On the Identifiability of Latent Action Policies
Not on the map. Why?
Abstract (arXiv)
We study the identifiability of latent action policy learning (LAPO), a framework introduced recently to discover representations of actions from video data. We formally describe desiderata for such representations, their statistical benefits and potential sources of unidentifiability. Finally, we prove that an entropy-regularized LAPO objective identifies action representations satisfying our desiderata, under suitable conditions. Our analysis provides an explanation for why discrete action representations perform well in practice.
Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 12 of its 16 references that are in the map. First posted on arXiv 1 Oct 2025.
Builds on
12 of 16 references.
References · 12
- 2016Categorical Reparameterization with Gumbel-Softmaxcited by 2,708
- 2024Genie: Generative Interactive Environmentscited by 706
- 2019Variational Autoencoders and Nonlinear ICA: A Unifying Frameworkcited by 494
- 2021Self-Supervised Learning with Data Augmentations Provably Isolates Content from Stylecited by 308
- 2024Latent Action Pretraining from Videoscited by 302
- 2021Disentanglement via Mechanism Sparsity Regularization: A New Principle for Nonlinear ICAcited by 158
- 2023Learning to Act without Actionscited by 116
- 2023Identifiability Guarantees for Causal Disentanglement from Soft Interventionscited by 84
- 2024Is Behavior Cloning All You Need? Understanding Horizon in Imitation Learningcited by 74
- 2022Function Classes for Identifiable Nonlinear Independent Component Analysiscited by 53
- 2023Nonlinear Independent Component Analysis for Principled Disentanglement in Unsupervised Deep Learningcited by 48
- 2024Nonparametric Partial Disentanglement via Mechanism Sparsity: Sparse Actions, Interventions and Sparse Temporal Dependenciescited by 39
Built on by
Citing papers per year: 2026: 2.
2 papers.