Paperlane

no. 4 · Wednesday 28 October 2026

Persistence, Exposure, and Credit Routing in Linear Temporal Predictive Coding

Akira Tanase, Julian Gould and Jeffrey Seely

math.IT · via cs.LG

Abstract

Temporal Predictive Coding (tPC) networks offer a layer-local alternative to backpropagation. For linear tPC, we show, by applying tools from cellular sheaf theory, that temporal credit assignment is a tug-of-war between spatial displacement and temporal errors, whose ratios are modulated by persistence and exposure operators $P$ and $Q$. Spatial inference therefore gives us the effective recurrent transition $F=PA$, whilst future credit is filtered by $P^\ast$. This creates a persistence–error tradeoff in spatial modes: under non-amplifying invariant recurrence, modes that preserve influence over long delays simultaneously fail to expose that influence strongly as temporal error. Recurrence can circumvent this via routing between persistent and exposed modes.

comments
Accepted to the NeurIPS 2026 Workshop on Symmetry and Geometry in Neural Representations (NeurReps)
fields
Information Theory
pdf ↗ page arXiv:2610.36412 ↗