Anatomy of a failure: When, how, and why deep vision fails in scientific domains

TL;DR

Deep learning often fails in scientific imaging due to modality-specific issues, especially with infrared data, leading to catastrophic prediction collapse despite rich information content.

Contribution

This paper reveals fundamental failure modes of deep learning in scientific imaging, highlighting the mismatch between data priors and model biases, and proposes a framework for modality-specific analysis.

Findings

DL models underperform on IR data despite its richness.

IR data priors interact poorly with DL's simplicity bias, causing collapse.

State-of-the-art robustification strategies do not prevent these failures.

Abstract

Mirroring its ubiquity in popular media and all human activities, the use of deep learning (DL) is rapidly growing in scientific imaging modalities. However, unlike everyday RGB pictures, pixels encode precise physicochemical properties in scientific imaging across potentially thousands of channels. While DL is well validated on human-centric RGB perceptual tasks, its effectiveness for scientific imaging remains uncertain. Here, we show that the naive application of DL frameworks to scientific images can lead to critical failures. We evaluate the use of DL for pathology, comparing RGB images of stained tissue with the quantitative and information-rich biochemical signatures of infrared (IR) imaging. Despite this informational advantage, DL models trained on IR data paradoxically underperform. We investigate this discrepancy to find that IR data priors interact poorly with the simplicity bias of DL, causing models to collapse to one-dimensional predictions. This constitutes a catastrophic DL failure because the model's representational capacity remains largely unused, while furthermore raising AI safety concerns and undermining the advantages of such scientific modalities. Notably, this problem persists even with state-of-the-art DL robustification strategies, which are primarily designed and validated for RGB imagery and thus inherit the same prior-bias mismatch. This work establishes a framework for understanding the limitations of generic DL in science and advocates for the study of modality-specific failure modes to guide the development of specialized, safe AI algorithms.