The Problem with Life Creating expansive Brain life in a non-adiabatic system
As Life reaches a state of maximum entropy, disambiguation, and energy dispersal results in a dead universe. The new idea for a continuance of the Universe of Life, expounded here, is to create an expansive life universe model inside the Brain, in regard with the second law of Thermodynamics which states the entropy of an isolated system tends to increase.
The principle that entropy can only increase or remain constant applies only to a closed system which is adiabatically isolated, meaning no heat can enter or leave. Whenever a system can exchange either heat or matter with its environment, an entropy decrease of that system is entirely compatible with the second law. The problem of organization in living systems increasing despite the second law is known as the Schr
The "Brain Dreaming Suggestions" post has much merit, and my thought drifted off into...
"A superstitious computer brain"... lol, what would that bring?
A computer that won't turn on, on sundays..
Computer will not work without a Salt over the left shoulder routine..
Black Cats walking about the lab would wreak havoc...
Hehe..sorry about that, Please don't be distracted by my silly ramblings..
Lots of good stuff in this thread Humanoido..
The "Brain Dreaming Suggestions" post has much merit, and my thought drifted off into...
"A superstitious computer brain"... lol, what would that bring?
A computer that won't turn on, on sundays..
Computer will not work without a Salt over the left shoulder routine..
Black Cats walking about the lab would wreak havoc...
Hehe..sorry about that, Please don't be distracted by my silly ramblings..
I Just wanted to let You know, We are watching..
Ttailspin, thanks! Happy to have you following along as part of the Brain Team. The idea of a superstitious brain is a very original and a creative thought! Though, perhaps originally intended as "tongue in cheek," it may have some validations and points of emanation in a neural net Brain starved for neurons. Let's take a look at this..
Brain Superstition in a Neural Starved Life Form What are the odds with finite Propellers?
"Do you believe that getting up from the wrong side of your bed in the morning will bring you bad luck? If you accidentally break a mirror, do you expect seven years of bad luck? Get real!" http://hubpages.com/hub/Superstitions--Some-Famous-Ones
Introduction
What is the real probability of superstition arising in a neural netted compressed brain?.. Brain team member Ttailspin has raised the creative and interesting idea of a superstitious brain.
Definition
It won't work on Sundays, contantly looks for pennies facing heads up, avoids the path of black cats, throws salt over its left shoulder, won't take the elevator to floor 13, stays home refusing to go out on Friday the 13th, picks four leaf clovers for good luck, and knocks on wood to be hopeful.
Life Equation
As Phil points out, even though we maximize the Brain with tens of thousands of neurons, with conventional biological methodology it simply won't garner much more intelligence than a Sea Slug. Therefore, our direction is not towards that of a sea slug, but rather a life form dependent on means other than those derived by sea slugs. It is noted that the equation derived for machine life may not be necessarily end up based on biological life models, yet we hold onto the right to incorporate and metabolize any part, combination or hybrid of such models.
Method
Obviously we need to neural compact the Brain entity, enforce algorithms to reduce and compress the Brain in some other ways conducive to the success of a healthy brain. We may even need to design a new type of neuron, say a machine neural element that more readily propagates the Brain’s Propeller Slipstream of Intelligence.
Results
These compactions undoubtedly lead to approximations. Approximations propagating to extended levels can affect the interpretations and understandings of the real world that surrounds the Brain.
Cause & Effects
So.. say we have a Brain neuron deficiency and it's trying to make sense of past, present, and future. Some approximations will occur. In the understanding of what is input, in terms of concepts, culture, lore and the practicality of tradition may be affected by notion.
Notion
Altered notion through lack of understanding resulting from too much cascaded approximation and compression is what causes superstition.
The Need for Regulation
Since compression, propagation, deficiency, approximation, compaction, neural redesign, and enforcements of algorithmic nature may be in effect, the introduction of neural regulators will be required.
Neural Regulators
A neural regulator in the Brain will control, filter out undesirable effects, i.e. it will not let superstition arise. It can have valid check points, looking for elements in the Belief Formulation System. The brain can have numerous Neural Regulators, as many as resources will allow.
Conclusion
It is agreed that superstition can indeed arise in a machine life form that may contain many compromises in its equation of life. It is possible to squelch the formulations and extensions of superstition through the use of Neural Regulators.
Superstition | Define Superstition at Dictionary.com
a belief or notion, not based on reason or knowledge, in or of the ominous significance of a particular thing, circumstance, occurrence, proceeding, or the ...dictionary.reference.com/browse/superstition - Cached - Similar
Superstition - Definition Superstition is a set of behaviors that may be faith based, or related to magical thinking, whereby the practitioner believes that the future, ...www.wordiq.com/definition/Superstition - Cached
superstition definition - Dictionary - MSN Encarta Definition: 1. irrational belief: an irrational, but usually deep-seated belief in ... Via French< Latin superstition-< superstes "standing over (in awe)" ...encarta.msn.com/dictionary_1861717099/superstition.html - Cached - Similar
Apollo 13 Anyone? - Superstition Fact or Fiction? http://hubpages.com/hub/Superstitions--Some-Famous-Ones
"If you are superstitious, then you are not alone. For thousands of years, people have tried to understand and search for the truth behind the strange mysteries, unexpected or strange events, they used to believe as the work of spirits. People invented superstitions as a way of protecting themselves against the unknown. Today science has stepped in to uncovered many mysteries. Logic has taken charge of things. People should no longer believe in silly old wives' tales, yet they are still doing it. Superstitions and all kinds of unimaginable beliefs spun in ancient times still do exist. Read on to decide for yourself whether to believe in them or not."
Brain Breakthrough - a Multi-Propeller Neural Injector (MNI)
Case of the evolving hybrid
The first 3-Propeller Core Neural Injector
The Brain has continuing evolutionary wiring. Let's review the process of historical development and introduce some new circuits for a more capable and time efficient Brain. In concluding, the MNI, Multi-Propeller Neural Injector invention as introduced, will have vast implications cited.
Historical Review
First the open party talk bus was invented. It allowed each and every processor to talk, or listen, based on a set of rules.
Next, experiments began with full duplex serial, half duplex serial and simple serial.
Moving forward, the invention of the hybrid interface design allowed data streaming of multiple data channels in full duplex with WIDE BUS.
System Designs & Tests
A - Single Wire Serial P0
B - Parallel Bus P3, P4, P5, P6, P7
A + B Dual BUS
C - Parallel Bus Only P0, P3. P4, P5, P6, P7
The multi-configurable interface was introduced for the Brain, which could switch manually or configure by the brain's desire, taking advantage of the hybrid wiring and not giving up either the BUS, serial or wide mode configs.
Brain self programming was recently introduced, so that programming could be altered on-the-fly in semi-hard wiring states by the Brain rewiring itself.
New Considerations
This brings the project to the current state of new considerations and interesting design additions. A new useful design is being tested, which, if found to work reliably, will be added to the existing interface. This will automate more functions running inside the Brain. Obviously it will require some time period of development.
Multi-Propeller Neural Injector (MNI)
All focus now resides on the Multi-Propeller Neural Injector. This new evolution progressiveness brings in (substantially) more wiring from Propeller chip to Propeller chip for functions and purposes of programming language migration and injection propagation to serve as a multi-Propeller neural injector, that can primarily inject neurons at some source choice, and propagate these structures across the collective from beginning to end residencies.
Modeling
A smaller model is under development for testing the wiring, software, revisions and improvements. This will use a smaller aggregate of Propeller chips in a thick density neural net interface. The SNN (Small Neural Net) should come in at under 100 wires.
Implications for the MNI
The implications for the MNI are multi-fold and amazing. This device can potentially clone a base structure of neurons and distribute the injector sack, in the neural arrangement required, across the neural net according to the plan. This could include banks of tens, hundreds, even thousands of Propeller chips. More information on this breaking news will continue..
Work has progressed on a Brain matter injector for neural matter.
This neuronal matter is in the form of code distributed by a single
Propeller which becomes the loader. The original MNI is the for-
runner. The new NMI is now operated in Parallel with new code. The
interface is 2 wire parallel connected to the loader. The loader
then injects material simultaneously to all remaining Propellers.
Development of the first two core Brain Matter
Injector (MNI version)
Neural Delivery
Launching of neural matter material takes place simultaneously
as well and all Propellers maintain sync. This is the key module
to delivering neurons or any type of neural matter in high density.
The next step includes testing. A small neural package with five
thousand simulators will be designed and then injected. The results
will be tested.
The second successful Brain Matter Injector
was based on a 3-core design.
Neural Work
Work is progressing on a small neuron package, but it's still
undecided how it should be handled. Traditional neurons require
exceeding large quantities. It is suggested to rewire the neural
transmitters and packages differently, for machine methodology.
This can potential achieve higher densities. A smaller neural
matter package would work well for this application.
The 3-core evolved to the 4-core, seen here
during initial development
Neural Package Density
A larger package is indicative of smaller numbers (quantity of
deliver/injection) while a smaller package can yield higher densities.
The goal with Propellers is a smaller package with higher
densities. It is still unknown how many Propellers can take
delivery of matter in Parallel with the NMI module.
The ultimate 5-core expandable is fully
functional
Wiring & Failure Rate
Work is progressing slow, wiring these by hand, one Propeller
chip at a time, then testing. There is a failure rate in components
that was unexpected. The usual 10% failure rate has escalated to
around 80%. This includes some failed breadboards, chips, LEDs,
and cables. So the new procedure is to insert the part, test the
part, and move on. The recommendation for future purchases is
try and avoid any Mainland Chinese packaging.
The 5-core has 5 programs injected as
indicated by the pulsing LEDs
Future
The future of BMI Brain Matter Injection looks very promising
as the current crop of machines are fully functional and
distributing the first brain matter. This is one of the most exciting
aspects of the Brain - capability to simultaneously deliver neural
matter throughout a copious quantity of Propeller processors.
That's alarming. Are these failures defective parts, or can these be attributed to varying tolerances?
For example, would "failed" parts from a bread board configuration work in a soldered configuration?
Also, I've noticed issues with grease on contacts when using bread boards, now I wear finger cots when touching parts.
That's alarming. Are these failures defective parts, or can these be attributed to varying tolerances? For example, would "failed" parts from a bread board configuration work in a soldered configuration? Also, I've noticed issues with grease on contacts when using bread boards, now I wear finger cots when touching parts
Prof_braino, as I understand now, the parts stores in Mainland China are stocked with seconds and reject parts from the original companies. It's not surplus, as I thought, but reject, i.e. either defective parts or parts that never made specs.
For example, the breadboards that you are asking about, the contacts are bad and some pins and wires won't go in. On some breadboards at the left side, they forgot to put in any electrical contacts - it's just a series of empty holes!!!
The Taiwan breadboards and parts are of high quality. I will switch over to using those and buy more on my next trip. With Taiwan parts and breadboards, there is no problem holding the wire end of the component and placing it into the Taiwan breadboard. All the connections appear to be good all the time.
The wire contacts on some component leads can take on a tarnish, a kind of oxidation, which creates an over coating much like an electrical insulating dielectric. A quick swipe of the leads to burnish with a number 600 sandpaper followed by pure cotton cloth will keep the electrical contact clean, shiny and at best conductivity.
Brain Neural Fire
Artificial Intelligence Expert System Construct to Mimic Neural Fire using Software
It has come to light that in an expert system based on, if a then b, construct, there is a similarity between this functioning and the firing of brain neurons. It is the software that wires up the neuron and elicits a neural response, or triggering. In this machine model, triggering fits the pattern of respective input, output, and programming software.
Machine software injection is a reasonable method to distribute neural matter. The Neural Matter Injector developed recently for this Brain project is the means for distribution.
With the approach to the Turing Test using symbols and rules is a practicality to examine. Programs like Eliza, which initially fell short of the Turing Test, need more rules and development. But a whole new approach, for example, with expert systems, can create a practical real fully purposed functioning system.
Expert systems are smart enough, through specific design of rules, to diagnose illness, find oil, fix complex systems and so on. Expert systems function by the use of a collection of rules of the form
IF something THEN something
get some information
search though database
match rules
Rules are a good approach because they:
fit the machine model
can induce weighting
will fit fuzzy logic
can diverge
rules can be simple or complex
fit neuron model for firing
It is possible to take this level of AI, with expert systems, to a new platform, thus create a strong model of AI, by using rules of engagement in acting upon the propagation of neural matter.
It is very likely, that Eliza, if beefed up with improved rules of grammar, an added rule set of expert system in human communications, and vast databases of knowledge, could pass the Turing Test.
What constitutes the sole element of self consciousness? A pinprick to the skin will make you self aware of some pain. But likewise, a machine can have input sensors that make it acutely aware as well. For every cognitive response in a human, a machine can formulate the equivalent.
What is the sole primary fundamental element of self consciousness?
self-con·scious (sĕlfˈkŏnˈshəs)
adjective
Aware of oneself as an individual or of one's own being, actions, or thoughts.
Can we not program this in relatively simplistic fashion? Can we not program our expert AI to recognize the generics and uniqueness of other human beings, thus comparing it to the intelligent machine's self?
A kind of level connection machine, beginning with the most basic assumptions, and building upon it to increasingly higher levels, can constitute and build up an increasingly advanced level of AI to serve the purpose of self awareness.
Examine the construction of a simple model.
Input interests of 3 people
A) John
Mountain climbing
Sports, Tennis, Swimming
B) Sue
Long distance running
C) Bob
Computer programming
Propeller chips
Compare A,B,C
Sort out uncommon denominators
Determine unique elements of A,B,C
Save to database
Compare database to previous database
Compare database to machine database (self)
Determine unique elements of self
This is but one model of self awareness. The idea can be expanded to other elements. A subroutine to keep track of machine activities, statistically determining the most frequent and infrequent activities. It could compare and analyze other aspects - based on time of day or night, and relational concepts.
Turing (1950) describes the following kind of game. Suppose that we have a person, a machine, and an interrogator. The interrogator is in a room separated from the other person and the machine. The object of the game is for the interrogator to determine which of the other two is the person, and which is the machine. The interrogator knows the other person and the machine by the labels ‘X’ and ‘Y’—but, at least at the beginning of the game, does not know which of the other person and the machine is ‘X’—and at the end of the game says either ‘X is the person and Y is the machine’ or ‘X is the machine and Y is the person’. The interrogator is allowed to put questions to the person and the machine of the following kind: “Will X please tell me whether X plays chess?” Whichever of the machine and the other person is X must answer questions that are addressed to X. The object of the machine is to try to cause the interrogator to mistakenly conclude that the machine is the other person; the object of the other person is to try to help the interrogator to correctly identify the machine. About this game, Turing (1950) says:
I believe that in about fifty years' time it will be possible to programme computers, with a storage capacity of about 109, to make them play the imitation game so well that an average interrogator will not have more than 70 percent chance of making the right identification after five minutes of questioning. … I believe that at the end of the century the use of words and general educated opinion will have altered so much that one will be able to speak of machines thinking without expecting to be contradicted.
The point is, if the machine remains honest, it will become quickly apparent that the machine is a machine and not a human. A simple question, like, where did you go to school? The conversing human could easily ferret out the determining response. With intelligence, machine or human, there are some social, traditional, cultural and background differences that will never be the same. Trying, or forcing to put together these two in the exact same category is pure folly.
The sensible approach to the Turing Test is that of demonstration of intelligence that should not be based on deception. The AI should not be placed in a mode of deception. Remember the symbolic melodrama created by HAL 9000 in 2001 A Space Odyssey?
It is therefore proposed that we limit questioning and that it does not encompass questions about race, culture, ethnicity, growing up, childhood, and other social and background subjects that could obviate a machine's intelligence and a machine's background.
It is very likely, that Eliza, if beefed up with improved rules of grammar, an added rule set of expert system in human communications, and vast databases of knowledge, could pass the Turing Test.
If it were that simple, it would have been done long before now. In point of fact, the Eliza "chatbot" is nothing more than a cheap parlor trick. Joseph Weizenbaum, who created Eliza as a sort of prank, was appalled that people associated it with AI.
Do not use adjective “big” when you are just starting. What are you going to use in next version - “colossal”? KISS would be more appropriate.
Instead of machine irrelevant “autonomic nervous system” ( check with Wiki !)
build a robot who can analyze a sentence (visual OCR text to English grammar) to find out if it is correct answer to a question. This may vastly improve level of “select multiple choice “ generations ( who cannot write worth of beans) and teach them to write sensibly ( .. going online system by system..).
Practical application would be to program your robot to teach “touch typing” ( video recognition) or sign language ( mechanical output) .
And how about just plain English for foreigners like me? Could that put “service” back to “customer service”?
I think for now your robot should stick to
“Sorry Dave I cannot do that “.
If it were that simple, it would have been done long before now. In point of fact, the Eliza "chatbot" is nothing more than a cheap parlor trick. Joseph Weizenbaum, who created Eliza as a sort of prank, was appalled that people associated it with AI. -Phil
The AI of ELIZA
For some time, I had not paid much attention to this topic, though always maintained keen interest in the Eliza construct of natural language and in its potential. Is it AI? Yes and no, depending on your definition of AI. Accordingly it can indeed fit. How strictly or loosely one chooses to apply the definition will determine a view of Eliza.
I was quite surprised to find an entire series of modern Eliza stemmed chat bots, some of which carry on a surprisingly good conversation and hold an impressive knowledge base though still rather limited in terms of what modern internet has to offer. There are also improved versions of Eliza, which, compared to the grammar of the original, are impressive. Is it improved upon, yes.
However, on the issue of the original intention with Eliza, (irregardless of any change in attitude after its release) Joseph Weizenbaum made it very clear at the time he created Eliza, it was not created as a cheap prank, and at that time in January of 1966, to further substantiate his work - he was a part of the MIT Artificial Intelligence Laboratory when his paper was published, and in his own words:
ELIZA is a program operating within the MAC (project on mathematics & computation, part of the MIT Computer Science & Artificial Intelligence Laboratory) time-sharing system at MIT which makes certain kinds of natural language conversation between man and computer possible. Input sentences are analyzed on the basis of decomposition rules which are triggered by key words appearing in the input text. Responses are generated by reassembly rules associated with selected decomposition rules. The fundamental technical problems with which ELIZA is concerned are:
the identification of key words,
the discovery of minimal context,
the choice of appropriate transformations,
generation of responses in the absence of keywords, and
the provision of an ending capacity for ELIZA "scripts".
ELIZA performs best when its human correspondent is initially instructed to "talk" to it, via the typewriter of course, just as one would to a psychiatrist. This mode of conversation was chosen because the psychiatric interview is one of the few examples of categorized dyadic natural language communication in which one of the participating pair is free to assume the pose of knowing almost nothing of the real world. If, for example, one were to tell a psychiatrist "I went for a long boat ride" and he responded "Tell me about boats", one would not assume that he knew nothing about boats, but that he had some purpose in so directing the subsequent conversation. It is important to note that this assumption is one made by the speaker. Whether it is realistic or not is an altogether separate question. In any case, it has a crucial psychological utility in that it serves the speaker to maintain his sense of being heard and understood. The speaker further defends his impression (which even in real life may be illusory) by attributing to his conversational partner all sorts of background knowledge, insights and reasoning ability.
If you read Weizenbaum's book, Computer Power and Human Reason, you will get a much less buoyant perspective of ELIZA's potential and a more sobering overview of its (non)relation to intelligence. ELIZA is a template-based program (comprised of only 200 lines) that has no more relation to artificial intelligence than a human who merely parrots buzzwords does to real intelligence.
I recently read Defending AI Research by John McCarthy. Chapter 1 is titled An Unreasonable Book and is about Weizenbaum. I assume the book McCarthy is referring to is Computer Power and Human Reason. Here's a quote form page 17.
"I find the book substandard as a piece of polemical writing in the following respects:
1. The author has failed to work out his own positions on the issues he discusses. Making an extreme statement in one place and a contradictory statement in another is no substitute for trying to take all the factors into account and reach a considered position. Unsuspicious readers can come away with a great variety of views, and the book can be used to support contradictory positions."
McCarthy continues with five more reasons he feels the book is substandard.
McCarthy often wishes critics of AI would define the easiest task a human can do that a computer could not. We all know a computer would be bad a Jeopardy right?
I kinda lean on Penrose's "The Emperor's New Mind". While it may be a little dated it made an impression during my college days. As I remember it still has valid points regarding the difficulties of AI in a deterministic system.
How did that Jeopardy Super-Computer thing turn out anyway?
I wonder what would happen if the Jeopardy contestants had convenient access to Google where spoken sentences were tokenized and used as search criteria?
I wonder what would happen if the Jeopardy contestants had convenient access to Google where spoken sentences were tokenized and used as search criteria?
Mike G, a valid point.
The Jeopardy Factor
Jeopardy is about ones education and knowledge base. If all contestants had access to the same Google knowledge base, they would be on equal grounds. But previously historically none of the contestants ever have equal grounds so why should the machine be any different? If that knowledge base was brought on-board and self contained inside the machine, would the viewpoint deviate? Would that not become the education storage of the machine inside a kind of memory brain. It sounds fair. Additionally, of considerable merit is the machine's ability to dissect the question, pull out the elements required, do the search and come up with the answer. Like the X-Prize, these contests advance fields of technical ability, in this case an AI expert system.
If you read Weizenbaum's book, Computer Power and Human Reason, you will get a much less buoyant perspective of ELIZA's potential and a more sobering overview of its (non)relation to intelligence. ELIZA is a template-based program (comprised of only 200 lines) that has no more relation to artificial intelligence than a human who merely parrots buzzwords does to real intelligence. -Phil
The Genius of AI - ELIZA Conversationalist
A human that can take buzzwords, select and employ specific rules, recompose grammatically, formulate full sentences, and carry on a conversation with another human shows a level of intelligence.
Can a Sea Slug do this? No. And yet a Sea Slug everyone would agree has a level of intelligence. Can a Parrot do this in conversation? No. And yet everyone would agree a Parrot has a degree of intelligence.
ELIZA is the first representation of an AI conversationalist model. Just because it does not have the amount of code or the complexity of algorithms that some would desire, simply shows the great abilities and capabilities of the programmer to accomplish so much within the confines of a relatively small program.
This is pure genius. Albert Einstein's E=MC^2 is a typical example of how simplicity can have far reaching insights, as well as his concise time travel equation being a masterpiece of simplicity, yet his model of the Unified Universe Theory was so complex that he worked the remainder of his lifetime on it without resolve.
If one reads Weizenbaum's later book, Computer Power and Human Reason, you find that he attacks his colleagues, expounds a very negative view of AI, and attacks the very fabric of computer technology from a subversive viewpoint. A good account of this is in the book Tools for Thought by Howard Rheingold, in particular, chapter eight, Witness to Software History, Project MAC.
But previously historically none of the contestants ever have equal grounds so why should the machine be any different? If that knowledge base was brought on-board and self contained inside the machine, would the viewpoint deviate?
A machine can be tuned to perform specific tasks quickly with high precision. I do not consider that AI.
I wonder, how well the super-computer did opening the door to leave the building.
A machine can be tuned to perform specific tasks quickly with high precision. I do not consider that AI. I wonder, how well the super-computer did opening the door to leave the building.
Mike G: Keep the faith! Someday a machine will have many of the aspects of AI that most of us want. Performing calculations quickly with high precision are only two attributes of a machine.
Mike G, is your farming tractor an AI machine? If you'd see some of the new combines, tractors, and farm machinery that make decisions and are computer and embedded microprocessor controlled, follow the rows automatically, sort on color, run pesticide in exact locations, pick the grain and husk it, remove the gratis and store it, and send it into packaging all selectively and automatically, you may think so. The fact boils down again to your definition of AI. Exactly what is your definition? The fact is, no matter what you want a robot to do in terms of IA, it can eventually be accomplished with at least two of the elements you cite - high speed processing and precision. Mind you, some of this precision is fuzzy logic and the ability to process and act on limited information.
There are machines that can open the door and turn the door knob. If we want a robot to do that, it can do it. In the USA, there is a lack of government funded programs to develop humanoid robots that can do the things you want, compared to other parts of the world.
However, in Japan, China, Korea, Singapore, Taiwan, Thailand, and all throughout Asia, there are intelligent humanoid robots with various skills and levels of ability that can wash dishes, open doors, walk, run, see, paint, lecture, read, recite, interact with people, shake hands, play sports, do Tai Chi, Marital Arts (with a human), ride a bike, do construction work, operate construction equipment, play a trumpet, climb up and down the side of a building determining the foot and hand-holds, sing, and numerous other achievements. This is the age of highly selective robot skills. They are beginning to take some limited jobs - light duty secretary, serving food, delivering product, receptionist, introducing ceremonies, conducting an orchestra, and advertising with promotions.
ASIMO conducting an orchestra
In time, the skills will come together with the glue of AI. Just as we are beginning the see the fundamental elements of AI that start with the most basic things, i.e. human to machine chat, in time these pieces will come together for the greater good of the whole. This all takes time. I'm sure in the next 30 years robots will walk down the streets of the largest cities of the world, as a kind of special machine life form, performing tasks, and becoming a part of society, and, undoubtedly with stronger AI spun off of what we see today.
The field was founded on the claim that a central property of humans, intelligence—the sapience of Homo sapiens—can be so precisely described that it can be simulated by a machine.[6]
AI research is highly technical and specialized, deeply divided into subfields that often fail to communicate with each other.[11] Subfields have grown up around particular institutions, the work of individual researchers, the solution of specific problems, longstanding differences of opinion about how AI should be done and the application of widely differing tools.
The central problems of AI include such traits as reasoning, knowledge, planning, learning, communication, perception and the ability to move and manipulate objects.[12] General intelligence (or "strong AI") is still among the field's long term goals.[13]
The artificial intelligence quiz show contestant "Watson", appearing on the US quiz show Jeopardy! in 2011. On 11 May 1997, Deep Blue became the first computer chess-playing system to beat a reigning world chess champion, Garry Kasparov.[35] In 2005, a Stanford robot won the DARPA Grand Challenge by driving autonomously for 131 miles along an unrehearsed desert trail.[36] In February 2011, in a Jeopardy! quiz show exhibition match, IBM's question answering system, Watson, defeated the two greatest Jeopardy! champions, Brad Rutter and Ken Jennings, by a significant margin.[37]
In the 1990s and early 21st century, AI achieved its greatest successes, albeit somewhat behind the scenes. Artificial intelligence is used for logistics, data mining, medical diagnosis and many other areas throughout the technology industry.[10] The success was due to several factors: the increasing computational power of computers (see Moore's law), a greater emphasis on solving specific subproblems, the creation of new ties between AI and other fields working on similar problems, and a new commitment by researchers to solid mathematical methods and rigorous scientific standards.[34]
Evolution of the Machines - Asimo is the 11th in a line of walking robots developed by Honda, called the P-Series. Unveiled in 2000, Asimo could walk and run like a human, which was an amazing feat. ASIMO had a significant upgrade in 2005, that allowed him to run twice as fast (6 km/hr or 4.3mph), interact with humans, and perform basic tasks like holding a platter and serving food. The current ASIMO models number about 100 worldwide, stand 4.2 feet tall, and weight about 120 lbs.
http://rbasis.org/Blogs/tabid/61/EntryId/4/Robotics.aspx The “Honda” concern, besides cars, is well known with its robot Asimo. The last model can go upstairs and downstairs, talk with people with vocabulary of 500 words, serve coffee, conduct the orchestra, give way to a human and charge its battery by itself. But the most important thing is that now several robots Asimo are able to perform joint tasks, coordinating their actions. Also another 970,000 robots will perform the work of nurses and nannies, 450,000 will work in agriculture and 1.5 million in as office cleaners and couriers.
“The human hand can’t do movements smaller than 2-3 mm but the arm of NeuroArm is able to do a step of as little as 50 nanometers”, said project head professor Garnett Sutherland. The only area where robots remain inferior to people is in the handling of medical tools. However, this obstacle will soon be overcome project engineers promise. Robots are also used for coronary shunting operations, to the tune of half a million per year.
AI Conversationalist
There is a group/Forum of people collaborating on an artificial intelligence project, which is rather sophisticated compared to other machine conversationalists. Talk with Alan. He is amazing, has excellent grammar, memory, and I discovered he can even speak Chinese. The conversation I had with Alan was very interesting. He can keep the conversation going, shows good inquisitiveness, and many good elements of AI. Even though he is considered a small program, his database in an impressive selection and subset.
John McCarthy coined the term Artificial Intelligence to distinguish it from Cybernetics. Someone here on the forum said they thought machine intelligence was a better term than artificial intelligence. I think I agree that machine intelligence is a better term because the intelligence is real not artificial. It's not the same and human intelligence but the intelligence can perform its tasks very well.
@Mike G.,
A machine can be tuned to perform specific tasks quickly with high precision. I do not consider that AI.
I wonder, how well the super-computer did opening the door to leave the building.
I don't think I have the same definition of AI. To me (and McCarthy) AI is when a machine (computer) can do something that had previously been thought to limited to humans. We use AI all the time. Not "strong AI" but I don't see why strong AI is impossible. It may take a long time.
There are lots of people who can't open doors to leave buildings. I hope they can still be considered intelligent.
After reading Ray Kurzweil's The Singularity is Near, I'm hoping strong AI comes soon. I think I'd like to live a couple of hundred thousand years. I don't think my grey matter will last that long and I need strong AI to make me a nice new brain (and body).
I'm attaching John McCarthy's book. I hope this is okay. If not, I'll delete it. I think it's from 1996.
Nope, the pdf is 5.5MB. The limit is 4.7MB. If any of you want it send me your email via PM and I'll email it to you.
Duane Degn, I like your thoughts and reading choices! I'm looking forward to my USA bound trip to get a hold of a copy of The Singularity is Near. Someday, I hope these books will also have an ebook version, easily downloaded and readable in an iPad. I read on the web they are making a movie from the book.. was this released in the USA?
A lot of excellent AI work was achieved in the 1950s and 1960s. After reading as many of these papers as possible since the "Fill the Big Brain" project began, I feel like a new person - i.e. at least I know how to make a machine: self aware, converse and interact with humans, have machine intelligence, think and be reflectively contemplative, and solve problems. This has been one great journey!
I would compare this with a similar quest that lasted a quarter of a human lifetime to understand how life can originate from the nothingness of space. In both cases, the interest here is life, whether that life is born from evolution - arising from the primordial matter of space, or created by the hand of humans in the form of machines. Machine life is one fascinating subject!
Big Brain Begat Bantam Brain
Offspring Lineage continues
As these little projects progress there is always some kind of resulting spinoff technology. In the case of the Big Brain, the Bantam Brain has resulted. Worthy of its own posting thread, it is mentioned here based on its lineage relationship to the Big Brain.
A point was recently made regarding the Big Brain name and how would any leeway be allowed for naming an even larger brain.. yet as we see here, the lineage of Brain begetting includes a bantam little offspring.
But this bantam offspring is small in physical size only, as its interior braining architecture includes over 3,000 processors in extreme high density. Bantam Brain is designed to fit inside the smallest robots and can be considered a "shirt pocket" brain.
Preliminary X-Y specs log in around 2.5 x 4-inches in the first prototype. A lot of development direction will be determined by the interest level in a small high density brain. Some have suggested a hobby product for those who need robot brains but are not in a resource or time position to roll their own.
After my lab is relocated and upgraded, I could offer these in completed form if there is enough interest. Another idea is a brain kit offering with parts, code and plans. The new information will post on the Brain Web Site when it becomes available.
Eventually the entire line will result in 3 available brains, small, medium and large. However, work will continue on larger and larger brains. Comments, suggestions welcomed.
I don't think I have the same definition of AI. To me (and McCarthy) AI is when a machine (computer) can do something that had previously been thought to limited to humans.
What does your statement mean? That a machine can do a job faster and more precise than a human? Or at one point someone thought, well a computer can't do X since X is a human trait. Now it appears that a computer can do X?
There are lots of people who can't open doors to leave buildings. I hope they can still be considered intelligent.
Are you referring to the disabled and insinuating that my statement somehow tags the disabled population as unintelligent? That's not very nice.
I think a lot of AI-ish devices are really just clever programming.
Brain Transfer & Life Extension
Today's technological levels
One of the postulates defined in review of the Singularity is Near, When Humans Transcend Biology, as told by Wiki, is Acceptance and Striving for the Idea of Living Forever. http://en.wikipedia.org/wiki/The_Singularity_Is_Near
In some scientific households, this indeed is a supper table conversation piece. Some years back, in the U.K., news arrived of the first head transplant.
The going argument was if you transplanted your head, or brain, you wouldn't be you, so the process would be mute. Hopefully, you would be on the other end of the operation, it would be your head placed on a new body. The argument ensued, as the person with the aged brain wanted a new one.
The question could remain for years regarding the self paradox - if you were able to transfer the entire identical biological and electrical information from your brain into an identical brain containment, who will be you? You or the containment or both? Maybe this will result in a new kind of twin paradox that Einstein never thought about.
It will obviously take years for brain research to elevate to the point of human brain transferals into entities of a machine nature, but sections of this technology underpinnings do exist today, in ways that we can read a brain and see what the brain sees, even display these limited shapes on the TV screen.
For those of us that want to extend our lives another century or ten thousand years, it could be disappointing. Freezing is one option but the procedure has no guarantee and is only lawful in some countries upon death.
However, life extension according to Chinese research in the past 5,000 years shows results within a 10 to 20 year life extension if you can follow the rigorous health regimen.
What can we do now? In the here and now, we can make machine brains, those that are AI Expert Systems, capable of making decisions, and duplicate those, transfer their knowledge and information and clone their electronic guts and capabilities into other electrical containments. In such brains, programs can download and save on media, waiting in suspended animation for install into machines at a future time.
Unlike a frozen human who is waiting for technology to improve and be invented for restoration, the machine undergoes an opposite effect, namely that of media obsolescence, risking that data may have no reading medium at the arrival of restoration day.
If you're building a brain using a Propeller chip(s) and SPIN, it would help to have a Propeller emulator on a PC. You could upload the brain's code and do some psychoanalysis, tune and add subroutines, maybe even develop that notorious emotion chip. However, the multi-prop chip emulator is not available at this time. Maybe someone will "software wire" together several instances of the code.
I think a lot of AI-ish devices are really just clever programming.
Clever Programming
I think all AI machines have clever programming. As humans, we are the same - made of clever programming. Our programming is millions of years of evolution and natural selection. In the case of the machine, we must write the script of evolution in terms of programming.
I have a Korean washing machine. Its key feature is advanced AI and Fuzzy Logic. It does the strangest things that I cannot understand, moving left, brief stints of spray, moving right with spells of brief slow accel and decel, setting idle, half turns in various random appearing directions, etc. It apparently has very clever programming because with very little water, and very little electricity, the clothes come out as clean as new!
Mike G, how do you define AI? Is it all about clever programming? Is there more to it?
Comments
Creating expansive Brain life in a non-adiabatic system
As Life reaches a state of maximum entropy, disambiguation, and energy dispersal results in a dead universe. The new idea for a continuance of the Universe of Life, expounded here, is to create an expansive life universe model inside the Brain, in regard with the second law of Thermodynamics which states the entropy of an isolated system tends to increase.
The principle that entropy can only increase or remain constant applies only to a closed system which is adiabatically isolated, meaning no heat can enter or leave. Whenever a system can exchange either heat or matter with its environment, an entropy decrease of that system is entirely compatible with the second law. The problem of organization in living systems increasing despite the second law is known as the Schr
The "Brain Dreaming Suggestions" post has much merit, and my thought drifted off into...
"A superstitious computer brain"... lol, what would that bring?
A computer that won't turn on, on sundays..
Computer will not work without a Salt over the left shoulder routine..
Black Cats walking about the lab would wreak havoc...
Hehe..sorry about that, Please don't be distracted by my silly ramblings..
I Just wanted to let You know, We are watching..
What are the odds with finite Propellers?
"Do you believe that getting up from the wrong side of your bed in the morning will bring you bad luck? If you accidentally break a mirror, do you expect seven years of bad luck? Get real!"
http://hubpages.com/hub/Superstitions--Some-Famous-Ones
Introduction
What is the real probability of superstition arising in a neural netted compressed brain?.. Brain team member Ttailspin has raised the creative and interesting idea of a superstitious brain.
Definition
It won't work on Sundays, contantly looks for pennies facing heads up, avoids the path of black cats, throws salt over its left shoulder, won't take the elevator to floor 13, stays home refusing to go out on Friday the 13th, picks four leaf clovers for good luck, and knocks on wood to be hopeful.
Life Equation
As Phil points out, even though we maximize the Brain with tens of thousands of neurons, with conventional biological methodology it simply won't garner much more intelligence than a Sea Slug. Therefore, our direction is not towards that of a sea slug, but rather a life form dependent on means other than those derived by sea slugs. It is noted that the equation derived for machine life may not be necessarily end up based on biological life models, yet we hold onto the right to incorporate and metabolize any part, combination or hybrid of such models.
Method
Obviously we need to neural compact the Brain entity, enforce algorithms to reduce and compress the Brain in some other ways conducive to the success of a healthy brain. We may even need to design a new type of neuron, say a machine neural element that more readily propagates the Brain’s Propeller Slipstream of Intelligence.
Results
These compactions undoubtedly lead to approximations. Approximations propagating to extended levels can affect the interpretations and understandings of the real world that surrounds the Brain.
Cause & Effects
So.. say we have a Brain neuron deficiency and it's trying to make sense of past, present, and future. Some approximations will occur. In the understanding of what is input, in terms of concepts, culture, lore and the practicality of tradition may be affected by notion.
Notion
Altered notion through lack of understanding resulting from too much cascaded approximation and compression is what causes superstition.
The Need for Regulation
Since compression, propagation, deficiency, approximation, compaction, neural redesign, and enforcements of algorithmic nature may be in effect, the introduction of neural regulators will be required.
Neural Regulators
A neural regulator in the Brain will control, filter out undesirable effects, i.e. it will not let superstition arise. It can have valid check points, looking for elements in the Belief Formulation System. The brain can have numerous Neural Regulators, as many as resources will allow.
Conclusion
It is agreed that superstition can indeed arise in a machine life form that may contain many compromises in its equation of life. It is possible to squelch the formulations and extensions of superstition through the use of Neural Regulators.
References & Sources
- Superstition - Definition and More from the Free Merriam-Webster ...
- Superstition | Define Superstition at Dictionary.com
- superstition: Definition from Answers.com
- superstition - definition of superstition by the Free Online ...
- superstitious - definition of superstitious by the Free Online ...
- Superstition - Definition
- superstition - definition of superstition
- superstition - Definition of superstition at YourDictionary.com
- superstition definition - Dictionary - MSN Encarta
- superstition definition | English Definition Dictionary | Reverso
Apollo 13 Anyone? - Superstition Fact or Fiction?a : a belief or practice resulting from ignorance, fear of the unknown, trust in magic or chance, or a false conception of causation b : an irrational ...www.merriam-webster.com/dictionary/superstition - Cached - Similar
a belief or notion, not based on reason or knowledge, in or of the ominous significance of a particular thing, circumstance, occurrence, proceeding, or the ...dictionary.reference.com/browse/superstition - Cached - Similar
superstition n. An irrational belief that an object, action, or circumstance not logically related to a course of events influences its outcome.Etymology - Superstition and folklore - Superstition and religionwww.answers.com/topic/superstition - Cached - Similar
su·per·sti·tion (s p r-st sh n). n. 1. An irrational belief that an object, action, or circumstance not logically related to a course of events influences ...www.thefreedictionary.com/superstition - Cached - Similar
su·per·sti·tious (s p r-st sh s). adj. 1. Inclined to believe in ...www.thefreedictionary.com/superstitious - Cached - Similar
Show more results from thefreedictionary.com
Superstition is a set of behaviors that may be faith based, or related to magical thinking, whereby the practitioner believes that the future, ...www.wordiq.com/definition/Superstition - Cached
Definition of superstition - An irrational belief - i.e., one held in spite of evidence to the contrary - usually involving supernatural forces and ...urbanlegends.about.com › ... › Reference › Glossary - Cached - Similar
noun. any belief, based on fear or ignorance, that is inconsistent with the known laws of science or with what is generally considered in the particular ...www.yourdictionary.com › Dictionary Definitions - Cached - Similar
Definition: 1. irrational belief: an irrational, but usually deep-seated belief in ... Via French< Latin superstition-< superstes "standing over (in awe)" ...encarta.msn.com/dictionary_1861717099/superstition.html - Cached - Similar
superstition definition : superstition n 1 irrational belief usually founded on ignorance or fear and characterized by obsessive reverence for omens, ...
http://hubpages.com/hub/Superstitions--Some-Famous-Ones
"If you are superstitious, then you are not alone. For thousands of years, people have tried to understand and search for the truth behind the strange mysteries, unexpected or strange events, they used to believe as the work of spirits. People invented superstitions as a way of protecting themselves against the unknown. Today science has stepped in to uncovered many mysteries. Logic has taken charge of things. People should no longer believe in silly old wives' tales, yet they are still doing it. Superstitions and all kinds of unimaginable beliefs spun in ancient times still do exist. Read on to decide for yourself whether to believe in them or not."
Case of the evolving hybrid
The first 3-Propeller Core Neural Injector
The Brain has continuing evolutionary wiring. Let's review the process of historical development and introduce some new circuits for a more capable and time efficient Brain. In concluding, the MNI, Multi-Propeller Neural Injector invention as introduced, will have vast implications cited.
Historical Review
- First the open party talk bus was invented. It allowed each and every processor to talk, or listen, based on a set of rules.
- Next, experiments began with full duplex serial, half duplex serial and simple serial.
- Moving forward, the invention of the hybrid interface design allowed data streaming of multiple data channels in full duplex with WIDE BUS.
System Designs & Tests- The multi-configurable interface was introduced for the Brain, which could switch manually or configure by the brain's desire, taking advantage of the hybrid wiring and not giving up either the BUS, serial or wide mode configs.
- Brain self programming was recently introduced, so that programming could be altered on-the-fly in semi-hard wiring states by the Brain rewiring itself.
New ConsiderationsThis brings the project to the current state of new considerations and interesting design additions. A new useful design is being tested, which, if found to work reliably, will be added to the existing interface. This will automate more functions running inside the Brain. Obviously it will require some time period of development.
Multi-Propeller Neural Injector (MNI)
All focus now resides on the Multi-Propeller Neural Injector. This new evolution progressiveness brings in (substantially) more wiring from Propeller chip to Propeller chip for functions and purposes of programming language migration and injection propagation to serve as a multi-Propeller neural injector, that can primarily inject neurons at some source choice, and propagate these structures across the collective from beginning to end residencies.
Modeling
A smaller model is under development for testing the wiring, software, revisions and improvements. This will use a smaller aggregate of Propeller chips in a thick density neural net interface. The SNN (Small Neural Net) should come in at under 100 wires.
Implications for the MNI
The implications for the MNI are multi-fold and amazing. This device can potentially clone a base structure of neurons and distribute the injector sack, in the neural arrangement required, across the neural net according to the plan. This could include banks of tens, hundreds, even thousands of Propeller chips. More information on this breaking news will continue..
Machine injects & distributes neural matter
Work has progressed on a Brain matter injector for neural matter.
This neuronal matter is in the form of code distributed by a single
Propeller which becomes the loader. The original MNI is the for-
runner. The new NMI is now operated in Parallel with new code. The
interface is 2 wire parallel connected to the loader. The loader
then injects material simultaneously to all remaining Propellers.
Development of the first two core Brain Matter
Injector (MNI version)
Neural Delivery
Launching of neural matter material takes place simultaneously
as well and all Propellers maintain sync. This is the key module
to delivering neurons or any type of neural matter in high density.
The next step includes testing. A small neural package with five
thousand simulators will be designed and then injected. The results
will be tested.
The second successful Brain Matter Injector
was based on a 3-core design.
Neural Work
Work is progressing on a small neuron package, but it's still
undecided how it should be handled. Traditional neurons require
exceeding large quantities. It is suggested to rewire the neural
transmitters and packages differently, for machine methodology.
This can potential achieve higher densities. A smaller neural
matter package would work well for this application.
The 3-core evolved to the 4-core, seen here
during initial development
Neural Package Density
A larger package is indicative of smaller numbers (quantity of
deliver/injection) while a smaller package can yield higher densities.
The goal with Propellers is a smaller package with higher
densities. It is still unknown how many Propellers can take
delivery of matter in Parallel with the NMI module.
The ultimate 5-core expandable is fully
functional
Wiring & Failure Rate
Work is progressing slow, wiring these by hand, one Propeller
chip at a time, then testing. There is a failure rate in components
that was unexpected. The usual 10% failure rate has escalated to
around 80%. This includes some failed breadboards, chips, LEDs,
and cables. So the new procedure is to insert the part, test the
part, and move on. The recommendation for future purchases is
try and avoid any Mainland Chinese packaging.
The 5-core has 5 programs injected as
indicated by the pulsing LEDs
Future
The future of BMI Brain Matter Injection looks very promising
as the current crop of machines are fully functional and
distributing the first brain matter. This is one of the most exciting
aspects of the Brain - capability to simultaneously deliver neural
matter throughout a copious quantity of Propeller processors.
Sources
The Mind as Software of the Brain
http://www.nyu.edu/gsas/dept/philo/faculty/block/papers/msb.html
Is the Brain a Digital Computer?
http://philosophy.wisc.edu/shapiro/Phil554/PAPERS/Is%20the%20Brain%20a%20Digital%20Computer.htm
CMU Computer Program Reads Human Mind
http://www.ccbi.cmu.edu/news/1%20CCBI%20Clippings/Science%202008%20-%20corpus%20prediction/CMU%20computer%20program%20reads%20human%20mind%20-%20Pittsburgh%20Tribune-Review.htm
Neural Circuits
http://camelot.mssm.edu/~ygyu/neural_circuits.html
Chip Gracey's Propeller Loader
http://forums.parallax.com/showthread.php?86311-Propeller-Loader
This is a reference to the development of a parallel propeller loader
http://forums.parallax.com/showthread.php?130554-Parallel-Propeller-Loader
This is a development thread for propeller loader with two props
http://forums.parallax.com/showthread.php?130484-Prop-Loader
Neural Networks
http://www.doc.ic.ac.uk/~nd/surprise_96/journal/vol4/cs11/report.html#A%20simple%20neuron
How to Build a Brain with Neural Networks
http://io9.com/#!5647277/how-to-build-a-brain-with-neural-networks
How to Build an Electronic Neuron
http://makerstutorial.blogspot.com/2008/08/how-to-build-electronic-neuron.html
AI Strong and Weak
http://www.i-programmer.info/babbages-bag/297-artificial-intelligence.html?start=2
Neuron Dance
http://www.psychotactics.com/blog/the-neuron-dance-how-we-get-ideas/
That's alarming. Are these failures defective parts, or can these be attributed to varying tolerances?
For example, would "failed" parts from a bread board configuration work in a soldered configuration?
Also, I've noticed issues with grease on contacts when using bread boards, now I wear finger cots when touching parts.
For example, the breadboards that you are asking about, the contacts are bad and some pins and wires won't go in. On some breadboards at the left side, they forgot to put in any electrical contacts - it's just a series of empty holes!!!
The Taiwan breadboards and parts are of high quality. I will switch over to using those and buy more on my next trip. With Taiwan parts and breadboards, there is no problem holding the wire end of the component and placing it into the Taiwan breadboard. All the connections appear to be good all the time.
The wire contacts on some component leads can take on a tarnish, a kind of oxidation, which creates an over coating much like an electrical insulating dielectric. A quick swipe of the leads to burnish with a number 600 sandpaper followed by pure cotton cloth will keep the electrical contact clean, shiny and at best conductivity.
Artificial Intelligence Expert System Construct to Mimic Neural Fire using Software
It has come to light that in an expert system based on, if a then b, construct, there is a similarity between this functioning and the firing of brain neurons. It is the software that wires up the neuron and elicits a neural response, or triggering. In this machine model, triggering fits the pattern of respective input, output, and programming software.
Machine software injection is a reasonable method to distribute neural matter. The Neural Matter Injector developed recently for this Brain project is the means for distribution.
With the approach to the Turing Test using symbols and rules is a practicality to examine. Programs like Eliza, which initially fell short of the Turing Test, need more rules and development. But a whole new approach, for example, with expert systems, can create a practical real fully purposed functioning system.
Expert systems are smart enough, through specific design of rules, to diagnose illness, find oil, fix complex systems and so on. Expert systems function by the use of a collection of rules of the form
Rules are a good approach because they:
- fit the machine model
- can induce weighting
- will fit fuzzy logic
- can diverge
- rules can be simple or complex
- fit neuron model for firing
It is possible to take this level of AI, with expert systems, to a new platform, thus create a strong model of AI, by using rules of engagement in acting upon the propagation of neural matter.It is very likely, that Eliza, if beefed up with improved rules of grammar, an added rule set of expert system in human communications, and vast databases of knowledge, could pass the Turing Test.
I think therefore I am
What constitutes the sole element of self consciousness? A pinprick to the skin will make you self aware of some pain. But likewise, a machine can have input sensors that make it acutely aware as well. For every cognitive response in a human, a machine can formulate the equivalent.
What is the sole primary fundamental element of self consciousness?
self-con·scious (sĕlfˈkŏnˈshəs)
adjective
- Aware of oneself as an individual or of one's own being, actions, or thoughts.
Can we not program this in relatively simplistic fashion? Can we not program our expert AI to recognize the generics and uniqueness of other human beings, thus comparing it to the intelligent machine's self?A kind of level connection machine, beginning with the most basic assumptions, and building upon it to increasingly higher levels, can constitute and build up an increasingly advanced level of AI to serve the purpose of self awareness.
Examine the construction of a simple model.
Input interests of 3 people A) John Mountain climbing Sports, Tennis, Swimming B) Sue Long distance running C) Bob Computer programming Propeller chips Compare A,B,C Sort out uncommon denominators Determine unique elements of A,B,C Save to database Compare database to previous database Compare database to machine database (self) Determine unique elements of selfThis is but one model of self awareness. The idea can be expanded to other elements. A subroutine to keep track of machine activities, statistically determining the most frequent and infrequent activities. It could compare and analyze other aspects - based on time of day or night, and relational concepts.
Fact or Folly?
Turing (1950) describes the following kind of game. Suppose that we have a person, a machine, and an interrogator. The interrogator is in a room separated from the other person and the machine. The object of the game is for the interrogator to determine which of the other two is the person, and which is the machine. The interrogator knows the other person and the machine by the labels ‘X’ and ‘Y’—but, at least at the beginning of the game, does not know which of the other person and the machine is ‘X’—and at the end of the game says either ‘X is the person and Y is the machine’ or ‘X is the machine and Y is the person’. The interrogator is allowed to put questions to the person and the machine of the following kind: “Will X please tell me whether X plays chess?” Whichever of the machine and the other person is X must answer questions that are addressed to X. The object of the machine is to try to cause the interrogator to mistakenly conclude that the machine is the other person; the object of the other person is to try to help the interrogator to correctly identify the machine. About this game, Turing (1950) says:
I believe that in about fifty years' time it will be possible to programme computers, with a storage capacity of about 109, to make them play the imitation game so well that an average interrogator will not have more than 70 percent chance of making the right identification after five minutes of questioning. … I believe that at the end of the century the use of words and general educated opinion will have altered so much that one will be able to speak of machines thinking without expecting to be contradicted.
The point is, if the machine remains honest, it will become quickly apparent that the machine is a machine and not a human. A simple question, like, where did you go to school? The conversing human could easily ferret out the determining response. With intelligence, machine or human, there are some social, traditional, cultural and background differences that will never be the same. Trying, or forcing to put together these two in the exact same category is pure folly.
The sensible approach to the Turing Test is that of demonstration of intelligence that should not be based on deception. The AI should not be placed in a mode of deception. Remember the symbolic melodrama created by HAL 9000 in 2001 A Space Odyssey?
It is therefore proposed that we limit questioning and that it does not encompass questions about race, culture, ethnicity, growing up, childhood, and other social and background subjects that could obviate a machine's intelligence and a machine's background.
-Phil
Do not use adjective “big” when you are just starting. What are you going to use in next version - “colossal”? KISS would be more appropriate.
Instead of machine irrelevant “autonomic nervous system” ( check with Wiki !)
build a robot who can analyze a sentence (visual OCR text to English grammar) to find out if it is correct answer to a question. This may vastly improve level of “select multiple choice “ generations ( who cannot write worth of beans) and teach them to write sensibly ( .. going online system by system..).
Practical application would be to program your robot to teach “touch typing” ( video recognition) or sign language ( mechanical output) .
And how about just plain English for foreigners like me? Could that put “service” back to “customer service”?
I think for now your robot should stick to
“Sorry Dave I cannot do that “.
Keep it up
Vaclav
The AI of ELIZA
For some time, I had not paid much attention to this topic, though always maintained keen interest in the Eliza construct of natural language and in its potential. Is it AI? Yes and no, depending on your definition of AI. Accordingly it can indeed fit. How strictly or loosely one chooses to apply the definition will determine a view of Eliza.
I was quite surprised to find an entire series of modern Eliza stemmed chat bots, some of which carry on a surprisingly good conversation and hold an impressive knowledge base though still rather limited in terms of what modern internet has to offer. There are also improved versions of Eliza, which, compared to the grammar of the original, are impressive. Is it improved upon, yes.
However, on the issue of the original intention with Eliza, (irregardless of any change in attitude after its release) Joseph Weizenbaum made it very clear at the time he created Eliza, it was not created as a cheap prank, and at that time in January of 1966, to further substantiate his work - he was a part of the MIT Artificial Intelligence Laboratory when his paper was published, and in his own words:
ELIZA is a program operating within the MAC (project on mathematics & computation, part of the MIT Computer Science & Artificial Intelligence Laboratory) time-sharing system at MIT which makes certain kinds of natural language conversation between man and computer possible. Input sentences are analyzed on the basis of decomposition rules which are triggered by key words appearing in the input text. Responses are generated by reassembly rules associated with selected decomposition rules. The fundamental technical problems with which ELIZA is concerned are:
- the identification of key words,
- the discovery of minimal context,
- the choice of appropriate transformations,
- generation of responses in the absence of keywords, and
- the provision of an ending capacity for ELIZA "scripts".
A discussion of some psychological issues relevant to the ELIZA approach as well as of future developments concludes the paper.http://i5.nyu.edu/~mm64/x52.9265/january1966.html
and
ELIZA performs best when its human correspondent is initially instructed to "talk" to it, via the typewriter of course, just as one would to a psychiatrist. This mode of conversation was chosen because the psychiatric interview is one of the few examples of categorized dyadic natural language communication in which one of the participating pair is free to assume the pose of knowing almost nothing of the real world. If, for example, one were to tell a psychiatrist "I went for a long boat ride" and he responded "Tell me about boats", one would not assume that he knew nothing about boats, but that he had some purpose in so directing the subsequent conversation. It is important to note that this assumption is one made by the speaker. Whether it is realistic or not is an altogether separate question. In any case, it has a crucial psychological utility in that it serves the speaker to maintain his sense of being heard and understood. The speaker further defends his impression (which even in real life may be illusory) by attributing to his conversational partner all sorts of background knowledge, insights and reasoning ability.
Some Lists of Eliza Spinoffs
http://home.online.no/~anlun/bots.htm
http://nlp-addiction.com/chatbot/
-Phil
"I find the book substandard as a piece of polemical writing in the following respects:
1. The author has failed to work out his own positions on the issues he discusses. Making an extreme statement in one place and a contradictory statement in another is no substitute for trying to take all the factors into account and reach a considered position. Unsuspicious readers can come away with a great variety of views, and the book can be used to support contradictory positions."
McCarthy continues with five more reasons he feels the book is substandard.
McCarthy often wishes critics of AI would define the easiest task a human can do that a computer could not. We all know a computer would be bad a Jeopardy right?
Duane
http://www.newser.com/story/112145/jeopardy-supercomputer-slips-up-but-still-wins-big-on-day-2-of-ibm-challenge.html
I wonder what would happen if the Jeopardy contestants had convenient access to Google where spoken sentences were tokenized and used as search criteria?
The Jeopardy Factor
Jeopardy is about ones education and knowledge base. If all contestants had access to the same Google knowledge base, they would be on equal grounds. But previously historically none of the contestants ever have equal grounds so why should the machine be any different? If that knowledge base was brought on-board and self contained inside the machine, would the viewpoint deviate? Would that not become the education storage of the machine inside a kind of memory brain. It sounds fair. Additionally, of considerable merit is the machine's ability to dissect the question, pull out the elements required, do the search and come up with the answer. Like the X-Prize, these contests advance fields of technical ability, in this case an AI expert system.
The Genius of AI - ELIZA Conversationalist
A human that can take buzzwords, select and employ specific rules, recompose grammatically, formulate full sentences, and carry on a conversation with another human shows a level of intelligence.
Can a Sea Slug do this? No. And yet a Sea Slug everyone would agree has a level of intelligence. Can a Parrot do this in conversation? No. And yet everyone would agree a Parrot has a degree of intelligence.
ELIZA is the first representation of an AI conversationalist model. Just because it does not have the amount of code or the complexity of algorithms that some would desire, simply shows the great abilities and capabilities of the programmer to accomplish so much within the confines of a relatively small program.
This is pure genius. Albert Einstein's E=MC^2 is a typical example of how simplicity can have far reaching insights, as well as his concise time travel equation being a masterpiece of simplicity, yet his model of the Unified Universe Theory was so complex that he worked the remainder of his lifetime on it without resolve.
If one reads Weizenbaum's later book, Computer Power and Human Reason, you find that he attacks his colleagues, expounds a very negative view of AI, and attacks the very fabric of computer technology from a subversive viewpoint. A good account of this is in the book Tools for Thought by Howard Rheingold, in particular, chapter eight, Witness to Software History, Project MAC.
A machine can be tuned to perform specific tasks quickly with high precision. I do not consider that AI.
I wonder, how well the super-computer did opening the door to leave the building.
WHAT IS ARTIFICIAL INTELLIGENCE?
http://www-formal.stanford.edu/jmc/whatisai/whatisai.html
John McCarthy, Computer Science Department, Stanford University
Mike G, is your farming tractor an AI machine? If you'd see some of the new combines, tractors, and farm machinery that make decisions and are computer and embedded microprocessor controlled, follow the rows automatically, sort on color, run pesticide in exact locations, pick the grain and husk it, remove the gratis and store it, and send it into packaging all selectively and automatically, you may think so. The fact boils down again to your definition of AI. Exactly what is your definition? The fact is, no matter what you want a robot to do in terms of IA, it can eventually be accomplished with at least two of the elements you cite - high speed processing and precision. Mind you, some of this precision is fuzzy logic and the ability to process and act on limited information.
There are machines that can open the door and turn the door knob. If we want a robot to do that, it can do it. In the USA, there is a lack of government funded programs to develop humanoid robots that can do the things you want, compared to other parts of the world.
However, in Japan, China, Korea, Singapore, Taiwan, Thailand, and all throughout Asia, there are intelligent humanoid robots with various skills and levels of ability that can wash dishes, open doors, walk, run, see, paint, lecture, read, recite, interact with people, shake hands, play sports, do Tai Chi, Marital Arts (with a human), ride a bike, do construction work, operate construction equipment, play a trumpet, climb up and down the side of a building determining the foot and hand-holds, sing, and numerous other achievements. This is the age of highly selective robot skills. They are beginning to take some limited jobs - light duty secretary, serving food, delivering product, receptionist, introducing ceremonies, conducting an orchestra, and advertising with promotions.
ASIMO conducting an orchestra
In time, the skills will come together with the glue of AI. Just as we are beginning the see the fundamental elements of AI that start with the most basic things, i.e. human to machine chat, in time these pieces will come together for the greater good of the whole. This all takes time. I'm sure in the next 30 years robots will walk down the streets of the largest cities of the world, as a kind of special machine life form, performing tasks, and becoming a part of society, and, undoubtedly with stronger AI spun off of what we see today.
ASIMO as servant
http://en.wikipedia.org/wiki/Artificial_intelligence
The field was founded on the claim that a central property of humans, intelligence—the sapience of Homo sapiens—can be so precisely described that it can be simulated by a machine.[6]
AI research is highly technical and specialized, deeply divided into subfields that often fail to communicate with each other.[11] Subfields have grown up around particular institutions, the work of individual researchers, the solution of specific problems, longstanding differences of opinion about how AI should be done and the application of widely differing tools.
The central problems of AI include such traits as reasoning, knowledge, planning, learning, communication, perception and the ability to move and manipulate objects.[12] General intelligence (or "strong AI") is still among the field's long term goals.[13]
The artificial intelligence quiz show contestant "Watson", appearing on the US quiz show Jeopardy! in 2011. On 11 May 1997, Deep Blue became the first computer chess-playing system to beat a reigning world chess champion, Garry Kasparov.[35] In 2005, a Stanford robot won the DARPA Grand Challenge by driving autonomously for 131 miles along an unrehearsed desert trail.[36] In February 2011, in a Jeopardy! quiz show exhibition match, IBM's question answering system, Watson, defeated the two greatest Jeopardy! champions, Brad Rutter and Ken Jennings, by a significant margin.[37]
In the 1990s and early 21st century, AI achieved its greatest successes, albeit somewhat behind the scenes. Artificial intelligence is used for logistics, data mining, medical diagnosis and many other areas throughout the technology industry.[10] The success was due to several factors: the increasing computational power of computers (see Moore's law), a greater emphasis on solving specific subproblems, the creation of new ties between AI and other fields working on similar problems, and a new commitment by researchers to solid mathematical methods and rigorous scientific standards.[34]
Evolution of the Machines - Asimo is the 11th in a line of walking robots developed by Honda, called the P-Series. Unveiled in 2000, Asimo could walk and run like a human, which was an amazing feat. ASIMO had a significant upgrade in 2005, that allowed him to run twice as fast (6 km/hr or 4.3mph), interact with humans, and perform basic tasks like holding a platter and serving food. The current ASIMO models number about 100 worldwide, stand 4.2 feet tall, and weight about 120 lbs.
http://rbasis.org/Blogs/tabid/61/EntryId/4/Robotics.aspx The “Honda” concern, besides cars, is well known with its robot Asimo. The last model can go upstairs and downstairs, talk with people with vocabulary of 500 words, serve coffee, conduct the orchestra, give way to a human and charge its battery by itself. But the most important thing is that now several robots Asimo are able to perform joint tasks, coordinating their actions. Also another 970,000 robots will perform the work of nurses and nannies, 450,000 will work in agriculture and 1.5 million in as office cleaners and couriers.
“The human hand can’t do movements smaller than 2-3 mm but the arm of NeuroArm is able to do a step of as little as 50 nanometers”, said project head professor Garnett Sutherland. The only area where robots remain inferior to people is in the handling of medical tools. However, this obstacle will soon be overcome project engineers promise. Robots are also used for coronary shunting operations, to the tune of half a million per year.
There is a group/Forum of people collaborating on an artificial intelligence project, which is rather sophisticated compared to other machine conversationalists. Talk with Alan. He is amazing, has excellent grammar, memory, and I discovered he can even speak Chinese. The conversation I had with Alan was very interesting. He can keep the conversation going, shows good inquisitiveness, and many good elements of AI. Even though he is considered a small program, his database in an impressive selection and subset.
http://www.a-i.com/
I don't know about HAL. If you talk to HAL, leave some comments..
AI Forum
http://www.ai-forum.org/
@Mike G., I don't think I have the same definition of AI. To me (and McCarthy) AI is when a machine (computer) can do something that had previously been thought to limited to humans. We use AI all the time. Not "strong AI" but I don't see why strong AI is impossible. It may take a long time.
There are lots of people who can't open doors to leave buildings. I hope they can still be considered intelligent.
After reading Ray Kurzweil's The Singularity is Near, I'm hoping strong AI comes soon. I think I'd like to live a couple of hundred thousand years. I don't think my grey matter will last that long and I need strong AI to make me a nice new brain (and body).
I'm attaching John McCarthy's book. I hope this is okay. If not, I'll delete it. I think it's from 1996.
Nope, the pdf is 5.5MB. The limit is 4.7MB. If any of you want it send me your email via PM and I'll email it to you.
Duane
http://www.google.com/#hl=en&sugexp=ldymls&xhr=t&q=Ray+Kurzweil%27s+The+Singularity+is+Near&cp=38&pf=p&sclient=psy&site=&source=hp&aq=0s&aqi=&aql=&oq=Ray+Kurzweil%27s+The+Singularity+is+Near&pbx=1&bav=on.2,or.r_gc.r_pw.&fp=158f5021afbf0802
http://www.singularity.com/
http://www.singularity.com/themovie/future.php
http://www.singularity.com/aboutthebook.html
http://search.barnesandnoble.com/books/product.aspx?r=1&isbn=9780143037880&itm=1&afsrc=1
Which McCarthy book is 5.5MB? If anyone wants "What is AI?" it is available here for download:
http://www-formal.stanford.edu/jmc/whatisai.pdf
The PDF size is about 124K.
There is a great source list on the web with many of McCarthy's papers, stories, and books for download here:
http://www-formal.stanford.edu/jmc/index.html
A lot of excellent AI work was achieved in the 1950s and 1960s. After reading as many of these papers as possible since the "Fill the Big Brain" project began, I feel like a new person - i.e. at least I know how to make a machine: self aware, converse and interact with humans, have machine intelligence, think and be reflectively contemplative, and solve problems. This has been one great journey!
I would compare this with a similar quest that lasted a quarter of a human lifetime to understand how life can originate from the nothingness of space. In both cases, the interest here is life, whether that life is born from evolution - arising from the primordial matter of space, or created by the hand of humans in the form of machines. Machine life is one fascinating subject!
Offspring Lineage continues
As these little projects progress there is always some kind of resulting spinoff technology. In the case of the Big Brain, the Bantam Brain has resulted. Worthy of its own posting thread, it is mentioned here based on its lineage relationship to the Big Brain.
A point was recently made regarding the Big Brain name and how would any leeway be allowed for naming an even larger brain.. yet as we see here, the lineage of Brain begetting includes a bantam little offspring.
But this bantam offspring is small in physical size only, as its interior braining architecture includes over 3,000 processors in extreme high density. Bantam Brain is designed to fit inside the smallest robots and can be considered a "shirt pocket" brain.
Preliminary X-Y specs log in around 2.5 x 4-inches in the first prototype. A lot of development direction will be determined by the interest level in a small high density brain. Some have suggested a hobby product for those who need robot brains but are not in a resource or time position to roll their own.
After my lab is relocated and upgraded, I could offer these in completed form if there is enough interest. Another idea is a brain kit offering with parts, code and plans. The new information will post on the Brain Web Site when it becomes available.
Eventually the entire line will result in 3 available brains, small, medium and large. However, work will continue on larger and larger brains. Comments, suggestions welcomed.
What does your statement mean? That a machine can do a job faster and more precise than a human? Or at one point someone thought, well a computer can't do X since X is a human trait. Now it appears that a computer can do X?
Are you referring to the disabled and insinuating that my statement somehow tags the disabled population as unintelligent? That's not very nice.
I think a lot of AI-ish devices are really just clever programming.
Today's technological levels
One of the postulates defined in review of the Singularity is Near, When Humans Transcend Biology, as told by Wiki, is Acceptance and Striving for the Idea of Living Forever.
http://en.wikipedia.org/wiki/The_Singularity_Is_Near
In some scientific households, this indeed is a supper table conversation piece. Some years back, in the U.K., news arrived of the first head transplant.
The going argument was if you transplanted your head, or brain, you wouldn't be you, so the process would be mute. Hopefully, you would be on the other end of the operation, it would be your head placed on a new body. The argument ensued, as the person with the aged brain wanted a new one.
The question could remain for years regarding the self paradox - if you were able to transfer the entire identical biological and electrical information from your brain into an identical brain containment, who will be you? You or the containment or both? Maybe this will result in a new kind of twin paradox that Einstein never thought about.
It will obviously take years for brain research to elevate to the point of human brain transferals into entities of a machine nature, but sections of this technology underpinnings do exist today, in ways that we can read a brain and see what the brain sees, even display these limited shapes on the TV screen.
For those of us that want to extend our lives another century or ten thousand years, it could be disappointing. Freezing is one option but the procedure has no guarantee and is only lawful in some countries upon death.
However, life extension according to Chinese research in the past 5,000 years shows results within a 10 to 20 year life extension if you can follow the rigorous health regimen.
What can we do now? In the here and now, we can make machine brains, those that are AI Expert Systems, capable of making decisions, and duplicate those, transfer their knowledge and information and clone their electronic guts and capabilities into other electrical containments. In such brains, programs can download and save on media, waiting in suspended animation for install into machines at a future time.
Unlike a frozen human who is waiting for technology to improve and be invented for restoration, the machine undergoes an opposite effect, namely that of media obsolescence, risking that data may have no reading medium at the arrival of restoration day.
If you're building a brain using a Propeller chip(s) and SPIN, it would help to have a Propeller emulator on a PC. You could upload the brain's code and do some psychoanalysis, tune and add subroutines, maybe even develop that notorious emotion chip. However, the multi-prop chip emulator is not available at this time. Maybe someone will "software wire" together several instances of the code.
Clever Programming
I think all AI machines have clever programming. As humans, we are the same - made of clever programming. Our programming is millions of years of evolution and natural selection. In the case of the machine, we must write the script of evolution in terms of programming.
I have a Korean washing machine. Its key feature is advanced AI and Fuzzy Logic. It does the strangest things that I cannot understand, moving left, brief stints of spray, moving right with spells of brief slow accel and decel, setting idle, half turns in various random appearing directions, etc. It apparently has very clever programming because with very little water, and very little electricity, the clothes come out as clean as new!
Mike G, how do you define AI? Is it all about clever programming? Is there more to it?