arXiv AI
Spectral Alignment in Forward-Backward Representations via Temporal Abstraction
Read the original on arXiv AI →The Flow has not summarised this story yet — read it at arXiv AI.
Related stories
arXiv Machine Learning
Jul 16
Factorized Spectral Representations for Reinforcement Learning
arXiv:2607. 13498v1 Announce Type: new Abstract: Learning a compact model of the world from interaction data is central to sample-efficient deep reinforcement learning.
arXiv AI
Jul 9
Can We Really Learn One Representation to Optimize All Rewards?
arXiv:2602. 11399v2 Announce Type: replace-cross Abstract: As unsupervised pretraining becomes increasingly ubiquitous in reinforcement learning, a more thorough theoretical understanding of these methods becomes of equal importance to their empirical success.
arXiv Machine Learning
Aug 31
Shift Before You Learn: Enabling Low-Rank Representations in Reinforcement Learning
The paper challenges the common assumption that the successor measure in reinforcement learning is approximately low-rank, showing instead that a low-rank structure emerges in a shifted successor measure that ignores initial transitions. It provides finite-sample guarantees for estimating this low-rank approximation, introduces Type II Poincaré inequalities to bound spectral recoverability, and links the necessary shift to the decay of high-order singular values and local mixing properties. Experiments confirm that shifting the successor measure improves goal-conditioned RL performance.
arXiv AI
Jul 20
The Terminal Representation in Reinforcement Learning
arXiv:2605. 31289v2 Announce Type: replace-cross Abstract: Representation learning is a powerful tool for spatio-temporal abstraction within reinforcement learning (RL).
arXiv Machine Learning
Aug 24
SPARCL: Spectral Partitioned Analytic Continual Learning
SPARCL introduces a spectral partitioned analytic continual learning method that addresses forgetting in analytic class‑incremental learning. By decomposing the running autocorrelation into a high‑energy core and a residual complement, SPARCL freezes core components for old classes and updates only the residual block, ensuring closed‑form updates with an invariance guarantee. Experiments on CIFAR‑100, CUB‑200, ImageNet‑R, and ImageNet‑A with a frozen ViT‑B/16 protocol show that SPARCL narrows the performance gap between classical analytic learners and strong representation matchers while complementing sparse feature‑decorrelation approaches.
arXiv Computer Vision
Sep 1
See the Change, Keep the Flow: Unsupervised Action Segmentation via Spectral-Temporal Representation Learning
arXiv:2608.29611v1 Announce Type: new Abstract: Unsupervised action segmentation aims to discover latent action categories and their temporal organization without action annotations. Optimal transpor...