Identifying and Benchmarking Natural Out-of-Context Prediction Problems
Abstract
Deep learning systems frequently fail at out-of-context (OOC) prediction, the problem of making reliable predictions on uncommon or unusual inputs or subgroups of the training distribution. To this end, a number of benchmarks for measuring OOC performance have recently been introduced. In this work, we introduce a framework unifying the literature on OOC performance measurement, and demonstrate how rich auxiliary information can be leveraged to identify candidate sets of OOC examples in existing datasets. We present NOOCh: a suite of naturally-occurring"challenge sets", and show how varying notions of context can be used to probe specific OOC failure modes. Experimentally, we explore the tradeoffs between various learning approaches on these challenge sets and demonstrate how the choices made in designing OOC benchmarks can yield varying conclusions.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
Lineage
Builds on10
- Out-of-Distribution Generalization via Risk Extrapolation (REx)David Krueger, Ethan Caballero, Jörn-Henrik Jacobsen, Amy Zhang et al.ICML 2021 · 1,163 citations
- The Pitfalls of Simplicity Bias in Neural NetworksHarshay Shah, Kaustav Tamuly, Aditi Raghunathan, Prateek Jain et al.NeurIPS 2020 · 503 citations
- Environment Inference for Invariant LearningElliot Creager, Jörn-Henrik Jacobsen, Richard S. ZemelICML 2021 · 454 citations
- Noise or Signal: The Role of Image Backgrounds in Object RecognitionKai Yuanqing Xiao, Logan Engstrom, Andrew Ilyas, Aleksander MadryICLR 2021 · 451 citations
- Evaluating Models' Local Decision Boundaries via Contrast SetsMatt Gardner, Yoav Artzi, Victoria Basmova, Jonathan Berant et al.EMNLP Findings 2020 · 260 citations
Related papers
- Detecting Semantic AnomaliesFaruk Ahmed, Aaron C. CourvilleAAAI 2020 · 93 citations
- In or Out? Fixing ImageNet Out-of-Distribution Detection EvaluationJulian Bitterwolf, Maximilian Müller, Matthias HeinICML 2023 · 154 citations
- ORIC: Benchmarking Object Recognition under Contextual Incongruity in Large Vision-Language ModelsZhaoyang Li, Zhan Ling, Yuchen Zhou, Litian Gong et al.CVPR 2026 · 2 citations
- A framework for benchmarking Class-out-of-distribution detection and its application to ImageNetIdo Galil, Mohammed Dabbah, Ran El-YanivICLR 2023 · 2 citations
- Understanding Failures in Out-of-Distribution Detection with Deep Generative ModelsLily H. Zhang, Mark Goldstein, Rajesh RanganathICML 2021 · 129 citations
