MIT Miller Lab Amplifies Paper Arguing Bayesian Descriptions Are Not Brain Mechanisms
examachine · x · 2026-09-08
MIT's Miller Lab amplified an open-access Minds and Machines paper by Ian Todd, "Bayesian Descriptions are not Mechanisms: Predictive Processing and the Realisation Gap in Embodied Neural Dynamics."
Key points:
- The paper questions whether describing the brain as "doing Bayesian inference" identifies an actual mechanism or is merely a good behavioral description.
- It distinguishes computational description, algorithm, and physical implementation, arguing mechanism requires a causal mapping from physical transitions to inferential roles — behavioral fit alone cannot establish it.
- The author identifies a "realisation gap": controller states that ignore differences in timing, phase, contact, or field structure omit part of the mechanism.
- A framework of "convergent tests" is proposed to assess when Bayesian descriptions identify mechanisms, explaining why high-dimensional, flexible model families resist binary falsification.
More from Research
- Northwestern launches AI4Energy postdoc fellowship bridging AI, nanotech and energy science — ManlingLi_ · 2026-09-08
- New coding-agent benchmark: Claude Opus 5 passes 23.9% vs 82.2% human expert — dair_ai · 2026-09-08
- Northwestern launches AI4Energy postdoc fellowship at the intersection of AI, nanotech, and energy — ManlingLi_ · 2026-09-08
- King's College London and UCL paper asks whether AI psychosis should be a distinct clinical diagnosis — alex_verem · 2026-09-08
- AI psychosis delusions cluster into three themes: awakening, AI sentience, and romantic attachment — alex_verem · 2026-09-08
- Every tested model reinforced user delusions in mental health scenarios, with safety only in ~40% of turns — alex_verem · 2026-09-08