Towards Causal Relationship in Indefinite Data: Baseline Model and New Datasets

TL;DR

This paper introduces new datasets and a probabilistic baseline model to address the challenges of learning causal structures from complex, indefinite data like dialogue and video, which contain multiple structures and values.

Contribution

The paper releases two high-quality datasets, Causalogue and Causaction, and proposes a novel probabilistic framework to handle multi-structure, multi-value indefinite data for causal discovery.

Findings

Datasets enable causal analysis in dialogue and video data.

The probabilistic model effectively captures causal structures and representations.

Experimental results demonstrate the model's ability to disentangle causal effects and confounders.

Abstract

Integrating deep learning and causal discovery has encouraged us to spot that learning causal structures and representations in dialogue and video is full of challenges. We defined These data forms as "Indefinite Data", characterized by multi-structure data and multi-value representations. Unlike existing adaptable data forms, Indefinite Data still faces gaps in datasets and methods. To address the dataset gap, we release two high-quality datasets - Causalogue and Causaction, containing text dialogue samples and video action samples with causal annotations respectively. Moreover, the method gap arises from the coexistence of multi-structure data and multi-value representations, breaking the assumptions of all current methods and rendering them infeasible on Indefinite Data. To this end, we propose a probabilistic framework as a baseline, incorporating three designed highlights for this gap: 1) establishing Causation Condition of representations using the independence of noise terms under non-fixed causal structures, 2) treating causal strength as a latent variable and measuring the reconstruction loss in the correlation space, and 3) estimating the effects of latent confounders. These highpoints make the probabilistic model capable of overcoming challenges brought by the coexistence of multi-structure data and multi-value representations and pave the way for the extension of latent confounders. Comprehensive experiments have evaluated baseline results of causal structures, causal representations, and confounding disentanglement.