Sequential Representation Learning via Static-Dynamic Conditional Disentanglement
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866916352945553408 |
|---|---|
| author | Simon, Mathieu Cyrille Frossard, Pascal De Vleeschouwer, Christophe |
| author_facet | Simon, Mathieu Cyrille Frossard, Pascal De Vleeschouwer, Christophe |
| contents | This paper explores self-supervised disentangled representation learning within sequential data, focusing on separating time-independent and time-varying factors in videos. We propose a new model that breaks the usual independence assumption between those factors by explicitly accounting for the causal relationship between the static/dynamic variables and that improves the model expressivity through additional Normalizing Flows. A formal definition of the factors is proposed. This formalism leads to the derivation of sufficient conditions for the ground truth factors to be identifiable, and to the introduction of a novel theoretically grounded disentanglement constraint that can be directly and efficiently incorporated into our new framework. The experiments show that the proposed approach outperforms previous complex state-of-the-art techniques in scenarios where the dynamics of a scene are influenced by its content. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2408_05599 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Sequential Representation Learning via Static-Dynamic Conditional Disentanglement Simon, Mathieu Cyrille Frossard, Pascal De Vleeschouwer, Christophe Machine Learning Artificial Intelligence Computer Vision and Pattern Recognition This paper explores self-supervised disentangled representation learning within sequential data, focusing on separating time-independent and time-varying factors in videos. We propose a new model that breaks the usual independence assumption between those factors by explicitly accounting for the causal relationship between the static/dynamic variables and that improves the model expressivity through additional Normalizing Flows. A formal definition of the factors is proposed. This formalism leads to the derivation of sufficient conditions for the ground truth factors to be identifiable, and to the introduction of a novel theoretically grounded disentanglement constraint that can be directly and efficiently incorporated into our new framework. The experiments show that the proposed approach outperforms previous complex state-of-the-art techniques in scenarios where the dynamics of a scene are influenced by its content. |
| title | Sequential Representation Learning via Static-Dynamic Conditional Disentanglement |
| topic | Machine Learning Artificial Intelligence Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2408.05599 |