Invariant Representation Learning for Generalizable Imitation - Equipe Robot interaction, Ambient system, Machine learning, Behaviour, Optimization
Preprints, Working Papers, ... Year : 2024

Invariant Representation Learning for Generalizable Imitation

Abstract

We address the problem of learning imitation policies that generalize across environments sharing the same underlying causal structure between the system dynamics and the task. We introduce a novel loss for learning invariant state representations that draws inspiration from adversarial robustness. Our approach is algorithm-agnostic and does not require knowledge of domain labels. Yet, evaluation in visual and non-visual environments reveals improved zero-shot generalization in the presence of spurious features compared to previous works.
Fichier principal
Vignette du fichier
ES2024-18-4.pdf (355.92 Ko) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-04613937 , version 1 (19-06-2024)
hal-04613937 , version 2 (09-09-2024)

Identifiers

  • HAL Id : hal-04613937 , version 2

Cite

Mohamed Khalil Jabri, Panagiotis Papadakis, Ehsan Abbasnejad, Gilles Coppin, Javen Shi. Invariant Representation Learning for Generalizable Imitation. 2024. ⟨hal-04613937v2⟩
33 View
65 Download

Share

More