Causal Representation Learning on High-Dimensional Data: Benchmarks, Reproducibility, and Evaluation Metrics

TL;DR

This paper reviews causal representation learning (CRL) on high-dimensional data, analyzing datasets, proposing a unified evaluation metric, and emphasizing reproducibility to advance model development and comparison.

Contribution

It critically analyzes current datasets, introduces an aggregate performance metric, and assesses reproducibility practices in CRL research.

Findings

Current datasets have notable limitations for CRL evaluation.

A new aggregate metric consolidates multi-directional performance.

Reproducibility issues are prevalent, with gaps in code availability and consistency.

Abstract

Causal representation learning (CRL) models aim to transform high-dimensional data into a latent space, enabling interventions to generate counterfactual samples or modify existing data based on the causal relationships among latent variables. To facilitate the development and evaluation of these models, a variety of synthetic and real-world datasets have been proposed, each with distinct advantages and limitations. For practical applications, CRL models must perform robustly across multiple evaluation directions, including reconstruction, disentanglement, causal discovery, and counterfactual reasoning, using appropriate metrics for each direction. However, this multi-directional evaluation can complicate model comparison, as a model may excel in some direction while under-performing in others. Another significant challenge in this field is reproducibility: the source code corresponding to published results must be publicly available, and repeated runs should yield performance consistent with the original reports. In this study, we critically analyzed the synthetic and real-world datasets currently employed in the literature, highlighting their limitations and proposing a set of essential characteristics for suitable datasets in CRL model development. We also introduce a single aggregate metric that consolidates performance across all evaluation directions, providing a comprehensive score for each model. Finally, we reviewed existing implementations from the literature and assessed them in terms of reproducibility, identifying gaps and best practices in the field.