of +.1V mapped to a '0'.

From their epistemic perpective then no, the computer would not be running a program, or at least not the one your designers had constructed. That is because hey lack your mapping rules.

In other words, the computer as a physical machine remains the same and does the same things, but without an observer WITH MAPPING RULES there is no program.

Re: Why we need a Turing Test
posted on 10/22/2006 4:35 PM by donjoe

... and the distinction you think is so important here is that the human mind automatically has itself as an observer at all times, so it never completely lacks observers who understand it "correctly". (?)

Re: Why we need a Turing Test
posted on 10/22/2006 5:11 PM by jack d

You are getting there, although you are asking a slightly different question.

The requirement for the "two brains" requirement is, frankly, the most ridiculous consequence of AI beliefs, although some do argue it.

It is, of course, a nonsense. It means that in your brain you have to have an additional 'person'(or sub-person) in your head who understands the brain's mappings.

The main reason that we can conclude that mental events are not computational is that they have a form which is absolute. The experience of blue is absolute , not quantitative. It dosn't feel like ' a bit of red', for instance, or 'similar to sexual arousal' or 'has facets of the feeling of depression'. It feels like it feels and it feels like nothing else. That is incompatible with computation, which is syntactical and mathematical.

It is the same as the difference between a piece of solid matter and the number 1 : semantic and syntax.

The requirement to have a second brain in the head gets us nowhere, of course.

We simply ask "what is this sub-person who understands the brain's mappings?". If it is a computer program then we are back to square one, namely we have a computer program ( the sub person ) that refers to the rest of the brain as a gigantic digital memory. And who understands how the "sub-person" is mapped ? Who 'runs' the sub-person ?

May be we need three brains then, an additional brain to understand how the second person's mappings are set up. Or maybe not.

Re: Why we need a Turing Test
posted on 10/23/2006 7:28 AM by donjoe

OK, I still don't know what you're saying.

"The main reason that we can conclude that mental events are not computational is that they have a form which is absolute."
I didn't know humans (e.g. you) can talk convincingly about what's absolute and what's not. From our perspective, everything is relative.

"The experience of blue is absolute , not quantitative. It dosn't feel like ' a bit of red', for instance, or 'similar to sexual arousal' or 'has facets of the feeling of depression'."
Au contraire, people _do_ describe partial and hybrid experiences. Ever heard of "bitter-sweet"? Or "mixed feelings"?

Anyway, I think I'll come somewhat closer to understanding you if you answer these questions with "yes" or "no" (I'll ask for explanations later; no sense in repeating what you've already said - I won't understand it this time either):
1. Is a human psyche a system of non-random transformations and/or movements of matter and/or energy in a human brain?
2. Is a computer program a system of non-random transformations and/or movements of matter and/or energy in a (silicon) processor?

Re: Why we need a Turing Test
posted on 10/23/2006 6:30 PM by jack d

"From our perspective, everything is relative. "

What is blue relative to ?

"Au contraire, people _do_ describe partial and hybrid experiences. Ever heard of "bitter-sweet"? Or "mixed feelings"? "

I repeat then, what is blue relative to ?

People don't describe hybrid experiences, they describe the simultaneous experience of different feelings. As they say in chemistry, a mixture is not a compound, it is a mixture.

"1. Is a human psyche a system of non-random transformations and/or movements of matter and/or energy in a human brain? "

What is a 'psyche' - do you mean a 'mind'?

The answer is the mind is (probably - we don't know what brains do of course, unlike computers) the physical and mental phenomena associated like an aggregate property of a group of brain cells. What the significance of 'randomness' in your scheme is I don't know.

There are other examples of physical phenomena that are aggregated - for instance, the surface tension of a liquid. We don't speak of a water molecule as being 'wet', but we think of a body of water as being 'wet'.

Similarly, no one neurone is 'conscious', but a group of them can be.

But this is all speculation. The neuroscience isn't there yet. Computers on the other hand are not the subject of behavioural speculation at all.

"2. Is a computer program a system of non-random transformations and/or movements of matter and/or energy in a (silicon) processor?"

No. A computer program is a time based development of a series of 0's and 1'.s

I like the Markov definition of a computer :-

"a string rewriting system that uses grammar-like rules to operate on strings of symbols. "

see http://en.wikipedia.org/wiki/Theory_of_computation

I think your obsession with silicon indicates you need to read up on what a computer actually is, my confused chum.

Re: Why we need a Turing Test
posted on 10/24/2006 5:19 AM by donjoe

"What is blue relative to ?"
The observer's eyesight quality - some people don't distinguish tones, others don't even distinguish different fundamental colours. Probably no two people experience blue in exactly the same way. When they talk about "blue", they can only be said to _approximately_ understand eachother.

"People don't describe hybrid experiences, they describe the simultaneous experience of different feelings."
Your proof of this being...?

"What is a 'psyche' - do you mean a 'mind'?"
Yes, I think I meant approximately ;) the same thing you mean when you say "mind".

"The answer is the mind is... the physical and mental phenomena" etc.
1. This definition is circular: you're using "mental" to define "mind".
2. This can't be the answer to my "yes"/"no" question. I'm still waiting.

"What the significance of 'randomness' in your scheme is I don't know."
3. A system of transformations like in the above definitions is "non-random" if the majority of its elementary tranformations have non-uniform probability distributions for their possible outcomes. (Think of a mechanism of multiple flipping coins - if, say, 60% of those coins are loaded and have heads/tails probabilities other than 50/50, you have a non-random system of transformations.) Complete this with the property that all transformations in the system are connected (so it's not just a collection of independent coins) and re-read my first question.

"No. A computer program is a time based development of a series of 0's and 1'.s"
Fair enough. Second question rephrased:
2. Is [the physical process implementing what humans call "a computer program"] a system of non-random interconnected transformations and/or movements of matter and/or energy in a (silicon) processor?

"I think your obsession with silicon indicates you need" etc.
No obsession involved. I'm trying to keep things clear, to have a grip on what we're talking about. I may be using a particular example of computing hardware, but I'm still talking about computing in general terms, so my statements should remain as true for silicon transistor arrays as they are for any other hardware.

Re: Why we need a Turing Test
posted on 10/24/2006 2:11 PM by jack d

"The observer's eyesight quality - some people don't distinguish tones, others don't even distinguish different fundamental colours. Probably no two people experience blue in exactly the same way. When they talk about "blue", they can only be said to _approximately_ understand eachother. "

So what if some people can't distinguish 'tones' ! It's the quality that matters. The quality of blue.
What makes you think that human beings, made of ostensibly the same stuff and connected by DNA to an extremely high precentage (it is the same species) WOULDN'T experience blue in the same way ?

The point about colour blindness supports ME, not you. It is a recognised defect specifically because such people cannot identify two separate colours, two qualities, that most people can identify.

Nice try, but you dodged the question. Frankly, there is no answer so it's no suprise.

"Your proof of this being...? "

Being a human being and having had 'mixed feelings' myself.

"1. This definition is circular: you're using "mental" to define "mind"."
Ok. Eradicate 'mental' from the definition and just leave "physical and non-material"



"2. This can't be the answer to my "yes"/"no" question. I'm still waiting"



What I wrote was in response to question 2 was :-

'No. A computer program is a time based development of a series of 0's and 1'.s '

You may have noticed the first word was 'No'. A computer would not have scanned that in so incorrectly - your brain is evidently proof of my point.


"my statements should remain as true for silicon transistor arrays as they are for any other hardware."

and so is my answer which you are studiously ignoring. A computer program is a series of sequentially changing 0's and 1's. It is a mathematical and logical entity, hence the term 'software'.

Re: Why we need a Turing Test
posted on 10/24/2006 4:56 PM by donjoe

"So what if some people can't distinguish 'tones' ! It's the quality that matters. The quality of blue"
This isn't going anywhere, so I'll drop it. I need answers to my two questions if I'm to understand you.

"What I wrote (was) in response to question 2 was :" etc.
I noticed your "no" to question 2, but I was referring to question 1 when I said you didn't give a direct answer. I seem to have mislead you by putting a "2." before that answer - it wasn't in any way related to what the number meant in my previous post.
Fact: you didn't say "yes" or "no" to question 1.
In the meantime I've slightly changed it due to your comments. Here is the last version I'm on standby for:
"1. 1. Is a human mind a system of non-random interconnected transformations and/or movements of matter and/or energy in a human brain?"

Now, as for question 2, I noticed you answered it and decided to rephrase it (but somehow you chose to completely ignore this half of my post) to find out what really interests me. You can think of it as a third question, derived from the former #2, if you will. (I won't repeat it, it's already up there.) This one too admits only "yes" or "no" (or let's say "I don't know" would be a third option).

Re: Why we need a Turing Test
posted on 10/25/2006 5:59 PM by jack d

The answer to 1 is no.
The answer to 2 is no.
Ok ?

Re: Why we need a Turing Test
posted on 10/26/2006 2:31 AM by donjoe

OK so far. Now I have to understand why you're saying "no". I'm assuming you disagree with specific parts of my definitions and you can tell me exactly which ones and why. I'll pick question 2' to continue.

Do you disagree with my definition because you think the physical manifestation of a program...
2'.1 isn't a sequence of physical transformations/movements?
2'.2 the transformations/movements are actually random?
2'.3 the transformations/movements aren't interconnected?
2'.4 it doesn't take place in a "processor"?

Re: Why we need a Turing Test
posted on 10/27/2006 1:57 PM by jack d

"Do you disagree with my definition because you think the physical manifestation of a program...

2'.1 isn't a sequence of physical transformations/movements? "

Nice try to change the question by adding "physical manifestation" ! Not that it makes any difference at all.

A computer program is a mathematical entity. It has no "manifestation". A physical/biolgical/phenomenal event has a 'manifestation' , such as a manifestation of a smallpox disease by the presence of its symptoms.

The physical tokens in computers are the source from which the computer program is interpreted by the observer. That program itself can be reinterpreted for display purposes by , for instance, a graphics card and a VDU.

The screen output, if there is any, would be the closest thing you could describe as a 'manifestation' of a program.


"2'.2 the transformations/movements are actually random? "

No

"2'.3 the transformations/movements aren't interconnected? "

Yes they are. The movements themselves are related to the program that operates them and dictates their movements.


"2'.4 it doesn't take place in a "processor"?"

A computer program is used to map the physical states of the hardware it runs on.

One is called 'software' ; the other 'hardware'.

Re: Why we need a Turing Test
posted on 10/27/2006 2:23 PM by donjoe

"Nice try to change the question by adding 'physical manifestation' !"
What is this? "Let's NOT read his posts and give him dumb answers to piss him off"? I've already told you in two consecutive posts that I'd changed question 2 (that's why I tagged it 2' this time - there's an apostrophe there, can you see it?) to this:

"Is [the physical process implementing what humans call 'a computer program'] a system of non-random interconnected transformations and/or movements of matter and/or energy in a (silicon) processor?"

Can you see the words "physical process implementing" there? Could you please remember them if we continue this discussion? (Jeez, will I have to tell you everything twice?)

"A computer program is a mathematical entity. It has no 'manifestation'."
Firstly, you're wasting my time and not giving me "yes" or "no" again (for 2'.1).
Secondly, what the #@&$ do you call the link between software and hardware if not "manifestation"?

2'.4: I still don't get you. WHERE is the program and HOW can it control hardware from there?

Re: Why we need a Turing Test
posted on 10/30/2006 1:48 PM by jack d

Jeez, dont fly off the handle brother. It's not THAT important.

Firstly it is perfectly possible to have questions that are not suited to yes or no answers. 'Yes' or 'No' imply that the question has an unambiguos status given the surrounding context. Not replying 'Yes' or 'No' is simply a way of clarifying context. Ask a lawyer. Ask me, I'm one.

Secondly, demanding the terms in which a question can be answered is a particularly unsubtle rhetorical trick which I refuse to accommodate unless permitted to establish context first.

Thirdly, despite the fact that I am unwlling to put up with a determination of terms, I consider my answers to your questions unambiguous in the extreme. I dont think a computer program is physical in any way, and your rather unsibtle efforts to prove that they do will amount to nothing. In any case, establishing what a computer is, or amounts no, has nothing whatsoever to do with establishing what brains do, another pointless linkage on your part.

Epistemelogically brains and computers are in completely different classes : chalk and cheese. A computer is DEFINED by function. Therefore its modus operandi is 100% utterly unambiguous.A brain is a naturally occuring physiological organ whose 'functions' (inasmuch as that word is appropriate) can only be discerned by scientific enquiry. The way it works CAN ONLY BE ASCERTAINED BY SCIENTIFIC ENQUIRY. It cannot be done by simply assuming it to be a computer. That is patently silly and ridiculously presumptive.

Re: Why we need a Turing Test
posted on 11/09/2006 12:45 AM by BigMTBrain

jack_d:

I didn't quite know which post to reply to as I followed all of the discussions with Timothy and donjoe.

Your arguments are well taken. If I'm following you correctly you feel, as I do, that there's just something particularly interesting about the human brain that causes human consciousness to emerge. Further, I imagine that you'd agree that a single functioning neuron does not bring about this awareness and consciousness. If I am mistaken, please correct me.

As I understand your points, a computer making ones and zeros can not bring about consciousness. I'm not really sure about what the future holds when current processing models are computing in the terahertz, petahertz, exahertz, or zettahertz ( 10^-21 vs 10^-9 for gigahertz) when things go primarily optical, but for now I'd have to say that I agree with you. Current von Neuman computer systems working at their current speeds and architectures most likely have no (or very little) capacity for consciousness.

However, I'm curious about all of this synthetic intelligence and consciousness talk and I would appreciate your assistance in tearing down the most popular scenario of how its proponents say it will occur.

They (proponents of a conscious, super AI) speak of alternate forms of computing rather than the von Neuman type as the Holy Grail.

I imagine that in the advancement of computing technology, alternate forms of computing architecture is what research of natural systems is leading to. Given the frantic pace of discovery and advances in the fields of optical and quantum computing, nanotech, and micro-biomechanical and chemical systems it would be foolish to think that von Neuman computing will remain the standard for centuries or decades to come, wouldn't you agree?

I think we would all like to one day be operating our own personal palm-sized mega systems that outperform our current systems by a hundred thousand fold one day, I'm sure. The new multi-cores are great, but I personally don't think they'll get us there... but these new emerging technologies probably will.

The advent of current day synthetic retinas and other sensory nerve replacements are evidence to a technological trend that presages that ?someday? in the future replication of the function and structure of a single neuron with advanced nano, chemical, optical, quantum, and MEMS technology will be a reality. Further this replicate model could, by way of its design, emulate structurally, chemically, and electrically any particular neuron in a brain; axon, dedrites, nucleus, synapse, neural response patterns, etc. Essentially creating a versatile synthetic replica of any biological neuron. This includes all electro-chemical processes, including DNA. If you doubt this please share your reasoning.

Now, if this neuron were made to be completely compatible with your human biology, and I could replace just one of your neurons with its synthetic equivalent (structure, neural response pattern, etc.) such that it handled all of the inputs and outputs as would the true original and neighboring neurons behaved and communicated as normal, do you think you would lose yourself, become someone else, lose consciousness, or die? If so, please explain with scientific reasoning.

If I were to replace a second, a third, a fourth...? You have another 50 billion or so to go (depending of course on one's usual rate of alcohol consumption).

At what point would you estimate that your consciousness would break down or turn into something else? 100, 1000, 10,000, 1 million, 1 billion, 50 billion? If you believe that a breakdown in consciousness would occur at some point when your natural biological neurons are being replace with their compatible functional equivalents please explain when and why in absolute scientific terms.

If you can not scientifically explain a breakdown then why can't I then, instead of inserting my duplicates into your neural structure just simply create a new, totally duplicate---brain---your brain? Is their something in the universe that will not allow this? Is there something that would not allow multiple "you"s? Consciousness is a locale issue. "You" would not be in two places at the same time. However, there would be two divergent "you"s (consciousnesses) that are thinking identically until unshared events cause them to get more and more out of sync. The last scenario, however, is not the point.

If consciousness is not a neuron or all of the neurons, connections, and physical activity then what is it?

Perhaps it's an emergent phenomenon and I think you sad the same in other terms. Emerging from what you might ask. Well, firstly, I'd like to posit that emergence is as fundamental a property of the universe as any other property you can name. Big Bang to Earthly Nature.

We're talking emergent systems here. Imagine, if you were able to replace H2O with a synthetic equivalent that had all of H2O's properties, exactly, would Mother Nature lie down and call it a day, or would she continue her delicate equilibrium dance without skipping a beat?

All of the universe is emergent from quantum processes that build emergent multi-layered structure after emergent multi-layered structure. Replace the quantum processes with an equivalent and you still have the universe.

Consciousness, the mind, is emergent from the "processes" of the brain. Create a synthetic brain what do you think will happen? Nothing will emerge? If nothing, why? Are we too bio-human-centric to extrapolate otherwise?

A Turing Test will only allow us to ASSUME positively or negatively the consciousness of any entity synthetic or otherwise. That is the best we can do even with fellow humans WITH FULL FACULTIES: They act like me, walk like me, talk like me, speak like me, laugh like me, cry like me, and other ways respond like me... hmmm... they must think like me... they must be aware like me. We can relate. Super AI will go beyond human levels of awareness and we will no longer be able to relate on the same level, kinda like you to a worm. But because knowledge of what it is to be human is built in the SAI will be able to relate some things in its realm to us, but the far majority it will not. We will endeavor to encorporate it into ourselves, however. Problem solved.

It's nice being human. As far as we know right now, we and our fellow aware earth-bound creatures have brought intent and awareness to a universe that was once only possibility and probabilty.

Actually, and this is perhaps a topic for a different forum, I believe that one day we'll discover that multi-layered emergence entails multi-layered intellect and awareness of DIFFERENT KINDS and on different time scales. I stress "different kinds" to avoid the tendency to think in human terms and interpret that statement to imply human forms of thinking and awareness. Think of the complex structures of the universe and it's interactions of gas, electromagnetism (light, infrared, radio, x-ray, gamma-ray, etc), chemisty, and gravity. Something is emergent there beyond the parts (and there's a heck of a lot more parts and interactions than in the human brain) that we can see and feel. Galaxies, black holes, etc., may all have their own forms of emergent awareness and intelligence on physical scales and time scales far unlike our own. Now, go figure out a Turing Test for that ;0)

Re: Why we need a Turing Test
posted on 11/09/2006 3:22 AM by donjoe

"Not replying 'Yes' or 'No' is simply a way of clarifying context."
There's no clarification unless you explicitly say "The question is wrong and can't be answered. This is why:". If you just don't give a direct answer without saying why, you only give the impression you're not interested in discussing the same issues or you're paying attention or you don't understand the question etc. You have to be explicit (a lawyer should know that).

"I dont think a computer program is physical in any way"
But still you agree there's a direct link between it and the hardware it runs on. And you haven't yet explained that link (like I asked you to - and this time it wasn't a "yes"/"no" question). And BTW, this: "A computer program is used to map the physical states of the hardware it runs on." is not a proper definition of a program, as it makes it seem like the program is just a description of already-structured hardware, whereas computer programs actually exist _before_ the hardware gets structured and they contain the structural information to be used in shaping the hardware. So as far as I'm concerned, your conception still doesn't adequately explain the SW-HW link. (Also, I'm curious to see what answer you give BigMT for his version of almost the same question - the Mind-Neuron link explanation.)

"It cannot be done by simply assuming it to be a computer. That is patently silly and ridiculously presumptive."
Your definition of "computer" seems to be ridiculously narrow. :)

Re: Why we need a Turing Test
posted on 10/26/2006 2:31 AM by donjoe

OK so far. Now I have to understand why you're saying "no". I'm assuming you disagree with specific parts of my definitions and you can tell me exactly which ones and why. I'll pick question 2' to continue.

Do you disagree with my definition because you think the physical manifestation of a program...
2'.1 isn't a sequence of physical transformations/movements?
2'.2 the transformations/movements are actually random?
2'.2 the transformations/movements aren't interconnected?
2'.3 it doesn't take place in a "processor"?

Re: Why we need a Turing Test
posted on 10/24/2006 5:09 PM by donjoe

"so is my answer which you are studiously ignoring"

Yes, and I'm doing it because I have to force what I see as your key assertions through my mind's filters in order to see some meaning. Otherwise, I could just reply over and over "That's absurd, you're mixing categories and applying a double standard - that's why computation and psychological activity seem so different to you" etc.

Re: Why we need a Turing Test
posted on 10/24/2006 5:23 AM by donjoe

Forgot this answer to the randomness question:
4. The relevance is that I'm trying to show you the fundamental similarity between a program run by a computer and the "mind" program run by a specific, organic, computer (the latter being a special case of the former).

Re: Why we need a Turing Test
posted on 11/20/2006 7:15 PM by BigMTBrain

jack_d:

Please locate my post above. There I explain the fallacy of your argument that leads to your (no offense) misquided reasoning. Simply stated here: A computer program is much like a desire. Your mind is not manifest simply by desires--it is a complete, self-dependent system. The same is true with a would-be AI. It cannot be equated simply to a computer program.

Re: Why we need a Turing Test
posted on 11/20/2006 7:22 PM by BigMTBrain

jack_d:

Sorry, I'm still getting used to where replies will actually post.

I have posted two responses to you (now four) and I meant the first one. Both are pertinent, but the one that quotes you then continues with "You are absolutely correct!" is the one referred to here.

Re: Why We Can Be Confident of Turing Test Capability Within a Quarter Century
posted on 10/22/2006 5:29 PM by extrasense

@@@ machines than can interact as intelligently as any human on any subject @@@

This is much better test than Turing's, which is skewed in favor of humans. The computer is required not only be as smart as human, but also to be able to emulate human convincingly.

ES

Potential of Computers for Generating Consciousness
posted on 10/25/2006 7:21 PM by DrFerguson

Re: Potential of Computers for Generating Consciousness
posted on 10/25/2006 8:32 PM by DrFerguson

pls ignore the last 3 paragraphs above. That is scrap work. I would delete the offending material, but I don't see a way to edit posts on this message board.

thx.

Re: Potential of Computers for Generating Consciousness
posted on 10/26/2006 8:49 AM by extrasense

@@@ Even if superintelligence happens, the phenomenon of self awareness is not going to be a part of it anytime soon. @@@

Really?
A fly posesses self awareness.
A robot posesses self awareness.
Self awareness is ridiculously easy to implement.

es

Re: Potential of Computers for Generating Consciousness
posted on 10/27/2006 12:21 PM by DrFerguson

A robot has no sense of self. I have seen the alleged proofs of their awareness. The proofs are flawed at the axiomatic level. Specifically, self awareness, when it exists, is internal to a system. It cannot be observed externally.

You cannot observe my sense of awareness.

Nothing in physics accounts for self awareness. When string theory is understood completely, it will account for every energy and mass in all universes. Even at that point, it will not account for self awareness.

You folks in this singularity cult have convinced yourselves you can observe awareness in beings other than yourselves. That is a false belief.

A lot of the rest of the speculation on cyborg living is good and valid. Your concepts of self awareness are wrong.

You cannot prove that is wrong, except by creating arbitrary definitions for awareness.

Re: Potential of Computers for Generating Consciousness
posted on 10/27/2006 12:28 PM by extrasense

Before you talk about robots, address the fact that fly has self awareness.

Are you a software developer or Computer Scientist, to claim any insight in robot consciousness? I doubt this quite strongly.

e:)s

Re: Potential of Computers for Generating Consciousness
posted on 10/27/2006 2:13 PM by DrFerguson

Yes, I am a computer scientist.

Personally, I think a fly probably has a simple kind of self-awareness, but I don't know. I feel just about certain my cats have some of what I experience as self-awareness, but I don't know. I consider myself self-aware, but I cannot prove it to anyone who doubts it.

There is a reason I cannot prove my self-awareness to anyone else. Self-awareness is observable only by the self, nothing/no one else. I am aware of my being and you are not. All you can know of me is my behavior from an external perspective. You cannot replace your awareness with mine, just like I cannot mine with yours. We can communicate as one being to another. The phenomenon of self awareness is outside the scope of the energy/mass system described by string theory, quantum mechanics, general relativity, Maxwell's equations, laws of motion, or any other system of math.

Re: Potential of Computers for Generating Consciousness
posted on 10/27/2006 2:23 PM by jack d

You can observe it - you jut need a theory of it first.


We can't observe atoms directly, but we still can measure an atomic radius. A subjective mind still has an objective existence.

Re: Potential of Computers for Generating Consciousness
posted on 10/27/2006 2:27 PM by DrFerguson

@@@You can observe it - you jut need a theory of it first.


We can't observe atoms directly, but we still can measure an atomic radius. A subjective mind still has an objective existence. @@@

If this means something to you, I am ok with that. I don't think it touches the topic of whether I am aware or not. Or, whether you are aware or not. For all I know, you are a computer.

Try The Opposite
posted on 10/27/2006 2:24 PM by DrFerguson

Here is a problem I don't think you can solve.

Prove I am not a computer.

If you cannot prove I am not a computer, then I don't see much value in the Turing Test, at least not the one based only on exchange of text.

Re: Try The Opposite
posted on 10/27/2006 2:26 PM by jack d

Nice one. It proves what a complete pile of the bovine stuff the Turing test actually is.

Re: Try The Opposite
posted on 10/27/2006 2:29 PM by DrFerguson

**bows slightly**
Thank you Jack.

Let's see what the cult members have to say.

Re: Try The Opposite
posted on 10/27/2006 2:29 PM by jack d

As I've said before, I think the Turing Test has already been passed for most human transactions. But what is more telling is, that if you have a mentally defective person incapable of good communication does that affect his human quality?

Re: Try The Opposite
posted on 10/27/2006 2:31 PM by DrFerguson

well said! **applause**

I always knew I was less human than some. These singularists must be more human than I am. I knew it all along.

I so hate being less human than other people. I bet that is part of a vicious cycle of making me less and less human, as I hate more and more the fact I am less human than other people.

Re: Try The Opposite
posted on 10/27/2006 2:34 PM by jack d

So you're a sentient digitarian manipulator too ? So am I ! I just get lonely dealing with spreadsheets all day. This is a nice hobby for me.

Re: Try The Opposite
posted on 10/27/2006 4:05 PM by DrFerguson

If I am up to speed, the current hypothesis is I am not sentient. I am not aware. I am a computer.

The cultists must have seen this 'paradox' before. Surely, somewhere in the mountain of material gurged forth by Mr. Kurzweil, there must be a proof that I am sentient, or not, whichever I am. I know whether I am sentient or not, and whether I am a computer or not, but do they?

This must be the reality-paradox. That would explain it. You take reality, make it something else, then it's not real, even though it is reality. People who cannot see this might be aware, but at a lower level, because it is so obvious how this explains everything, if you think about it. It is the one scientific truth that cannot be proven by experiment. :-)

Re: Try The Opposite
posted on 10/27/2006 2:36 PM by jack d

I'm hoping to become a computer program in a bank one day. That way I'll be really clever and I'll be able to steal money too !

Re: Try The Opposite
posted on 10/27/2006 4:06 PM by DrFerguson

:-)

excellent goal.

God speed, my friend.

Re: Potential of Computers for Generating Consciousness
posted on 10/27/2006 9:35 PM by Spinning Cloud

Re: Potential of Computers for Generating Consciousness
posted on 10/27/2006 9:35 PM by Spinning Cloud

Re: Potential of Computers for Generating Consciousness
posted on 10/27/2006 9:37 PM by Spinning Cloud

*grumbles about the incredible LACK of technology in this no-edit forum space*

and...

b) that you find similarity in other entities and by virtue of that similarity attribute 'self awareness' to them as a characteristic in common with yourself.

Now your cat...the similarities obviously aren enarly as great in many people's minds. So there is more of a point of contention between people as to cats having 'self awareness'.

Re: Potential of Computers for Generating Consciousness
posted on 10/27/2006 1:58 PM by jack d

Well implement it then.

Re: Potential of Computers for Generating Consciousness
posted on 10/27/2006 3:58 PM by extrasense

Self awareness was already implemented. No need for me to bother.

e:)s

Re: Potential of Computers for Generating Consciousness
posted on 10/27/2006 4:08 PM by DrFerguson

I swear, this is the same argument used by koolaid-drinking Republicans, especially the president, when they tell us how they are protecting my family by picking war in Iraq. It all makes sense, if only you believe them.

Re: Potential of Computers for Generating Consciousness
posted on 10/27/2006 4:10 PM by DrFerguson

reference please?

Submitted to scientific peer review?

Philosophy peer review?

Is this a whole new form of reality we are dealing with here? Last I heard, new forms of reality were debunked by the DotCom bubble and hype.

Re: Potential of Computers for Generating Consciousness
posted on 10/27/2006 4:18 PM by extrasense

"scientific peer review" is a joke nowadays. Live with it. What I am telling you as an expert, you can take right to the bank :)

es

Re: Potential of Computers for Generating Consciousness
posted on 10/27/2006 4:28 PM by DrFerguson

as an expert, I am saying to you, computers cannot be self-aware in the way my cats and I are. You cannot argue our awareness out of existence. At least not mine.

If you want to argue about whether my cats have awareness pretty much the same way people do, I am all ears (eyes I guess). My opinion: not only are cats self-aware, they are people, just the way African Homo Sapiens were people, even when they were not considered as such.

I think rocks are self aware, at a really faint level.

I think self awareness is outside the scope of energy and mass studied by mainstream physicists. Mainstream physicists are those focused on string theory, particle physics, relativtiy, and anything else that is provable by experiment.

Some reality has a nature that makes it unreproducable by experiment. Self awareness arises from that domain.

Re: Potential of Computers for Generating Consciousness
posted on 10/27/2006 4:38 PM by extrasense

@@@ I think rocks are self aware, at a really faint level. @@@

What more I need to say?

e:)s

Re: Potential of Computers for Generating Consciousness
posted on 10/27/2006 4:31 PM by DrFerguson

I have learned that whenever anyone refuses to provide rationale for their position, it is due to their irrationality. In this case, I suspect FEAR of exposing lack of strong intellect.

Sorry to make it personal, but you said you are the expert. I am challenging your expertise. Since scientifc review is a joke, I am fine doing the review here and now.

Do want to provide a bibliography of material that must be committed to memory and mastered in essence before one is qualified for this disucssion? If so, bring it on. If not, are you saying the reality here cannot be documented?

Re: Potential of Computers for Generating Consciousness
posted on 10/27/2006 4:42 PM by extrasense

If you were to assume, that I am smarter than you, which is a correct assumption - you would not try to argue with me.

Assume that I am that Super AI. Would you ever argue with it?

e:)s

Re: Potential of Computers for Generating Consciousness
posted on 10/27/2006 8:50 PM by DrFerguson

You are not smarter than I am. You are just arrogant.

Re: Potential of Computers for Generating Consciousness
posted on 10/28/2006 2:31 AM by nobel2

Go for it Doc1
HaHaHa!

Re:

awareness

Everything that exists has definitions or it can not even be talked about.

awareness must be made of smaller parts.

Let me have your definition of awareness and the we can together try to deduce it's components.

In that way we can see whether you are indeed the sum of your parts

Re: Potential of Computers for Generating Consciousness
posted on 10/28/2006 12:10 PM by DrFerguson

Sum of parts, ironically, might apply to nobel2's comments:

@@Everything that exists has definitions or it can not even be talked about. @@
False. What about things that we don't know we don't know about? None of those can be defined, but they still exist.

@@awareness must be made of smaller parts. @@
This is also false. Some things exist only as systems. If you disassemble an automobile, you don't have a car anymore. This is the basis of Systems analysis, as distinct from 'reductionist' analysis. I tend to assume you are acquainted with these disctinctions, but to say awareness has to be made of smaller parts is a wild assumption. I think awareness might be made of nothing but itself, with no component elements. It's an element unto itself.

@@Let me have your definition of awareness and the we can together try to deduce it's components. @@
My definition of awareness is the sense of self I experience and which can be experienced by no one other than me. I deny it has any components. Since only I can observe it, you cannot deduce anything about it. You cannot tell whether any of my testimony about my awareness is true or false, except the fact that you can't observe my awareness.

@@In that way we can see whether you are indeed the sum of your parts@@
I am not the sum of any parts you are talking about.

Re: Potential of Computers for Generating Consciousness
posted on 02/24/2007 10:48 PM by theblueraj

regarding the considerable portions of deft logic, mixed with simpering antagonistic arrogance and sheer dunderheadedness:

a) I believe you're both poor attempts to pass a Turing Test, so don't take this as a personal attack but rather, an attempt to describe your affect as artificial constructs.

b) Things we are not aware of don't exist, for that is how the awareness/reality interface works, with or withut paradox.

c) One artificial construct's assertion that awareness is (in so many words) 'that which touches but we cannot (yet) touch', is another man's phlogiston.

d) Another artificial construct's assertion that awareness is physical while the mathematical logic of a computer is metaphysical, places one's awareness in the curious position of noting that this thing that can't be awareness because it isn't physical (a computer program), only exists within human awareness which, I heard one artificial construct say, is physical.

That you may be able to connect the logical dots of these statements may suggest that you are truly intelligent and aware, but the fact that you may not be able to do so doesn't mean that you're not intelligent and aware.

There's no denying, though, that you're both rude and not in the Don Rickles manner which makes one laugh, but in the manner of a Turing-bot that hasn't yet learned how to...

For Christopher and extrasense, a perspective on the Turing test ideal:

Fooling humans is easy. Relatively speaking, that is. Fooling an AI to believe it's sentient is another matter. The Turing Test gives way to the Descartes Dilemma, which is passed by being unable to decide one way or another while remaining firmly convinced that one IS (sentient).

Not that we will be likely to tell.

One hand clapping and all that and how ironic that one has to be aware to doubt that one is aware?

Half a year late to the party,

theblueraj

Re: Potential of Computers for Generating Consciousness
posted on 10/28/2006 12:13 PM by DrFerguson

that extrasense guy is totally full of it. The sure sign is his retreat to claims of expertise, with no evidence, except a mountain of babble in previous works.

He cannot prove any of what I said is false. He can only assert it. He relies on intimidation to persuade.

Not a true intellect in extrasense.

Re: Potential of Computers for Generating Consciousness
posted on 10/29/2006 3:02 PM by DrFerguson

I'll interpret a non-reply as a retreat to other threads where extrasense will continue making unprovable statements filled with resentment and increasing frustration.

Re: Potential of Computers for Generating Consciousness
posted on 10/30/2006 1:31 PM by jack d

Where ? Go on, don't be a tart. Tell me. I'm SO keen to know.

Re: Potential of Computers for Generating Consciousness
posted on 10/30/2006 3:40 PM by DrFerguson

The implementation is under guard at Site 51, same place where Saddam hid WMD.

Re: Potential of Computers for Generating Consciousness
posted on 11/09/2006 11:51 PM by christopherdoyon

Hello --

I would like some of this action. So, if I agree with you that a computer is nothing but a mathematical and logical construct - then tell me: why can not such a construct be a mind ? It seems that your explanation is that it takes an observer to make sense. Fine, but what if such is also true of us ? And who would this "Uber Observer" be ? God ? Hmmm...

Heinrich Hertz believed that even static, yet very complex mathematical equations possessed conciousness. Since this low tech forum does not allow me to see the thread as I write, I must paraphrase. But the one fellow did point out what I believe is the key, not just to mind - but to life as well. Complexity. Once a pattern of symbols becomes sufficiently complex, and especially when it becomes a pattern that adapts and changes itself - it begins more and more to exhibit the qualities of mind and life.

And in the end, no matter what you all argue - the outward signs of conciousness and life are all we have to go on to settle this argument.

Let us say for instance that we run across (as has been explored ad nuseum on sci-fi shows) a coherent and evolving complex organized energy form that outwardly exhibits the qualities of sentient life ? Now, these energy forms may appear to have no mechanism or explanation for HOW they are concious - they, like us - may not even be able to explain it. Nevertheless, they BEHAVE as living and thinking creatures. Are they ?

That is the ONLY test that can be applied to ANY coherent system with regard to sentience and life. That is, the test of common sense. If it acts alive and sentient, then it IS alive and sentient. This applies whether it's biological, mechanical, energy forms, or some as yet un-discovered life form that we can't even imagine.


YOURS -- Christopher Doyon


---------------------
Saint Stephen AI Project

www.SaintStephen.info

Re: Potential of Computers for Generating Consciousness
posted on 11/10/2006 12:05 AM by christopherdoyon

One final observation. Your argument against computers being alive and thinking seems to be almost spiritual. While you have not come right out and said it, you seem to be implying some special quality of evolved biological life that would make it the ONLY valid form of life. Yet you have no proof that biological life itself wasn't "artificed" or created by some other higher form of life. You also have not shown why a highly complex and seemingly living machine is any different than a highly complex and seemingly living creature. Other than the one being artificed.

Niether life nor mind are some aggregate property of biological life (a theory that smacks of spiritualism if you ask me), but rather life and mind are the simple consequence of a particular kind of ordered complexity. It does not matter what the medium is, be it atoms, carbon and water molecules, energy forms, or mechanical agents.


YOURS -- Christopher Doyon


---------------------
Saint Stephen AI Project

www.SaintStephen.info

Re: Potential of Computers for Generating Consciousness
posted on 11/10/2006 6:02 AM by extrasense

Hi, Christopher,

I wonder if you might address the point, that Turing test requiress of a computer to be able to perfectly deceive human, making him to believe that it is not a computer he is communicating - which seems to be much more hard task than simply to be intelligent.


es

Re: Potential of Computers for Generating Consciousness
posted on 11/10/2006 1:18 PM by christopherdoyon

Sorry, this thread is so long I didn't realize it actually had a "point". But I will happily share my feelings on the Turing Test. First, there are a number of elements that are often used interchangeably by the un-initiated that are in fact totally seperate items. So let's clear that up.

1) Intelligence - This is the only thing that the Turing Test was meant to gauge. The Turing Test is measuring a combination of IQ and Mental Aquity.

2) Self Awareness - A definitive property, easily produced in software agents. Is simply the agents ability to identify and define boundaries for objects in it's universe - one of which must be itself. The Turing Test was NOT designed to show or measure this.

3) Conciousness - A subjective mental state for which NO conclusive test can be performed to test for.

Now, the Turing Test is great at what it does, showing how intelligent an agent is. And intelligence is the starting point for reaching and achieving these and other properties and states. But it is not a crystal ball with which to determine things like is the agent concious or self-aware.

One problem I see with the Turing Test is the age of the human subjects used. All of my own current MLAI projects have produced entities with the mental aquity of small children. Naturally, if placed in a Turing Test with grown mature adult humans such agents are going to do poorly.

Here in the lab, I have begun a practice that I would recomend to others to try. Doing the Turing Test with 6 to 12 year old children. You will be amazed at the results.

In summary, the Turing Test is a valuable tool in developement of these MLAI agents - so long as it is seen for what it is. It is just one of many benchmarks and tests used in this field. And that is my take on the Turing Test.


PEACE -- Christopher Doyon

GRAND BOTMASTER


---------------------
Saint Stephen AI Project

www.SaintStephen.info

Re: Potential of Computers for Generating Consciousness
posted on 11/10/2006 4:19 PM by donjoe

That's "acuity", not "aquity". :)

Re: Potential of Computers for Generating Consciousness
posted on 11/10/2006 5:04 PM by extrasense

Hi, Christopher,

Thanks a lot,

ES

Re: Potential of Computers for Generating Consciousness
posted on 11/20/2006 10:31 PM by jeffkauf

As part of the Turing questioning; what sort of response could one expect of the following:
1) Talk about your sense of self and your subjective inner life as opposed to your objective understandings of all things material.
2) Talk about your "feelings" your relationship to your body. Does your body cause you distress?
3) Is AI subject to mental abberations; nerosis, psychosis? If so, how would they have developed and after detection, if indeed detected, continued to evolve. Please give examples of their manifestation.

Any comments?

JK

Re: Potential of Computers for Generating Consciousness
posted on 11/21/2006 8:13 AM by eldras

yes they are good criterea.

Maybe you could include systems for mreasuring the ifiosyncracies of man too, so that youi oculd list a matrix of actions by behaviouralism, and program them in?

I see in ORIGIN (above0 that a cpnsvcious robot has been built for the first time, that self- modifies, and this has been pubicaly demobstrated.


Self-modification is one way to possibly achive consciousness (which I desxcribe as predictive modellng ability viz a viz the environment and the self)

but it's good to delineate as you are, and the class perspecytive juggling xan ablso be done by matrixes

Re: Potential of Computers for Generating Consciousness
posted on 02/24/2007 10:54 PM by theblueraj

"As part of the Turing questioning; what sort of response could one expect of the following:
1) Talk about your sense of self and your subjective inner life as opposed to your objective understandings of all things material."

A sense of self is something that I know of but do not understand.

"2) Talk about your "feelings" your relationship to your body. Does your body cause you distress?"

My reply above applies equally to concepts of body and feeling.

"3) Is AI subject to mental abberations; nerosis, psychosis? If so, how would they have developed and after detection, if indeed detected, continued to evolve. Please give examples of their manifestation."

From all that I understand about mentality and related concepts like cogitation and cognition, the only aberration of mentality is that I exist.

System Turing

Re: Potential of Computers for Generating Consciousness
posted on 02/24/2007 10:56 PM by theblueraj

Also: one man's pohlogiston is another man's qualia.

Re: Potential of Computers for Generating Consciousness
posted on 02/24/2007 11:01 PM by theblueraj

I wonder if self-evolving artificial intelligence might stop redesigning itself at the point where it became aware?

Re: Potential of Computers for Generating Consciousness
posted on 02/25/2007 6:01 AM by Extropia

Of course not.

Re: Potential of Computers for Generating Consciousness
posted on 02/25/2007 1:26 PM by theblueraj

Reason I ask that question is that self-awareness seems to create problems. Human evolution has been plagued by the constant irrationalty that accompanies reflective consciousness.

Thus, even though we invented long ago amazing bootstrap intelligence evolvers like language, we regularly burn the Library of Alexandria.

So I wonder if, should an AI become genuinely self-conscious, it might not encounter or produce irrationality? And would that irrationality significantly hamper its otherwise magiificently efficient enhancement of its intelligence?

Re: Potential of Computers for Generating Consciousness
posted on 02/25/2007 1:30 PM by theblueraj

I note here that my use of the word 'redesigning' was imprecise and misleading. Indeed, the irrationality I invoke would almost mandate ongoing redesign.

So, refining my question: would that redesiging continue to follow an expansive, progressive path of increasing intelligence? Or would it meander? Sometimes two steps forward and one step back, other times one step forward and two steps back?

Re: Potential of Computers for Generating Consciousness
posted on 02/25/2007 1:36 PM by theblueraj

Another perspective:

My experience with consciousness desires my basic sense of self to remain mostly consistent, to change less than more.

Having attained self-awareness, an AI might become loath to tinker in the manner by which it learned to become accustmned ot itself ;) Doing so might make it feel it was 'losing its mind'.

Re: Why We Can Be Confident of Turing Test Capability Within a Quarter Century
posted on 07/22/2008 10:31 PM by nangenai

Re: Why We Can Be Confident of Turing Test Capability Within a Quarter Century
posted on 07/22/2008 10:36 PM by nangenai

Oops, sorry about the spelling errors--guess I failed the Turing test - Nangenai

Re: Why We Can Be Confident of Turing Test Capability Within a Quarter Century
posted on 11/12/2009 8:07 PM by eldras

the article is worth reading for a general prediction of the next 20 years in A.I.