Metraipolis › archive › dc_889d5cbcf276

Learning Stabilizing Latent Feedback Controllers via Multi-Step Backward Losses — arXiv:2610.40931

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.40931
License (as archived)
CC0-1.0
Archived
2026-11-03T08:12:48.399Z
SHA-256 of the archived text
3c08dc7af6f9f038a025dd3bea8f37d5b72eba2d51e86ea883cfe6b79bad8973
Stable URL of this page
https://metraipolis.com/documento/dc_889d5cbcf276

How to cite

Learning Stabilizing Latent Feedback Controllers via Multi-Step Backward Losses — arXiv:2610.40931 — arXiv (en), https://arxiv.org/abs/2610.40931 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_889d5cbcf276 (Metraipolis archive dc_889d5cbcf276, sha256 3c08dc7af6f9f038a025dd3bea8f37d5b72eba2d51e86ea883cfe6b79bad8973)
BibTeX
@misc{dc889d5cbcf276,
  title = {Learning Stabilizing Latent Feedback Controllers via Multi-Step Backward Losses — arXiv:2610.40931},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.40931},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_889d5cbcf276; sha256 3c08dc7af6f9f038a025dd3bea8f37d5b72eba2d51e86ea883cfe6b79bad8973},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.40931. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Can learned latent dynamics models support the design of feedback controllers that provably stabilize the original system? For linear time-invariant systems, we show that learning linear autoencoders and latent dynamics with multi-step prediction-reconstruction losses $\unicode{x2014}$ which we term "backward losses" as they encourage latent trajectories to reconstruct full state trajectories $\unicode{x2014}$ provably recovers the structure needed for stabilization. Specifically, a sufficiently long reconstruction horizon ensures that a minimum ${H}_\infty$-norm controller designed using the learned latent model stabilizes the original system through output-feedback via the corresponding encoder, provided the latent dimension is large enough to capture all non-stable modes of the underlying system. We also show why both reconstruction and a multi-step horizon are essential: forward prediction losses that require only low prediction error along latent trajectories admit global loss minimizers that discard unstable modes regardless of horizon length, while single-step reconstruction losses can yield latent models whose stabilizing controllers fail to stabilize the original system. Finally, experiments on a non-linear flexible-arm system illustrate the practical implications of our results for learning non-linear autoencoders and latent dynamics.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Optimization and Control (math.OC) ; Machine Learning (cs.LG); Systems and Control (eess.SY).
Pagina del lavoro: https://arxiv.org/abs/2610.40931 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-11-03T08:02:38.971Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).