Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

The Future of AI: Going Beyond Deep Learning, Watson, and the Semantic Web

361 views

Published on

These slides, based on a presentation at distinguished lecture at IBM Almaden in March, 2017 explore some of the challenges to machine learning and some recent work. It is a newer version of the slides originally presented at IJCAI 2016.

Published in: Technology
  • Be the first to comment

  • Be the first to like this

The Future of AI: Going Beyond Deep Learning, Watson, and the Semantic Web

  1. 1. Tetherless World Constellation, RPI The Future of AI: Going Beyond Deep Learning, Watson, and the Semantic Web Jim Hendler Tetherless World Professor of Computer, Web and Cognitive Sciences Director, Institute for Data Exploration and Applications Rensselaer Polytechnic Institute http://www.cs.rpi.edu/~hendler @jahendler (twitter) Major talks at: http://www.slideshare.net/jahendler
  2. 2. Tetherless World Constellation, RPI Knowledge and Learning Knowledge representation in the age of Deep Learning, Watson, and the Semantic Web Jim Hendler Tetherless World Professor of Computer, Web and Cognitive Sciences Director, Institute for Data Exploration and Applications Rensselaer Polytechnic Institute http://www.cs.rpi.edu/~hendler Major talks at: http://www.slideshare.net/jahendler
  3. 3. Tetherless World Constellation, RPI Talk derives in part from a recent book
  4. 4. Tetherless World Constellation, RPI New Journal: Data Intelligence Knowledge graph is one of the topics we are interested in, please consider submitting a paper! (handout in your conference bag)
  5. 5. Tetherless World Constellation, RPI What has happened? • Several important AI technologies have moved through “knees in the curve” bringing much of the attention to AI again –Deep Learning (eg AlphaGo, vision processing) –Associative learning (eg Watson) –Semantic Web (eg search and schema.org)
  6. 6. Tetherless World Constellation, RPI A) Deep Learning “phase transition” in capabilities of neural networks w/machine power
  7. 7. Tetherless World Constellation, RPI Trained on lots of categorized images Imagenet: Duck Imagenet: Cat
  8. 8. Tetherless World Constellation, RPI Impressive results Increasingly powerful techniques have yielded incredible results in the past few years
  9. 9. Tetherless World Constellation, RPI B) Associative knowledge (text mining/QA) © IBM, used with permission.
  10. 10. Tetherless World Constellation, RPI Impressive Results Watson showed the power of “associative knowledge”
  11. 11. Tetherless World Constellation, RPI C) Semantic Web
  12. 12. Tetherless World Constellation, RPI “Knowledge graphs” mined from extracted and learned data
  13. 13. Tetherless World Constellation, RPI Impressive Results Google finds embedded metadata on >30% of its crawl – Guha, 2015
  14. 14. Tetherless World Constellation, RPI Not just Google…
  15. 15. Tetherless World Constellation, RPI Summary: AI has done some way cool stuff • Deep Learning: neural learning from data with high quality, • Watson: Associative learning from data with high quality • Semantic Web/Knowledge Graph: Graph links formation from extraction, clustering and learning but, there are still problems…
  16. 16. Tetherless World Constellation, RPI Combining these technologies
  17. 17. Tetherless World Constellation, RPI still a long way to go
  18. 18. Tetherless World Constellation, RPI Many intermediate steps (P. Norvig, WWW 4/2016, w. permission)
  19. 19. Tetherless World Constellation, RPI Why did knowledge graph need “”Human Judgments”? Association ≠ Correctness P. Mika, 2014 w.permission Michelangelo Leonardo Raphael Donnatello
  20. 20. Tetherless World Constellation, RPI Knowledge Representation? • A knowledge representation (KR) is most fundamentally a surrogate, a substitute for the thing itself, used to enable an entity to determine consequences by thinking rather than acting, i.e., by reasoning about the world rather than taking action in it. • It is a set of ontological commitments, i.e., an answer to the question: In what terms should I think about the world? • It is a fragmentary theory of intelligent reasoning, expressed in terms of three components: (i) the representation's fundamental conception of intelligent reasoning; (ii) the set of inferences the representation sanctions; and (iii) the set of inferences it recommends. • It is a medium for pragmatically efficient computation, i.e., the computational environment in which thinking is accomplished. One contribution to this pragmatic efficiency is supplied by the guidance a representation provides for organizing information so as to facilitate making the recommended inferences. • It is a medium of human expression, i.e., a language in which we say things about the world. R. Davis, H. Shrobe, P. Szolovits (1993)
  21. 21. Tetherless World Constellation, RPI KR: Human Expression Cute kid story: first two words
  22. 22. Tetherless World Constellation, RPI Telling cats from ducks doesn’t need KR !
  23. 23. Tetherless World Constellation, RPI “Saying things about the world” does "If I was telling it to a kid, I'd probably say something like 'the cat has fur and four legs and goes meow, the duck is a bird and it swims and goes quack’. " How would you explain the difference between a duck and a cat to a child? Woof
  24. 24. Tetherless World Constellation, RPI Learning semantic inferencing Bassem Makni, 2018 Phd (now at IBM Yorktown)
  25. 25. Tetherless World Constellation, RPI In noisy data Bassem Makni, 2018 Phd (now at IBM Yorktown)
  26. 26. Tetherless World Constellation, RPI The challenge of “background knowledge” What is the relationship between this man and this woman?
  27. 27. Tetherless World Constellation, RPI AI systems coming along well… What is the relationship between this man and this woman? Deep learning produced Scene Graph w/relationships (Klawonn & Heims, 2018)
  28. 28. Tetherless World Constellation, RPI But the challenges remain What is the relationship between this man and this woman? Deep learning produced Scene Graph w/relationships (Klawonn, 2018) Seeing the bride adds significant information that cannot be easily learned w/o background knowledge
  29. 29. Tetherless World Constellation, RPI A major problem to deploying AI in key areas
  30. 30. Tetherless World Constellation, RPI Adding knowledge to scene graphs Matthew Klawonn, PhD, 2019
  31. 31. Tetherless World Constellation, RPI KR: Surrogate knowledge? Which could you sit in? What is most likely to bite what? Which one is most likely to become a computer scientist someday? …
  32. 32. Tetherless World Constellation, RPI “Surrogate” knowledge Which could you sit in? What is most likely to bite what? Which one is most likely to become a computer scientist someday? How would they go about doing it?
  33. 33. Tetherless World Constellation, RPI KR: Recommended vs. Possible inference Which one would you save if the house was on fire?
  34. 34. Tetherless World Constellation, RPI Ethical AI systems need certainty Which one would you save if the house was on fire? Would you use a robot baby-sitter without knowing which of the three possibilities it would choose?
  35. 35. Tetherless World Constellation, RPI Human-Aware AI • Context is key – AI learning systems still perform best in well-defined contexts (or trained situations, or where their document set is complete, etc.) – Humans are good at recognizing context and deciding when extraneous factors don’t make sense • Or add extra “inferencing” (the bride example)
  36. 36. Tetherless World Constellation, RPI The challenge • If we want to implement KR systems on top of neural and associative learners we have an issue – The numbers coming out of Deep Learning and Associative graphs are not probabilities – They don’t necessarily ground in human-meaningful symbols • ”sub-symbolic” learning … • Association by clustering … • Errorful extraction …
  37. 37. Tetherless World Constellation, RPI The challenges • Can we avoid throwing out the reasoning baby with the grounding bathwater? – Long-term planning – Rules that need to be followed – Human Interaction • Even if computers don’t need to be symbolic communicators, WE DO!!! – Background knowledge (context) is symbolic
  38. 38. Tetherless World Constellation, RPI Human-AI interaction • Evidence that “centaurs” win – Human(s) and computer(s) teams currently beat either at chess (Go centaurs underway) – Anecdotal evidence that humans w/Watson top choices outperform Watson or human alone at Jeopardy – Medical study (diagnostic): • Doctor w/computer outperformed just doctor, just computer, two doctors
  39. 39. Tetherless World Constellation, RPI And this matters! “There was no rule about how long we were allowed to think before we reported a strike … but we knew that every second of procrastination took away valuable time, that the Soviet Union’s military and political leadership needed to be informed without delay. All I had to do was to reach for the phone; to raise the direct line to our top commanders — but I couldn’t move. I felt like I was sitting on a hot frying pan … when people start a war, they don’t start it with only five missiles …” We must all strive to be like Petrov and learn to trust the combination of AI training and human intuition. Stanislav Petrov: The man who saved the world

×