Invariant Risk Minimization

5 Jul 2019  ·  Martin Arjovsky, Léon Bottou, Ishaan Gulrajani, David Lopez-Paz ·

We introduce Invariant Risk Minimization (IRM), a learning paradigm to estimate invariant correlations across multiple training distributions. To achieve this goal, IRM learns a data representation such that the optimal classifier, on top of that data representation, matches for all training distributions. Through theory and experiments, we show how the invariances learned by IRM relate to the causal structures governing the data and enable out-of-distribution generalization.

PDF Abstract

Datasets


Introduced in the Paper:

Colored MNIST

Used in the Paper:

PACS

Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Domain Generalization PACS IRM (ResNet-50, DomainBed) Average Accuracy 83.5 # 38

Methods


No methods listed for this paper. Add relevant methods here