A Sparsity Principle for Partially Observable Causal Representation Learning
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866916288288260096 |
|---|---|
| author | Xu, Danru Yao, Dingling Lachapelle, Sébastien Taslakian, Perouz von Kügelgen, Julius Locatello, Francesco Magliacane, Sara |
| author_facet | Xu, Danru Yao, Dingling Lachapelle, Sébastien Taslakian, Perouz von Kügelgen, Julius Locatello, Francesco Magliacane, Sara |
| contents | Causal representation learning aims at identifying high-level causal variables from perceptual data. Most methods assume that all latent causal variables are captured in the high-dimensional observations. We instead consider a partially observed setting, in which each measurement only provides information about a subset of the underlying causal state. Prior work has studied this setting with multiple domains or views, each depending on a fixed subset of latents. Here, we focus on learning from unpaired observations from a dataset with an instance-dependent partial observability pattern. Our main contribution is to establish two identifiability results for this setting: one for linear mixing functions without parametric assumptions on the underlying causal model, and one for piecewise linear mixing functions with Gaussian latent causal variables. Based on these insights, we propose two methods for estimating the underlying causal variables by enforcing sparsity in the inferred representation. Experiments on different simulated datasets and established benchmarks highlight the effectiveness of our approach in recovering the ground-truth latents. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2403_08335 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | A Sparsity Principle for Partially Observable Causal Representation Learning Xu, Danru Yao, Dingling Lachapelle, Sébastien Taslakian, Perouz von Kügelgen, Julius Locatello, Francesco Magliacane, Sara Machine Learning Artificial Intelligence Causal representation learning aims at identifying high-level causal variables from perceptual data. Most methods assume that all latent causal variables are captured in the high-dimensional observations. We instead consider a partially observed setting, in which each measurement only provides information about a subset of the underlying causal state. Prior work has studied this setting with multiple domains or views, each depending on a fixed subset of latents. Here, we focus on learning from unpaired observations from a dataset with an instance-dependent partial observability pattern. Our main contribution is to establish two identifiability results for this setting: one for linear mixing functions without parametric assumptions on the underlying causal model, and one for piecewise linear mixing functions with Gaussian latent causal variables. Based on these insights, we propose two methods for estimating the underlying causal variables by enforcing sparsity in the inferred representation. Experiments on different simulated datasets and established benchmarks highlight the effectiveness of our approach in recovering the ground-truth latents. |
| title | A Sparsity Principle for Partially Observable Causal Representation Learning |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2403.08335 |