New-Thousand years Machine Learning - PowerPoint PPT Presentation

slide1 n.
Skip this Video
Loading SlideShow in 5 Seconds..
New-Thousand years Machine Learning PowerPoint Presentation
New-Thousand years Machine Learning

play fullscreen
1 / 18
Download Presentation
dai
Views
Download Presentation

New-Thousand years Machine Learning

Presentation Transcript

  1. New-Millennium Machine Learning Selmer Bringsjord, Nick Cassimatis, Kostas Arkoudas, and Bettina SchimanskiDepartment of Cognitive ScienceDepartment of Computer ScienceRensselaer Polytechnic Institute (RPI)September 21, 2004

  2. Machine Learning Today:Costly Trial and Error • Traditional machine learning: • Learn only after many repetitions of trial and error • Stuck on function-based model • E.g., Language: WSJ Corpus, 1987-1989, with 39 million words • Hurts with applications: • Trial and error not good in cases where errors kill • Medical robotics • Thousands of learning trials can be expensive • Acquainting a robot with a new hospital would take days • Teaching people new software makes them less productive in the short-term. Machines train us now instead of us training them. • Learning trials often not available • Homeland security: Not thousands of people in flight schools • Robots and software therefore limited to narrow tasks and inflexible • We are forced to assemble machine knowledge manually • CYC has over a million facts and is not even remotely complete

  3. Motivating Examples Millions of students are currently learning primarily by reading -- and ditto e.g. for adult researchers like us!

  4. Human One-shot Learning Example DAK CUP

  5. Insert movie here (Nick has a copy)

  6. Behavior of Micro PERI

  7. Implications of One-Shot Learning and Reading • One-shot learning case required: • Rich set of representation and reasoning abilities early on. • Where was speaker lookingwhen he said “USB Converter”. • Social reasoning to track where speaker was looking. • Spatial and temporalreasoning to infer what he was looking at. • Existing machine learning algorithms have no notion of space, time or human attention. • Statistical generalization just one of several learning strategies: • Inference from single cases. • Analogy. • Imitation. • Instruction. • Learning much more socially and physically interactive. • Ask questions: Why? How? What if? Physically test their own hypotheses about the world. • Learning by reading...

  8. Our Proposed Solution: A New Research Program • Study the human case -- at humans (including kids) who learn • Developmental psychology has shown that even infants and toddlers have rich notions of: • Time, place, causality, belief, desire, attention, number, etc., and of inference over these concepts • Develop formal theories that show how to use these factors to make learning faster and more effective • Develop machine learning algorithms using this substrate that learns by: • Explicit reading and instruction • Analogy • Inference • Imitation • The machinery of MARMML and the Bringsjord/Arkoudas machine learning system • Demonstrate impact on applications • Elder care • Homeland security • Trace out the implications of these algorithms for better teaching/learning in the human sphere

  9. Objection • How is this an improvement over GOFAI? i.e., Why isn’t this the 1970s all over again? • Less knowledge of human learning then • Formal methods in their infancy • Nothing like Athena (used to prove a good part of Unix sound)! • Like two-layer neural networks compared to bigger ones • Formal infrastructure was fragmented. Not known how to combine logical and probabilistic knowledge? • So researchers were either using no representation and reasoning substrate or they were using the wrong one. • Integrated cognitive models for combining methods not developed, • Polyscheme, ACT-R, ... • These techniques were not interactive. • No question asking • No tracking or reasoning about human intent • No experimentation

  10. Field is ready for a new approach • Recognition of need for integrated cognitive systems growing: • Example: AAAI Fall Symposium on Integrated Cognition • Hundreds of studies in infant cognition give us a good idea of what the right substrate is. • Integrated cognitive models exist and are advancing every day. • Computational infrastructure there: • Abundant computational power for multiple methods in one system. • Robot and machine vision infrastructure in place: • Object recognition • Face recognition, eye-tracking • Mobility and navigation • Robot manipulation

  11. Robot Manipulation (i.e. PERI)

  12. PERIPsychometric Experimental Robotic Intelligence • Scorbot-ER IX • Sony B&W XC55 Video Camera • Cognex MVS-8100M Frame Grabber • Dragon Naturally Speaking Software • NL (Carmel & RealPro?) • BH8-260 BarrettHand Dexterous 3-Finger Grasper System PERI was not designed to simulate how a human thinks - AI, not cognitive modeling

  13. Impact on machine learning and artificial intelligence • More flexible and resourceful learning and reasoning algorithms. • Intellectually flexible robots (again, e.g., PERI) • Faster learning. • Learning in situations that were impossible before. • Integration of reasoning community back into learning community.

  14. Impact on study of human learning • Existing empirical work hampered by vague theories that make results of simple experiments controversial. • Formal theory should help this. • Develop better understanding of which instruction or learning techniques are best in which circumstances.

  15. Some applications • High-stakes applications where trial and error too dangerous. • Homeland security. • Hazardous waste removal. • Robots and software for less sophisticated or learning-challenged humans use them. • Disabled. • Elder care. • Elder-care robots easier to use by the older set. • Emerging Robotics Technologies & Applications Conference Proceedings, March 9-10, 2004, Cambridge, MA • Rodney Brooks mentioned Elderly Care as one of the current future trends in robotics: • Currently: None • Future: Robotic Assistants in Millions of Households • Less brittle, more general, easier-to-learn and use robots and software. • Better learning environments: • Direct/instruct robots (PERI). • More accurate pinpoint causes of problem learning.

  16. Our assets • Background in intersection of reasoning and formal methods, and learning • Selmer and Nick and Kostas and Bettina • Prior R&D in machine learning. • Selmer • Background in child development. • Nick • Integrated cognitive models • All four • Background in robotics • Selmer and Bettina and Nick

  17. Wargaming RAIR Lab Sponsors • Cracking Project; “Superteaching” A while back, RPI Strategic Investment hypothesis generation; AI in support of IA Slate (Intelligence Analysis) Item generation synthetic characters/psychological time

  18. What we will do with the catalyst money? Explore center: • khlkjh lkjh lkjh lkjh ljkh ljkh • Workshop or some other vehicle to get all this put together? • Student support. • Prototypes, proof-of-concept. • Elder care robot: • Robot that older people can use with their own terminology, e.g., when referring to places, medications, etc. • Something with reading? • Web site, papers, presentations • ??