Itβs forgivable =) We just do the best we can with what we have (i.e., resource rational) π€£
31.07.2025 23:56 β π 2 π 0 π¬ 0 π 0@maxkw.bsky.social
professor at university of washington and founder at csm.ai. computational cognitive scientist. working on social and artificial intelligence and alignment. http://faculty.washington.edu/maxkw/
Itβs forgivable =) We just do the best we can with what we have (i.e., resource rational) π€£
31.07.2025 23:56 β π 2 π 0 π¬ 0 π 0Max giving a talk w the slide in OP
lol this may be the most cogsci cogsci slide I've ever seen, from @maxkw.bsky.social
"before I got married I had six theories about raising children, now I have six kids and no theories"......but here's another theory #cogsci2025
Quantifying the cooperative advantage shows why humans, the most sophisticated cooperators, also have the most sophisticated machinery for understanding the minds of others. It also offers principles for building more cooperative AI systems. Check out the full paper!
www.pnas.org/doi/10.1073/...
Finally, when we tested it against memory-1 strategies (such as TFT and WSLS) in the iterated prisoner's dilemma, the Bayesian Reciprocator: expanded the range where cooperation is possible and dominated prior algorithms using the *same* model across simultaneous & sequential games.
22.07.2025 06:03 β π 5 π 0 π¬ 1 π 0Even in one-shot games with observability, the Bayesian Reciprocator learns from observing others' interactions and enables cooperation through indirect reciprocity
22.07.2025 06:03 β π 6 π 0 π¬ 1 π 0In dyadic repeated interactions in the Game Generator, the Bayesian Reciprocator quickly learns to distinguish cooperators from cheaters, remains robust to errors, and achieves high population payoffs through sustained cooperation.
22.07.2025 06:03 β π 6 π 0 π¬ 2 π 0Instead of just testing on repeated prisoners' dilemma, we created a "Game Generator" which creates infinite cooperation challenges where no two interactions are alike. Many classic games, like the prisonerβs dilemma or resource allocation games, are just special cases.
22.07.2025 06:03 β π 8 π 0 π¬ 1 π 0It uses theory of mind to infer the latent utility functions of others through Bayesian inference and an abstract utility calculus to work across ANY game.
22.07.2025 06:03 β π 4 π 0 π¬ 1 π 0We introduce the "Bayesian Reciprocator," an agent that cooperates with others proportional to its belief that others share its utility function.
22.07.2025 06:03 β π 6 π 0 π¬ 1 π 0Classic models of cooperation like tit-for-tat are simple but brittle. They only work in specific games, can't handle noise and stochasticity and don't understand others' intentions. But human cooperation is remarkably flexible and robust. How and why?
22.07.2025 06:03 β π 6 π 0 π¬ 1 π 0This project was first presented back in 2018 (!) and was born from a collaboration between Alejandro Vientos, Dave Rand @dgrand.bsky.social & Josh Tenenbaum @joshtenenbaum.bsky.social
22.07.2025 06:03 β π 7 π 0 π¬ 1 π 0Our new paper is out in PNAS: "Evolving general cooperation with a Bayesian theory of mind"!
Humans are the ultimate cooperators. We coordinate on a scale and scope no other species (nor AI) can match. What makes this possible? π§΅
www.pnas.org/doi/10.1073/...
As always, CogSci has a fantastic lineup of workshops this year. An embarrassment of riches!
Still deciding which to pick? If you are interested in building computational models of social cognition, I hope you consider joining @maxkw.bsky.social, @dae.bsky.social, and me for a crash course on memo!
Very excited for this workshop!
17.07.2025 04:42 β π 14 π 2 π¬ 0 π 0Promotional image for a #CogSci2025 workshop titled βBuilding computational models of social cognition in memo.β Organized and presented by Kartik Chandra, Sean Dae Houlihan, and Max Kleiman-Weiner. Scheduled for July 30 at 8:30 AM in room Pacifica I. The banner features the conference theme βTheories of the Past / Theories of the Future,β and the dates: July 30βAugust 2 in San Francisco.
#Workshop at #CogSci2025
Building computational models of social cognition in memo
ποΈ Wednesday, July 30
π Pacifica I - 8:30-10:00
π£οΈ Kartik Chandra, Sean Dae Houlihan, and Max Kleiman-Weiner
π§βπ» underline.io/events/489/s...
'Cross-environment Cooperation Enables Zero-shot Multi-agent Coordination'
@kjha02.bsky.social Β· Wilka Carvalho Β· Yancheng Liang Β· Simon Du Β·
@maxkw.bsky.social Β· @natashajaques.bsky.social
doi.org/10.48550/arX...
(3/20)
AI DOOM
Settling in for my flight and apparently A.I. DOOM is now a movie genre between Harry Potter and Classics. Nothing better than an existential crisis with pretzels and a ginger ale.
29.06.2025 22:52 β π 6 π 0 π¬ 0 π 0Thanks to the Diverse Intelligence Community for all these inspiring days & impressions in Sydney ππ» @chriskrupenye.bsky.social @katelaskowski.bsky.social @divintelligence.bsky.social @maxkw.bsky.social
28.06.2025 03:46 β π 16 π 3 π¬ 0 π 0And a more detailed thread from the lead authors
Tianyi (Alex) Qiu and Zhonghao He, who both did an incredible job with this work: x.com/Tianyi_Alex_...
Check out the paper here:
arxiv.org/abs/2506.06166
LLMs learn beliefs and values from human data, influence our opinions, and then reabsorb those influenced beliefs, feeding them back to users again and again. We call this the "Lock-In Hypothesis" and develop theory, simulations, and empirics to test it in our latest ICML paper!
09.06.2025 20:23 β π 30 π 6 π¬ 1 π 0Congrats Fred! Awesome news!
08.05.2025 07:27 β π 2 π 0 π¬ 0 π 0Excited to speak about some new work on Bayesian Cooperation at this workshop! Join us virtually
28.04.2025 21:14 β π 10 π 3 π¬ 0 π 0Now out in JPSP βΌοΈ
"Inference from social evaluation" with Zach Davis, Kelsey Allen, @maxkw.bsky.social, and @julianje.bsky.social
π (paper): psycnet.apa.org/record/2026-...
π (preprint): osf.io/preprints/ps...
Our new paper (first one of my PhD!) on cooperative AI reveals a surprising insight: Environment Diversity > Partner Diversity.
Agents trained in self-play across many environments learn cooperative norms that transfer to humans on novel tasks.
shorturl.at/fqsNN%F0%9F%...
Awesome new work from my lab led by @kjha02.bsky.social scaling cooperative AI! True cooperation requires adapting to both unfamiliar partners and novel environments. Agents trained with CEC get us closer to agents that can act with general cooperative principles rather than memorized strategies.
19.04.2025 06:24 β π 3 π 0 π¬ 0 π 0How AlphaGo like architectures can explain human insight. Out now in Cognition!
14.03.2025 15:26 β π 10 π 1 π¬ 0 π 0my paper with max, @maxkw.bsky.social, tuomas, and @fierycushman.bsky.social out in cognition at long last www.sciencedirect.com/science/arti...
We explain why humans and successful AI planners both fail on a certain kind of problem that we might describe as requiring insight or creativity
Accepted as a Spotlight in ICLR2025!
13.02.2025 23:43 β π 10 π 0 π¬ 0 π 0Emergent transition from code to natural language for reasoning tasks when RL tuning a language model for math. Interesting to consider implications for "Language of Thought" style theories in cognition.
hkust-nlp.notion.site/simplerl-rea...