Do-Operation Guided Causal Representation Learning with Reduced Supervision Strength

TL;DR

This paper introduces a causal representation learning framework that uses do-operation to reduce supervision needs by leveraging causal invariances, with new metrics for evaluation and demonstrated improvements on synthetic and real data.

Contribution

It proposes a novel do-operation based method for causal representation learning that reduces supervision requirements and introduces improved evaluation metrics.

Findings

Outperforms state-of-the-art methods on synthetic datasets.

Demonstrates effectiveness on real-world datasets.

Provides new metrics for causal representation evaluation.

Abstract

Causal representation learning has been proposed to encode relationships between factors presented in the high dimensional data. However, existing methods suffer from merely using a large amount of labeled data and ignore the fact that samples generated by the same causal mechanism follow the same causal relationships. In this paper, we seek to explore such information by leveraging do-operation to reduce supervision strength. We propose a framework that implements do-operation by swapping latent cause and effect factors encoded from a pair of inputs. Moreover, we also identify the inadequacy of existing causal representation metrics empirically and theoretically and introduce new metrics for better evaluation. Experiments conducted on both synthetic and real datasets demonstrate the superiorities of our method compared with state-of-the-art methods.