mlsanityData to 14 Oct 2026

A call for better unit testing for invariant risk minimisation

arXiv 2106.03234 ↗First posted 6 Jun 2021cs.LG cs.AI stat.MLcited by 0

Not on the map. Why?

Abstract (arXiv)

In this paper we present a controlled study on the linearized IRM framework (IRMv1) introduced in Arjovsky et al. (2020). We show that IRMv1 (and its variants) framework can be potentially unstable under small changes to the optimal regressor. This can, notably, lead to worse generalisation to new environments, even compared with ERM which converges simply to the global minimum for all training environments mixed up all together. We also highlight the isseus of scaling in the the IRMv1 setup. These observations highlight the importance of rigorous evaluation and importance of unit-testing for measuring progress towards IRM.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 10 of its 14 references that are in the map. First posted on arXiv 6 Jun 2021.

Builds on

10 of 14 references.

References · 10

  1. 2019Invariant Risk Minimizationcited by 1,854
  2. 2020WILDS: A Benchmark of in-the-Wild Distribution Shiftscited by 1,375
  3. 2019Distributionally Robust Neural Networks for Group Shifts: On the Importance of Regularization for Worst-Case Generalizationcited by 1,247
  4. 2018Recognition in Terra Incognitacited by 658
  5. 2021Does Invariant Risk Minimization Capture Invariance?cited by 116
  6. 2020When is invariance useful in an Out-of-Distribution Generalization problem ?cited by 52
  7. 2016Does Distributionally Robust Supervised Learning Give Robust Classifiers?cited by 38
  8. 2021Linear unit-tests for invariance discoverycited by 27
  9. 2020Linear Regression Games: Convergence Guarantees to Approximate Out-of-Distribution Solutionscited by 7
  10. 2019Feature Noise Induces Loss Discrepancy Across Groupscited by 2

Built on by

Not cited yet.