Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Human–computer interaction wikipedia , lookup
Visual Turing Test wikipedia , lookup
Intelligence explosion wikipedia , lookup
Philosophy of artificial intelligence wikipedia , lookup
Wizard of Oz experiment wikipedia , lookup
Embodied cognitive science wikipedia , lookup
Ethics of artificial intelligence wikipedia , lookup
Existential risk from artificial general intelligence wikipedia , lookup
AI IN THE NEWS (Spring 2004) This is more or less a chronological list of links that were sent to me during the course, with only a little editing. The contributor’s name is highlighted either in yellow (we have not looked at the site yet in class), or in turquoise (if we have looked at the site). Hi Morre, Attached is a link to an AI / Neural network for dealing with the spam problem. It has some nice ideas. http://www.generation5.org/content/2004/KillSpam.asp The mother site of this article looks like a professional AI site as well: www.generation5.org Rani BEN: .מורי שלום רב זה אתר שמתעסק בעניינים שבקדמת הבמה בכל מני תחומים. . יש שם הרבה מאמרים מעניינים.(טכנולוגיים בעיקר) ובניהם בינה מלאכותית ההפניה הראשונה שם היא.מרכיב מומלץ בחום הוא "המוח" זה לינק בתוך האתר בנוגע . מקווה שתהנה מהאתרAI. ל http://www.kurzweilai.net BEN: Hi Morre. Here is the link to the "common sense" draft by Minsky. If you would like i can send you the rest of the drafts of his book. (there are 7 chapters). Best Regard Ben Friedman http://web.media.mit.edu/~minsky/E6/eb6.html http://web.media.mit.edu/~minsky/ Hi Morre, This is the article I would like to present in the next AI class. Thanks, Liran. Within three to five years, we could see a search engine that could extract specific facts, draw inferences and organize those facts based on a few key words http://www.computerworld.com/softwaretopics/software/appdev/story/0,10801,86988 ,00.html Rani: http://www.globetechnology.com/servlet/ArticleNews/gtnews/TGAM/20030322/FCR OBO Hi Morre, This is a link that describes a baby bot / AI upbrought in MIT. I made a Tamzit of the page: "working from the bottom up and organizing itself, like life does, into greater and greater complexity." They plan before long to give Ripley a verbal-comprehension test administered to children. They expect that, as he is taught increasingly abstract words, he will increasingly develop the capacity for abstract reasoning. "Ripley, how heavy is this?" "How heavy is what?" A metal block is placed in Ripley's mouth. He flexes his neck up and down, hefting the block, as a human would flex his arm to test the weight of an object. "It is very heavy." "It's perfectly aware and capable of knowing how heavy things are," says Kaiyuh Hsiao, 25. "I simply haven't programmed it to feel pain [when it lifts something too heavy]. The same way as a human, its motors heat -- its muscles get sore -- and I think there's a very great analogue there, but we haven't chosen to make use of it yet." Ripley learns language by looking at an object, touching it and hearing the word for it. In the media lab it is called "grounding." The sight and touch of an object is linked -- grounded -- to the sound of thespoken word. It is exactly how a human infant learns her first words. Ripley is about to have his few fragments of canned, programmed speech deleted -- the equivalent of George W. Bush losing his Teleprompter. Thereafter, Ripley will speak to humans only with the language he has learned from his environment. If it isn't out there for Ripley to experience somehow, he won't have a word for it. Even the grammar of language helps. "By virtue of Ripley's perceptual system," Dr. Roy says, "the world is parsed into objects -- bean bags, balls, etc. -- that have properties: Blue. Large. Heavy. This representation of the world as objects with properties forms the basis for nouns and adjectives. "This is an important point -- that the perceptual system provides a starting point for the language system to create parts of speech, which in turn is the basis for learning syntax. Over time, the learner can use this starting point . . . [to] understand that nouns reach beyond objects. Think of 'the heat,' 'the flight.' " "By the time you have traced all connections from language to non-linguistic aspects of the human -- perception, motor control, planning, visual imagery, and so on -- there will be no aspect of a human left disconnected from language," says Dr. Roy. "The human genome is the language genome." The virtual simulator with which Ripley is now being equipped, says graduate student Nick Mavridis, 29, from Greece, will enable the robot to see himself bend down to muzzle a beanbag, giving him an experience of his body. The simulator also will allow Ripley to retain an image of objects he saw a moment ago but which are now, for example, behind him. If he is told to find a green beanbag that is behind him, he will know where to locate it by "looking" into his simulator. In linguistics terms, it is the third side of what is known as the semiotic triangle: There's the real object (the green beanbag), the label (the name of the beanbag) and the thought (the image of the beanbag the simulator provides). Likewise, if a ball rolls past him, the simulator will enable Ripley to predict where it is rolling to. Which means Ripley will have to understand the laws of physics. Eventually, he will distinguish between animate and inanimate objects -- a distinction, Mr. Mavridis notes, that language doesn't always make. "We say, 'The car is coming,' when it is the driver of the car making it come." Rani: Quantum computing gets a step closer http://www.nature.com/nsu/040308/040308-8.html Hi, It is an article about an attempt to implement computerized gaze control for face perception. http://www.cs.umass.edu/~mahadeva/papers/book-chapter.htm Rakefet All failed at the race. http://www.cnn.com/2004/TECH/ptech/03/14/darpa.race/index.html Rani Hi Morre, How easy it is for us to open a free email account at Yahoo! ? well, not as easy as it is for an automated program on other free mail sites. http://www.ynet.co.il/articles/0,7340,L-2889927,00.html This is a related link for the previous email (reverse Turing Test) http://www.captcha.net/ Rani MORRE: THIS IS AN EMAIL I RECEIVED RECENTLY CONTAINING A PROPOSAL FOR A SYMPOSIUM ON METACOGNITION IN COMPUTATION (AAAI – American Association for Artificial Intelligence). Dear friends and colleagues, We intend to propose a AAAI Symposium on Metacognition in Computation. As part of the proposal process, AAAI asks that we submit the names of individuals interested in the topic; you have been identified as someone likely to be in this category. If you would indeed be interested in seeing a symposium on the topic of Metacognition in Computation, please reply to this mail, and we will put your name on the list. Expressions of interest do not commit one to actually participating in the symposium. A draft of the symposium description is pasted below. Suggestions for improving the description, or for specific topics which you think should be included, are most welcome. Sincerely yours, Mike Anderson, University of Maryland, College Park Mike Cox, Wright State University Tim Oates, University of Maryland, Baltimore County Don Perlis, University of Maryland, College Park Proposal for Symposium on Metacognition in Computation The importance of metacognition in human thinking, learning, and problem solving is well established. Humans use metacognitive monitoring and control to choose goals, assess their own progress, and, if necessary, adopt new strategies for achieving those goals, or even abandon a goal entirely. Absent-minded Professor Doe, for instance, almost always forgets his lunch. He has an adequate recovery plan for this: he simply goes to the school cafeteria. However, as the school cafeteria is expensive, this strategy is wasteful. Thus, Professor Doe employs metacognitive reflection, realizes the frequency with which he forgets his lunch, and adopts a special strategy to help him remember: he sticks a note on his mirror. In a similar vein, students preparing for--or taking--an exam will make judgements about the relative difficulty of the material to be covered, and use this to choose study strategies, or which questions to answer first. Not surprisingly, in these cases, accuracy of metacognitive judgements correlates with academic performance. Thus, understanding human metacognition has been an important part of work on automated tutoring systems, and has led to the design of methods for using computer assistants to help improve human metacognition. Recently, however, there has been growing interest in trying to create, and investigate the potential benefits of, intelligent systems which are *themselves* metacognitive. It is thought that systems that monitor themselves, and proactively respond to problems, can perform better, for longer, with less need for (expensive) human intervention. Thus has IBM widely publicized their "autonomic computing" initiative, aimed at developing computers which are (in their words) self-aware, self-configuring, selfoptimizing, self-healing, self-protecting, and self-adapting. More ambitiously, it is hypothesized that metacognitive awareness may be one of the keys to developing truly intelligent artificial systems. DARPA's recent Cognitive Information Processing Technology initiative, for instance, foregrounds reflection (along with reaction and deliberation) as one of the three pillars required for flexible, robust AI systems. On the other side of the coin, it has also been established that metacognition can actually interfere with performance. That is, metacognition is no panacea, and therefore one of the issues which requires further inquiry is the scope and limits of its usefulness. Furthermore many researchers still argue over the most useful definition of metacognition. Is it just cognition about cognition, or is it more useful to include such things as monitoring the outcomes of one's own actions in the world? The proposed workshop is intended to bring together researchers from computer science, cognitive science, linguistics, psychology, philosophy, etc., interested in exploring, reporting on methods for, and evaluating the worth of, implementing metacognition in AI systems. Possible topics include: -Reports on implemented metacognitive systems -Computationally tractable models of human metacognition -Methods for evaluating metacognitive systems -Methods for implementing metacognition in heterogeneous systems -Evaluation of different architectures for implementing metacognition -Domains and/or problems for which metacognition is useful/essential -Formal and/or knowledge-representation issues in metacognition -The limits of metacognition (including cost/benefit analyses) eof ------------------------------------------------------------Michael L. Anderson Post-Doctoral Research Associate Institute for Advanced Computer Studies University of Maryland College Park, MD 20742 http://www.cs.umd.edu/~anderson CHECK OUT ANDERSON'S HOME PAGE – HE DOES VERY INTERESTING WORK, ON AI AND METACOGNITION, BUT MAINLY ON EMBODIED AI/COGNITION MORRE: Artificial Intelligence (Journal) - Most downloaded articles January - December 2003: http://www1.elsevier.com/homepage/sac/downloads/00043702.htt Hi Morre, here's a nice game I found online.. http://q.20q.net/q.cgi 20Q.net is an experiment in artificial intelligence. The program is very simple but its behavior is complex. Everything that it knows and all questions that it asks were entered by people playing this game. 20Q.net is a learning system; the more it is played, the smarter it gets. I played with it and it guessed right what I thought of after 30 questions. after revealing the answer it writes a summary where it written what questions I answered 'wrongly' according to its knowledge and what it doesn't know yet about the object. See you Tuesday, Merrav. http://www.globeandmail.com/servlet/ArticleNews/TPStory/LAC/20040117/ROBOT 17//? query=robot+scientist the robotic scientist. have we mentioned it before? if not, i could delve into it more to learn the details. TOMER ,מורי שלום יש כאן דוגמאות מעניינות למודלים רובוטיים המנסים. Biorobotics -האתר שלהלן עוסק ב .Embedded Intelligence -לחקות תופעות ביולוגיות ועשויות להוות כיוון ל .http://lib.haifa.ac.il/www/searchengines_fr.html ,בברכה .גיא Hi all, For Shirley's referat, here is a short summary of Searle's Chinese room argument and several of the most prominent replies (from the Internet Encyclopedia of Philosophy): http://www.iep.utm.edu/c/chineser.htm And here is Christopher Green's criticism of AI (as analogous to Disneyland) in the Psycoloquy e-journal archive. In the page that is displayed, the first link is to the target article itself, and the remaining links are to replies and replies to replies: http://www.cogsci.ecs.soton.ac.uk/cgi/psyc/ptopic?topic=Ai-cognitivescience&submit=View+Topic Morre