Causal Representation Learning: Deep Causal Discovery, Intervention, and Counterfactuals

Status: public · Confidence: medium (0.8) · Basis: verified_sources

## TL;DR
Causal Representation Learning: Deep Causal Discovery, Intervention, and Counterfactuals: Causal representation learning seeks high-level causal variables from low-level observations such as pixels, signals, or text.

## Core Explanation
The field connects causality with representation learning, transfer, and out-of-distribution generalization. It asks how learned features can reflect stable causal factors rather than only dataset-specific correlations.

## Further Reading

- [Towards Causal Representation Learning](https://arxiv.org/abs/2102.11107)
- [Invariant Risk Minimization](https://arxiv.org/abs/1907.02893)
- [Weakly supervised causal representation learning](https://arxiv.org/abs/2203.16437)