Self-Supervised Learning of Local Predictive Directions as a Neural Computation
Abstract
Natural sensory dynamics are globally nonlinear but often exhibit simple local geometry, with nearby trajectories evolving preferentially along a small number of predictive directions. We ask how neural circuits could discover this geometry directly from temporal experience. We introduce Local Predictive Directions (LPD), a self-supervised framework that learns local directions whose projections maximally predict near-future trajectory evolution. LPD applies past-future canonical correlation analysis to temporal differences, making the learned representation invariant to translations of the underlying state while avoiding estimation of equilibria or a global dynamical model. To represent systems with multiple predictive regimes, we introduce a non-negative similarity-matching objective, with offline and online algorithms, that partitions trajectory segments into predictive domains, each associated with a distinct local predictive direction. The online algorithm admits a natural neural-network interpretation: the resulting activities provide soft domain assignments, while synaptic weights encode the corresponding local predictive filters. On stochastic nonlinear dynamical systems, LPD recovers the local predictive geometry and partitions trajectories into distinct dynamical regimes. On naturalistic sensory streams, the online algorithm produces center-surround spatial receptive fields and ON/OFF sustained and transient temporal filters resembling those of early visual neurons. Our results suggest that predictive learning can provide a self-supervised mechanism by which neural circuits discover the local geometry of environmental dynamics and represent it in receptive fields.
Related
- Metabolite-centric Knowledge Graph Unveils Predictive and Mechanistic Insights through Self-supervised Multimodal Learning
- Predictive coding networks capture human neural representations missing in supervised DNNs
- Neural traveling waves in cortex: network mechanisms and potential roles in neural computation
- Neural mechanisms for resilient sensory computation under metabolic stress
- Bio-CPC: Collective Predictive Computation in Biological Organization
BibTeX
@misc{zenari2026selfsupervised,
title = {{Self-Supervised Learning of Local Predictive Directions as a Neural Computation}},
author = {Marco Zenari and Xuehao Ding and Daniel Burnham and Austin R. Ellis-Mohr and Anirvan M. Sengupta and Dmitri B. Chklovskii},
howpublished = {arXiv},
year = {2026},
month = nov,
doi = {10.48550/arxiv.2611.07809},
eprint = {2611.07809},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2611.07809},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Self-Supervised Learning of Local Predictive Directions as a Neural Computation
AU - Marco Zenari
AU - Xuehao Ding
AU - Daniel Burnham
AU - Austin R. Ellis-Mohr
AU - Anirvan M. Sengupta
AU - Dmitri B. Chklovskii
PY - 2026
DA - 2026/11/05
PB - arXiv
DO - 10.48550/arxiv.2611.07809
UR - https://arxiv.org/abs/2611.07809
AB - Natural sensory dynamics are globally nonlinear but often exhibit simple local geometry, with nearby trajectories evolving preferentially along a small number of predictive directions. We ask how neural circuits could discover this geometry directly from temporal experience. We introduce Local Predictive Directions (LPD), a self-supervised framework that learns local directions whose projections maximally predict near-future trajectory evolution. LPD applies past-future canonical correlation analysis to temporal differences, making the learned representation invariant to translations of the underlying state while avoiding estimation of equilibria or a global dynamical model. To represent systems with multiple predictive regimes, we introduce a non-negative similarity-matching objective, with offline and online algorithms, that partitions trajectory segments into predictive domains, each associated with a distinct local predictive direction. The online algorithm admits a natural neural-network interpretation: the resulting activities provide soft domain assignments, while synaptic weights encode the corresponding local predictive filters. On stochastic nonlinear dynamical systems, LPD recovers the local predictive geometry and partitions trajectories into distinct dynamical regimes. On naturalistic sensory streams, the online algorithm produces center-surround spatial receptive fields and ON/OFF sustained and transient temporal filters resembling those of early visual neurons. Our results suggest that predictive learning can provide a self-supervised mechanism by which neural circuits discover the local geometry of environmental dynamics and represent it in receptive fields.
ER -