Metraipolis › archive › dc_71fc55682b49
Truncated backpropagation through time induces a two-sided gradient turnpike property — arXiv:2611.07415
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.07415
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-06T14:51:10.596Z
- SHA-256 of the archived text
944d70a79ff0f515c3e03a325881825230e42f958d946a1567663a38b00c0f75- Stable URL of this page
https://metraipolis.com/documento/dc_71fc55682b49
How to cite
Truncated backpropagation through time induces a two-sided gradient turnpike property — arXiv:2611.07415 — arXiv (en), https://arxiv.org/abs/2611.07415 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_71fc55682b49 (Metraipolis archive dc_71fc55682b49, sha256 944d70a79ff0f515c3e03a325881825230e42f958d946a1567663a38b00c0f75)
BibTeX
@misc{dc71fc55682b49,
title = {Truncated backpropagation through time induces a two-sided gradient turnpike property — arXiv:2611.07415},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.07415},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_71fc55682b49; sha256 944d70a79ff0f515c3e03a325881825230e42f958d946a1567663a38b00c0f75},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.07415. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Truncated backpropagation through time reduces the computational burden of training recurrent models by restricting gradient computations to shorter temporal contexts. This induces two distinct boundary effects: missing past information perturbs the forward state evolution, while truncating the backward recursion omits future information. This motivates decoupling the interval over which gradient contributions are averaged for the parameter update from the surrounding temporal context over which they are computed. Under exponential decay of the state-transition Jacobians and mild regularity conditions, this leads to a two-sided gradient turnpike property: the error between truncated- and full-context gradient contributions decays exponentially with the distance from both context boundaries. We derive bounds for the averaged update gradient, yielding guarantees on gradient alignment and approximate full-context stationarity; under unbiased interval sampling and a Polyak-Lojasiewicz condition, these further imply a bound on the resulting optimality gap. Numerical experiments on the Wiener-Hammerstein benchmark illustrate the role of past and future context and the trade-off between predictive performance and computational cost. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno). Categorie arXiv: Systems and Control (eess.SY) ; Machine Learning (cs.LG); Optimization and Control (math.OC). Pagina del lavoro: https://arxiv.org/abs/2611.07415 — 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 2026-11-06T14:21:42.659Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).