Causal Component Analysis
Dec 10, 2023
Speakers
Liang Wendong
Speaker · 0 followers
Armin Kekic
Speaker · 1 follower
Julius von Kügelgen
Speaker · 0 followers
Simon Buchholz
Speaker · 0 followers
Michel Besserve
Speaker · 0 followers
Luigi Gresele
Speaker · 0 followers
Bernhard Scholkopf
Speaker · 0 followers
About
Independent Component Analysis (ICA) aims to recover independent latent variables from observed mixtures thereof. Causal Representation Learning (CRL) aims instead to infer causally related (thus often statistically _dependent_) latent variables, together with the unknown graph encoding their causal relationships. We introduce an intermediate problem termed _Causal Component Analysis (CauCA)_. CauCA can be viewed as a generalization of ICA, modelling the causal dependence among the latent components, and as a special case of CRL. In contrast to CRL, it presupposes knowledge of the causal graph, focusing solely on learning the unmixing function and the causal mechanisms. Any impossibility results regarding the recovery of the ground truth in CauCA also apply for CRL, while possibility results may serve as a stepping stone for extensions to CRL. We characterize CauCA identifiability from multiple datasets generated through different types of interventions on the latent causal variables. As a corollary, this interventional perspective also leads to new identifiability results for nonlinear ICA—a special case of CauCA with an empty graph—requiring strictly fewer datasets than previous results. We introduce a likelihood-based approach using normalizing flows to estimate both the unmixing function and the causal mechanisms, and demonstrate its effectiveness through extensive synthetic experiments in the CauCA and ICA setting.Independent Component Analysis (ICA) aims to recover independent latent variables from observed mixtures thereof. Causal Representation Learning (CRL) aims instead to infer causally related (thus often statistically _dependent_) latent variables, together with the unknown graph encoding their causal relationships. We introduce an intermediate problem termed _Causal Component Analysis (CauCA)_. CauCA can be viewed as a generalization of ICA, modelling the causal dependence among the latent compon…
Organizer
NeurIPS 2023
Account · 651 followers
Like the format? Trust SlidesLive to capture your next event!
Professional recording and live streaming, delivered globally.
Sharing
Recommended Videos
Presentations on similar topic, category or speaker
A One-Size-Fits-All Approach to Improving Randomness in Paper Assignment
Watch later
Yixuan Xu, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Efficient Hyper-parameter Optimization with Cubic Regularization
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Regret-Optimal Model-Free RL for Discounted MDPs with Short Burn-In Time
Watch later
Xiang Ji, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Polyhedron Attention Module: Learning Adaptive-order Interactions
Watch later
Tan Zhu, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Entropic Neural Optimal Transport via Diffusion Processes
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Continuous-time Graph Representation with Sequential Survival Process
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%