← Back to all news
arXiv Computation and Language By Jiwan Chung, Heechan Choi, Seon Joo Kim

Rethinking State Tracking in Recurrent Models Through Error Control Dynamics

Read the original on arXiv Computation and Language →

The paper argues that error control is as crucial as expressive capacity for state tracking in recurrent models. It proves that affine recurrent networks, including State‑Space Models and Linear Attention, cannot correct errors along state‑separating subspaces once they preserve state representations, leading to finite‑horizon solutions governed by accumulated error. Empirical results on group state‑tracking tasks show that tracking collapses when the distinguishability ratio falls below a readability threshold, and this crossing predicts downstream accuracy failure.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Computation and Language.

Related stories

arXiv Machine Learning
Aug 17

The Expressive Limits of Diagonal SSMs for State-Tracking

arXiv:2603. 01959v2 Announce Type: replace Abstract: State-Space Models (SSMs) have recently been shown to achieve strong empirical performance on a variety of long-range sequence modeling tasks while remaining efficient and highly-parallelizable.

By Mehran Shakerinava, Behnoush Khavari, Siamak Ravanbakhsh, Sarath Chandar
More like this →
arXiv Machine Learning
Jun 26

Learning State-Tracking from Code Using Linear RNNs

arXiv:2602. 14814v3 Announce Type: replace Abstract: Over the last years, state-tracking tasks, particularly permutation composition, have become a testbed to understand the limits of sequence models architectures like Transformers and RNNs (linear and non-linear).

By Julien Siems, Riccardo Grazzi, Korbinian P\"oppel, Kirill Kalinin, Hitesh Ballani, Babak Rahmani
llms
More like this →
arXiv AI
Oct 2

Measuring the Stability Assumption Behind Action Chunking

The paper investigates how small action errors evolve when using action chunking in behavioural cloning. By injecting errors at each state and observing their growth under open‑loop (no replanning) and closed‑loop (replanning) regimes, the authors classify states as contracting, expanding, or unresolved. Across twelve manipulation tasks, they find that stable states are rare, error amplification is common, and that short‑horizon fitting can overestimate long‑horizon propagation. Predictors trained on camera and proprioceptive data can recover open‑loop stability but only partially capture closed‑loop dynamics, indicating that standard imitation learning does not reliably produce policies that contract errors when perturbed.

By Aryan Goyal
ragroboticsbenchmarks
More like this →