Szczegóły publikacji
Opis bibliograficzny
Goal-conditioned decision transformer for multi-goal offline reinforcement learning / Paweł GAJEWSKI, Dominik ŻUREK, Marcin PIETROŃ, Kamil FABER // W: Computational Science – ICCS 2026 : 26th International Conference : Hamburg, Germany, June 29–July 1, 2026 : proceedings , Pt. 2 / eds. Philipp Neumann [et al.]. — Cham : Springer, cop. 2026. — ( Lecture Notes in Computer Science ; ISSN 0302-9743 ; LNCS 16784 ). — ISBN: 978-3-032-29923-9; e-ISBN: 978-3-032-29924-6. — S. 298–306. — Bibliogr., Abstr. — Publikacja dostępna online od: 2026-06-27
Autorzy (4)
Słowa kluczowe
Dane bibliometryczne
| ID BaDAP | 168883 |
|---|---|
| Data dodania do BaDAP | 2026-08-28 |
| DOI | 10.1007/978-3-032-29924-6_22 |
| Rok publikacji | 2026 |
| Typ publikacji | materiały konferencyjne (aut.) |
| Otwarty dostęp |  |
| Wydawca | Springer |
| Konferencja | International Conference on Computational Science 2026 |
| Czasopismo/seria | Lecture Notes in Computer Science |
Abstract
Reinforcement learning (RL) in robotics faces significant hurdles regarding sample efficiency and generalization across varying goals. While Offline RL mitigates the need for costly online interactions, its integration with goal-conditioned policies and transformer-based architectures remains underexplored. We introduce a Goal-Conditioned Decision Transformer adapted for offline multi-goal robotics. By explicitly incorporating goal states into the sequence modeling framework, our approach efficiently solves varying tasks using only pre-collected data. We validate this method on a newly released offline dataset for the Franka Emika Panda platform. Experimental results demonstrate that our approach outperforms state-of-the-art online baselines in complex tasks and maintains robustness in sparse-reward settings, even with limited expert demonstrations.
