As machine learning models are introduced into every aspect of our lives, and potential benefits become abundant, so do possible catastrophic failures. One of the most common failure scenarios when deploying machine learning models in the wild, which could possibly lead to dire consequences in extreme cases, is the reliance of models on apparently unnatural or irrelevant features.
The issue comes up in a variety of applications: from the reliance of detection models for X-rays on scanner types and marks made by technicians in the hospital, through visual question answering models being sensitive to linguistic variations in the questions, the list of examples for such undesirable behaviors keeps growing.In examples like these, the undesirable behavior stems from the model exploiting a spurious correlation.

Following last year's workshop on Spurious Correlations, Invariance and Stability (SCIS), it is apparent that work on spurious correlations is a long-term effort that spans communities such as fairness, causality-inspired ML, and domains such as NLP, healthcare and many others. Hence we hope that this year's workshop, the second edition of SCIS, will help facilitate this long term effort across communities. The workshop will feature talks by top experts doing methodological work on dealing with spurious correlations, and an extended poster session to allow extensive discussion on work submitted to the workshop.