On the Identification of Temporal Causal Representation with Instantaneous Dependence
Li, Zijian Shen, Yifan Zheng, Kaitao Cai, Ruichu Song, Xiangchen Gong, Mingming Chen, Guangyi Zhang, Kun
Li, Zijian
Shen, Yifan
Zheng, Kaitao
Cai, Ruichu
Song, Xiangchen
Gong, Mingming
Chen, Guangyi
Zhang, Kun
Supervisor
Department
Machine Learning
Embargo End Date
Type
Conference proceeding
Date
2025
License
Language
English
Collections
Research Projects
Organizational Units
Journal Issue
Abstract
Temporally causal representation learning aims to identify the latent causal process from time series observations, but most methods require the assumption that the latent causal processes do not have instantaneous relations. Although some recent methods achieve identifiability in the instantaneous causality case, they require either interventions on the latent variables or grouping of the observations, which are in general difficult to obtain in real-world scenarios. To fill this gap, we propose an IDentification framework for instantaneOus Latent dynamics (IDOL) by imposing a sparse influence constraint that the latent causal processes have sparse time-delayed and instantaneous relations. Specifically, we establish identifiability results of the latent causal process up to a Markov equivalence class based on sufficient variability and the sparse influence constraint by employing contextual information. We further explore under what conditions the identification can be extended to the causal graph. Based on these theoretical results, we incorporate a temporally variational inference architecture to estimate the latent variables and a gradient-based sparsity regularization to identify the latent causal process. Experimental results on simulation datasets illustrate that our method can identify the latent causal process. Furthermore, evaluations on human motion forecasting benchmarks indicate the effectiveness in real-world settings. Source code is available at https://github.com/DMIRLAB-Group/IDOL. © 2025 13th International Conference on Learning Representations, ICLR 2025. All rights reserved.
Citation
Z. Li et al., “On the Identification of Temporal Causal Representation with Instantaneous Dependence,” International Conference on Representation Learning, vol. 2025, pp. 35203–35242, May 2025
Source
13th International Conference on Learning Representations, ICLR 2025
Conference
13th International Conference on Learning Representations, ICLR 2025
Keywords
Subjects
Source
13th International Conference on Learning Representations, ICLR 2025
Publisher
International Conference on Learning Representations, ICLR