#132 Bayesian Cognition and the Future of Human-AI Interaction, with Tom Griffiths

Proudly sponsored by PyMC Labs, the Bayesian Consultancy. Book a call, or get in touch!

Check out Hugo’s latest episode with Fei-Fei Li, on How Human-Centered AI Actually Gets Built

Our theme music is « Good Bayesian », by Baba Brinkman (feat MC Lars and Mega Ran). Check out his awesome work!

Visit our Patreon page to unlock exclusive Bayesian swag ;)

Takeaways:

Computational cognitive science seeks to understand intelligence mathematically.
Bayesian statistics is crucial for understanding human cognition.
Inductive biases help explain how humans learn from limited data.
Eliciting prior distributions can reveal implicit beliefs.
The wisdom of individuals can provide richer insights than averaging group responses.
Generative AI can mimic human cognitive processes.
Human intelligence is shaped by constraints of data, computation, and communication.
AI systems operate under different constraints than human cognition. Human intelligence differs fundamentally from machine intelligence.
Generative AI can complement and enhance human learning.
AI systems currently lack intrinsic human compatibility.
Language training in AI helps align its understanding with human perspectives.
Reinforcement learning from human feedback can lead to misalignment of AI goals.
Representational alignment can improve AI's understanding of human concepts.
AI can help humans make better decisions by providing relevant information.
Research should focus on solving problems rather than just methods.

Chapters:

00:00 Understanding Computational Cognitive Science

13:52 Bayesian Models and Human Cognition

29:50 Eliciting Implicit Prior Distributions

38:07 The Relationship Between Human and AI Intelligence

45:15 Aligning Human and Machine Preferences

50:26 Innovations in AI and Human Interaction

55:35 Resource Rationality in Decision Making

01:00:07 Language Learning in AI Models

Fler avsnitt av Learning Bayesian Statistics