CoChat

Socially situated artificial intelligence enables learning from human interaction

Open access
Krishna, Ranjay, Lee, Donsuk, Fei-Fei, Li, Bernstein, Michael S. · 2022 · Proceedings of the National Academy of Sciences
Cited by 63

The DOI matches its Crossref record; 19 of the 19 references it cites were found in a registry or an index.

Abstract

Regardless of how much data artificial intelligence agents have available, agents will inevitably encounter previously unseen situations in real-world deployments. Reacting to novel situations by acquiring new information from other people—socially situated learning—is a core faculty of human development. Unfortunately, socially situated learning remains an open challenge for artificial intelligence agents because they must learn how to interact with people to seek out the information that they lack. In this article, we formalize the task of socially situated artificial intelligence—agents that seek out new information through social interactions with people—as a reinforcement learning problem where the agent learns to identify meaningful and informative questions via rewards observed through social interaction. We manifest our framework as an interactive agent that learns how to ask natural language questions about photos as it broadens its visual intelligence on a large photo-sharing social network. Unlike active-learning methods, which implicitly assume that humans are oracles willing to answer any question, our agent adapts its behavior based on observed norms of which questions people are or are not interested to answer. Through an 8-mo deployment where our agent interacted with 236,000 social media users, our agent improved its performance at recognizing new visual information by 112%. A controlled field experiment confirmed that our agent outperformed an active-learning baseline by 25.6%. This work advances opportunities for continuously improving artificial intelligence (AI) agents that better respect norms in open social environments.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1073/pnas.2115730119

Cites

19 of 19 references verified0 citation errors · 0 not checkable

Statistics

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 19 references

Research works printed but not linked: 43

  1. M. L. Gray, S. Suri, Ghost Work: How to Stop Silicon Valley from Building a New Global Underclass (Eamon Dolan Books, 2019).
  2. 10.1145/3191513 · DOI · Paper page
  3. G. Marcus, E. Davis, Rebooting AI: Building Artificial Intelligence We can Trust (Pantheon, 2019).
  4. G. Marcus Deep learning: A critical appraisal. arXiv [Preprint] 13 August 2018.
  5. T. Winograd, Thinking Machines: Can There Be? Are We? (Department of Computer Science, Stanford University, 1987).
  6. 10.1037/11193-000 · DOI
  7. J. Newson, The growth of shared understandings between infant and caregiver. (Cambridge University Press, Cambridge, United Kingdom, 1979), pp. 207–222.
  8. 10.18653/v1/D17-1259 · DOI · Paper page
  9. 10.18653/v1/D16-1127 · DOI · Paper page
  10. A. L. Thomaz, C. Breazeal, Reinforcement Learning with Human Teachers: Evidence of Feedback and Guidance with Implications for Learning Performance (Aaai, Boston, MA, 2006), vol. 6, pp. 1000–1005.
  11. 10.1016/j.artint.2007.09.009 · DOI · Paper page
  12. M. Cakmak A. L. Thomaz “Designing robot learners that ask good questions” in Proceedings of the Seventh Annual ACM/IEEE International Conference on Human-Robot Interaction H. A. Yanco A. Steinfeld V. Evers O. Chadwicke Jenkins Eds. (ACM Boston MA 2012) pp. 17–24. · DOI · Paper page
  13. 10.1007/s10458-006-0005-z · DOI
  14. 10.1109/CVPR.2018.00009 · DOI
  15. C. L. Breazeal “Sociable machines: Expressive social exchange between humans and robots ” PhD thesis Massachusetts Institute of Technology Cambridge MA (2000).
  16. A. Athalye L. Engstrom A. Ilyas K. Kwok “Synthesizing robust adversarial examples” in International Conference on Machine Learning J. G. Dy A. Krause Eds. (PMLR Stockholm Sweden 2018) pp. 284–293.
  17. G. Neff, P. Nagy, Talking to bots: Symbiotic agency and the case of Tay. Int. J. Commun. 10, 17 (2016).
  18. 10.1109/TAMD.2010.2051030 · DOI · Paper page
  19. 10.1109/DEVLRN.2010.5578865 · DOI
  20. A. Guillory, J. A. Bilmes, “Simultaneous learning and covering with adversarial noise” in ICML (2011), vol. 11, pp. 369–376.
  21. Z. Zhao . “Recommending what video to watch next: A multitask ranking system” in Proceedings of the 13th ACM Conference on Recommender Systems (2019) pp. 43–51. · DOI · Paper page
  22. E. J. Webb D. T. Campbell R. D. Schwartz L. Sechrest Unobtrusive Measures (Sage Publications 1999) vol. 2 . · DOI · Paper page
  23. L. Von Ahn L. Dabbish “Labeling images with a computer game” in Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (ACM 2004) pp. 319–326. · DOI · Paper page
  24. R. B. Cialdini, Influence: The Psychology of Persuasion, vol. 3 (A. Michel Port Harcourt, 1987).
  25. M. Dontcheva R. R. Morris J. R. Brandt E. M. Gerber “Combining crowdsourcing and learning to improve engagement and performance” in Proceedings of the SIGCHI Conference on Human Factors in Computing Systems M. Jones P. A. Palanque A. Schmidt T. Grossman Eds. (ACM Toronto ON Canada 2014) pp. 3379–3388. · DOI
  26. P. G. Ipeirotis E. Gabrilovich “Quizz: Targeted crowdsourcing with a billion (potential) users” in Proceedings of the 23rd International Conference on World Wide Web C.-W. Chung A. Z. Broder K. Shim T. Suel Eds. (ACM Seoul Republic of Korea 2014) pp. 143–154. · DOI · Paper page
  27. N. Jaques . “Social influence as intrinsic motivation for multi-agent deep reinforcement learning” in International Conference on Machine Learning (PMLR) P. A. Ortega D. J. Strouse J. Z. Leibo N. de Freitas Eds. (Long Beach CA 2019) pp. 3040–3049.
  28. N. Rabinowitz . “Machine theory of mind” in International Conference on Machine Learning (PMLR) J. G. Dy A. Krause Eds. (Stockholm Sweden 2018) pp. 4215–4224.
  29. I. Serban A. Sordoni Y. Bengio A. Courville J. Pineau “Building end-to-end dialogue systems using generative hierarchical neural network models” in Proceedings of the AAAI Conference on Artificial Intelligence D. Schurrmans M. P. Wellman Eds. (AAAI Press Phoenix AZ 2016) vol. 30 pp. 3776–3784. · DOI · Paper page
  30. B. M. Hill, “Almost Wikipedia: Eight early encyclopedia projects and the mechanisms of collective action” in Essays on Volunteer Mobilization in Peer Production (Tech. Rep., Massachusetts Institute of Technology, Cambridge, MA, 2013), pp. 1–38.
  31. 10.3102/0013189X11427083 · DOI · Paper page
  32. R. E. Kraut, P. Resnick, Encouraging contribution to online communities in Building successful online communities: Evidence-based social design. (MIT Press, Cambridge, MA, 2011), pp. 21–76.
  33. 10.1007/978-3-322-84540-5_13 · DOI · Paper page
  34. M. R. Morris J. Teevan K. Panovich “What do people ask their social networks and why? A survey study of status message q&a behavior” in Proceedings of the SIGCHI Conference on Human Factors in Computing Systems E. D. Mynatt . Eds. (Atlanta GA 2010) pp. 1739–1748. · DOI · Paper page
  35. H. Narayanan, S. Mitter, “Sample complexity of testing the manifold hypothesis” in Advances in Neural Information Processing Systems, J. D. Lafferty, C. K. I. Williams, J. S. Taylor, R. S. Zemel, A. Culotta, Eds. (Vancouver, BC, Canada, 2010), pp. 1786–1794.
  36. G. K. Zipf, Human Behavior and the Principle of Least Effort (Cambridge University Press, 1949).
  37. 10.4324/9780203781036 · DOI · Paper page
  38. 10.1016/j.tics.2004.02.007 · DOI · Paper page
  39. 10.1017/CBO9780511807572 · DOI · Paper page
  40. 10.1109/CVPR.2019.00211 · DOI · Paper page
  41. J. Li, A. H. Miller, S. Chopra, M. Ranzato, J. Weston, “Learning through dialogue interactions by asking questions” in International Conference on Learning and Representation (Toulon, France, 2017).
  42. T. Shen A. Kar S. Fidler “Learning to caption images through a lifetime by asking questions” in Proceedings of the IEEE International Conference on Computer Vision (Seoul Republic of Korea 2019) pp. 10393–10401. · DOI
  43. J. Schulman F. Wolski P. Dhariwal A. Radford O. Klimov Proximal policy optimization algorithms. arXiv [Preprint] 13 August 2018.

Cited by

See all 63

Metadata

Authors
Krishna, R., Lee, D., Fei-Fei, L., & Bernstein, M. S.
Venue
Proceedings of the National Academy of Sciences
Published
2022-09-19
Volume
119
Issue
39
Pages
e2115730119-e2115730119
ISSN
  • 0027-8424 (ISSN-L)
  • 1091-6490
OpenAlex
W4296343374
Cited by
63
References
19
FWCI
4.44
Type
journal article
Language
en
Open access
hybrid
Funders
  • Brown Institute of Media Innovation (NA)
  • Toyota Research Institute (NA)
  • Snapchat Research (NA)
  • National Science Foundation (1900638)

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Socially situated artificial intelligence enables learning from human interaction. https://app.cochat.ai/papers/10.1073/pnas.2115730119
The paper
Krishna, R., Lee, D., Fei-Fei, L., & Bernstein, M. S. (2022). Socially situated artificial intelligence enables learning from human interaction. Proceedings of the National Academy of Sciences. https://doi.org/10.1073/pnas.2115730119

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.