Theoretic analyses of internal model dynamics in the brain
Talkat Understanding Higher-Level Intelligence from AI, Psychology, and Neuroscience Perspectives · 2024-06-24..06-28·
summary
Argues that canonical internal world models are unnecessary and often suboptimal because time is the scarce resource for biological agents and chronologically simulating futures wastes it. Supports this with hippocampal replay analyses showing jumpy versus diffusive off‑line trajectories, mouse schizophrenia models where replay loses spatial structure, a human semantic foraging study using word2vec with 26 patients and 26 controls showing larger semantic jumps, and a computational runtime‑optimal gradient inference model that prioritizes bottlenecks and matches human route planning heuristics.
AI-written summary · about
chapters (9)
- 0:00Internal model dynamics intro
- 1:31Critique of internal model hypothesis
- 5:28Neural correlates of internal stimulation
- 10:55Hippocampal activity in schizophrenia
- 15:41Semantic foraging in schizophrenia
- 18:13Teleological simulation alternative
- 22:50Taxi drivers and planning
- 27:27Summary and open questions
- 29:16Q&A
related talks
01h 03mPlenary Discussion: What Is Missing in Current Theories of Brain Computation?
43mCan Cognitive Neuroscience Provide a Theory of Deep Learning...
53mGeneralization insights from actual cognition
01h 13mComputation in Networks of Neurons in the Brain II
53mPlenary Discussion: How Can One Test/Falsify Current Theories of Brain Computation?
47mDynamic Neural Network Structures Through Stochastic Rewiring
45mMachine Learning, Deep Networks and Interpretability in Systems, Cognitive and...
58mStrong Generalization from Small Brains and No Training Data