Temporally Disentangled Representation Learning under Unknown Nonstationarity

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Song, Xiangchen, Yao, Weiran, Fan, Yewen, Dong, Xinshuai, Chen, Guangyi, Niebles, Juan Carlos, Xing, Eric, Zhang, Kun
Natura: Preprint
Pubblicazione: 2023
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866917738364010496
author Song, Xiangchen
Yao, Weiran
Fan, Yewen
Dong, Xinshuai
Chen, Guangyi
Niebles, Juan Carlos
Xing, Eric
Zhang, Kun
author_facet Song, Xiangchen
Yao, Weiran
Fan, Yewen
Dong, Xinshuai
Chen, Guangyi
Niebles, Juan Carlos
Xing, Eric
Zhang, Kun
contents In unsupervised causal representation learning for sequential data with time-delayed latent causal influences, strong identifiability results for the disentanglement of causally-related latent variables have been established in stationary settings by leveraging temporal structure. However, in nonstationary setting, existing work only partially addressed the problem by either utilizing observed auxiliary variables (e.g., class labels and/or domain indexes) as side information or assuming simplified latent causal dynamics. Both constrain the method to a limited range of scenarios. In this study, we further explored the Markov Assumption under time-delayed causally related process in nonstationary setting and showed that under mild conditions, the independent latent components can be recovered from their nonlinear mixture up to a permutation and a component-wise transformation, without the observation of auxiliary variables. We then introduce NCTRL, a principled estimation framework, to reconstruct time-delayed latent causal variables and identify their relations from measured sequential data only. Empirical evaluations demonstrated the reliable identification of time-delayed latent causal influences, with our methodology substantially outperforming existing baselines that fail to exploit the nonstationarity adequately and then, consequently, cannot distinguish distribution shifts.
format Preprint
id arxiv_https___arxiv_org_abs_2310_18615
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Temporally Disentangled Representation Learning under Unknown Nonstationarity
Song, Xiangchen
Yao, Weiran
Fan, Yewen
Dong, Xinshuai
Chen, Guangyi
Niebles, Juan Carlos
Xing, Eric
Zhang, Kun
Machine Learning
In unsupervised causal representation learning for sequential data with time-delayed latent causal influences, strong identifiability results for the disentanglement of causally-related latent variables have been established in stationary settings by leveraging temporal structure. However, in nonstationary setting, existing work only partially addressed the problem by either utilizing observed auxiliary variables (e.g., class labels and/or domain indexes) as side information or assuming simplified latent causal dynamics. Both constrain the method to a limited range of scenarios. In this study, we further explored the Markov Assumption under time-delayed causally related process in nonstationary setting and showed that under mild conditions, the independent latent components can be recovered from their nonlinear mixture up to a permutation and a component-wise transformation, without the observation of auxiliary variables. We then introduce NCTRL, a principled estimation framework, to reconstruct time-delayed latent causal variables and identify their relations from measured sequential data only. Empirical evaluations demonstrated the reliable identification of time-delayed latent causal influences, with our methodology substantially outperforming existing baselines that fail to exploit the nonstationarity adequately and then, consequently, cannot distinguish distribution shifts.
title Temporally Disentangled Representation Learning under Unknown Nonstationarity
topic Machine Learning
url https://arxiv.org/abs/2310.18615