Architected Interpretability: Structured Bottlenecks Make Brain-Dynamics Models Causally Explainable
PlisSergey · x · 2026-10-28
Sergey Plis 预告 11 月 10 日在美国数学与统计创新研究所（IMSI）的演讲，主题为「瓶颈中的信息：面向脑动态的架构化可解释性（ArchInterp）」。
深度学习虽能仅凭神经影像预测诊断状态，但准确预测本身并不解释背后的脑过程。演讲将介绍一种补充性思路：在编码器与解码器之间加入结构化瓶颈，强制预测所用的信息经过一个可检查、可约束的表示。在功能神经影像工作中，该瓶颈是表示脑区间交互的图，使交互成为模型计算的显式部分。
报告还将讨论：图边的含义及在何种假设下支持因果解释，以及这一路径如何推动「用预测模型来发展与检验脑动态—行为—障碍关系的解释」。
Related event: Sergey Plis to Speak on Architected Interpretability for Brain Dynamics(2 posts)→
More from Research
- Can AI search automatically discover cross-field math connections like Fermat's proof? — IgorCarron · 2026-10-28
- New bioRxiv paper reframes enzyme function prediction as conditional generation — KevinKaichuang · 2026-10-28
- New generative model predicts enzyme reaction mechanisms straight from sequence — KevinKaichuang · 2026-10-28
- GenAI is creeping into bestselling novels, study featured in Nature — TuhinChakr · 2026-10-28
- Nathan Lambert: open models' demise exaggerated, but AA index flattered them — natolambert · 2026-10-28
- Local NCM benchmarks: 4x faster for one request, 4-8x slower at 64 concurrent — Prompt Engineering · 2026-10-28