Identifiable Exchangeable Mechanisms for Causal Structure and Representation Learning
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866916603859304448 |
|---|---|
| author | Reizinger, Patrik Guo, Siyuan Huszár, Ferenc Schölkopf, Bernhard Brendel, Wieland |
| author_facet | Reizinger, Patrik Guo, Siyuan Huszár, Ferenc Schölkopf, Bernhard Brendel, Wieland |
| contents | Identifying latent representations or causal structures is important for good generalization and downstream task performance. However, both fields have been developed rather independently. We observe that several methods in both representation and causal structure learning rely on the same data-generating process (DGP), namely, exchangeable but not i.i.d. (independent and identically distributed) data. We provide a unified framework, termed Identifiable Exchangeable Mechanisms (IEM), for representation and structure learning under the lens of exchangeability. IEM provides new insights that let us relax the necessary conditions for causal structure identification in exchangeable non--i.i.d. data. We also demonstrate the existence of a duality condition in identifiable representation learning, leading to new identifiability results. We hope this work will pave the way for further research in causal representation learning. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_14302 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Identifiable Exchangeable Mechanisms for Causal Structure and Representation Learning Reizinger, Patrik Guo, Siyuan Huszár, Ferenc Schölkopf, Bernhard Brendel, Wieland Machine Learning Artificial Intelligence Identifying latent representations or causal structures is important for good generalization and downstream task performance. However, both fields have been developed rather independently. We observe that several methods in both representation and causal structure learning rely on the same data-generating process (DGP), namely, exchangeable but not i.i.d. (independent and identically distributed) data. We provide a unified framework, termed Identifiable Exchangeable Mechanisms (IEM), for representation and structure learning under the lens of exchangeability. IEM provides new insights that let us relax the necessary conditions for causal structure identification in exchangeable non--i.i.d. data. We also demonstrate the existence of a duality condition in identifiable representation learning, leading to new identifiability results. We hope this work will pave the way for further research in causal representation learning. |
| title | Identifiable Exchangeable Mechanisms for Causal Structure and Representation Learning |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2406.14302 |