1 / 25

Modeling Progress in AI

Modeling Progress in AI. Miles Brundage Consortium for Science, Policy, and Outcomes Arizona State University. Overview. Motivations Conceptual challenges/approaches Algorithmic vs. hardware/data/etc. progress Different cognitive domains and/vs. general AI Applications Go Atari

Télécharger la présentation

Modeling Progress in AI

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Modeling Progress in AI Miles Brundage Consortium for Science, Policy, and Outcomes Arizona State University

  2. Overview • Motivations • Conceptual challenges/approaches • Algorithmic vs. hardware/data/etc. progress • Different cognitive domains and/vs. general AI • Applications • Go • Atari • Future of work • Future directions Slide 2 of 25

  3. Motivations • The future of AI matters… • Why not model it rigorously? As done with… Slide 3 of 25

  4. Motivations (cont’d) • Concepts and existence proofs from other domains • Methodological clues • Model-based (Armstrong et al., 2014), • Quantitative (Mullins, 2012), • Short-term (Ibid) • Simple extrapolations often beat expert intuitions (Roper et al., 2011) Slide 4 of 25

  5. Relevant Literatures • AI evaluation • Natural intelligence evaluation • Technology forecasting • Technology roadmapping Eryilmaz et al. 2015 Slide 5 of 25

  6. Limitations • Not integrated • Focused on one agent (vs. overall AI community) • Static (vs. dynamic) Slide 6 of 25

  7. Challenges • “Progress in AI is linear/exponential” • Y-axis? • Unit/level(s) of analysis? • Metric(s)? • Confounding variables Clark, 2015 Bowling et al., 2015 Ginesh, 2011 Slide 7 of 25

  8. Proposed Framework • AI performance benefits from algorithmic progress, greater compute power, more/better data, and human input (including research, programming, feedback, etc.) • Vs… • Algorithmic AI progress is an increase in the ability to efficiently convert computing power to performance in a given cognitive domain or in a range of domains, controlling for human input and data Slide 8 of 25

  9. Proposed Framework (cont’d) • Rigorous definitions of intelligence, tasks, definitions, abilities, difficulty, etc. needed • Many foundations developed in Legg and Hutter, 2007; Goertzel, 2009; Hernandez-Orallo, 2016 • Need empirical study of different dimensions of AI performance increase Slide 9 of 25

  10. Role of Hardware • Speed and/vs. level (Carroll, 1993) • Ctotal = Cquality + Cspeedup • Algorithmic progress and hardware progress enable Pareto optimal improvements • Multiple reference levels He et al., 2015 with annotations Slide 10 of 25

  11. Role of Hardware (cont’d) • Research pace speedup • ~1/2 of historical performance improvement in six domains (Grace 2013) • Multiple reference levels (say, 2) which increase over time Slide 11 of 25

  12. Abilities • Internally correlated, externally distinct task classes • General intelligence is superset • Within-ability proficiency: • “The ability of an individual subject to perform a specified kind of task is the difficulty E at which the probability is 1/2 that he will do that task” • (Thurstone, 1937) Slide 12 of 25

  13. Preliminary Synthesis • Perception • Manipulation • Language processing • Learning • Planning • Social cognition • Inference • Steps toward rigor • task difference measure (Hernandez-Orallo, 2016) • Empirical analysis Slide 13 of 25

  14. (Deep) Learning Hegemony? • How to model interdependencies and synergies? • Perception • Manipulation • Language processing • Learning • Planning • Inference • Social cognition Slide 14 of 25

  15. Go • “Large, sudden jump in strength” • Jon Diamond, President of British Go Association, quoted by Jack Clark of Bloomberg • It was an improvement, but a smaller one than it was made out to be by some, after controlling for: • Hardware • Effort • Data Silver et al., 2016 Silver et al., 2016 Slide 15 of 25

  16. Go (cont’d) • Expert judgment vs. simple extrapolations • Forecast by Yamashita in 2011: cross-over in 4 years • Compare to Facebook’s darkfmcts3 and Zen19X Silver et al., 2016 data, extrapolated to 5 mins/turn or more hardware Slide 16 of 25

  17. Atari • Less human input for same performance or increased performance for same input = progress • DQN uses raw pixels • No hyperparameter adjustment across games • Exponentially improving performance, faster training Slide 17 of 25

  18. Atari (cont’d) • Human-scaled performance in 6 DeepMind papers, 2015-2016 Mnih et al., 2016 data • 2015 = approximate crossover point for human level as well as classical planning performance Slide 18 of 25

  19. Atari (cont’d) Steady trend for >5 years • Some games still subhuman, but decreasing in # Mnih et al., 2015, with annotations Data from several publications, details available upon request Slide 19 of 25

  20. AI Progress and Jobs:Current Theories • Murnane/Levy 2004: • Routine vs. non-routine • Autor2013: • Novelty of tasks • Brynjolfsson/McAfee 2014: • Creativity • Rus 2015: • Perception/manipulation (of some types) • Abstraction, creativity • Frey and Osborne 2013: • Social intelligence • Creative intelligence • Perception and manipulation Slide 20 of 25

  21. Issues with Current Theories • Performance and algorithmic progress not disentangled • Insufficient attention to: • robustness considerations • Interrelationships between abilities • generality Slide 21 of 25

  22. Example: Frey and Osborne, 2013 • Reasons for skepticism re: resilience of perception/manipulation, social intelligence, and creative intelligence in face of: • New datasets • Disproportionate effort • Hardware • Deep RL progress calls into question independence of abilities

  23. Future Directions • Principled demarcations of abilities and possible interrelationships • More epirical analysis of progress rates • Monte Carlo simulations of future ability levels Slide 23 of 25

  24. Acknowledgments • Thanks to David Guston, Joanna Bryson, Erik Fisher, Mark Gubrud, Daniel Dewey, Katja Grace, Kaj Sotala, Brad Knox, Vincent Mueller, Beau Cronin, Adi Wiezel, David Dalrymple, Adam Elkus, Jose Hernandez-Orallo, an anonymous reviewer, and others for helpful comments on this and related work. Slide 24 of 25

  25. Thanks! • www.milesbrundage.com • miles.brundage@asu.edu Slide 25 of 25

More Related