What could go wrong? Discovering and describing failure modes in computer vision
This work addresses the brittleness of deep learning models for practitioners, aiming to improve safety and interpretability, though it is incremental as it builds on existing vision-and-language methods.
The paper tackles the problem of predicting and describing failure modes of computer vision models using natural language, formalizing Language-Based Error Explainability (LBEE) and proposing metrics and solutions that isolate nontrivial sentences for specific error causes like unseen objects or adverse conditions.
Deep learning models are effective, yet brittle. Even carefully trained, their behavior tends to be hard to predict when confronted with out-of-distribution samples. In this work, our goal is to propose a simple yet effective solution to predict and describe via natural language potential failure modes of computer vision models. Given a pretrained model and a set of samples, our aim is to find sentences that accurately describe the visual conditions in which the model underperforms. In order to study this important topic and foster future research on it, we formalize the problem of Language-Based Error Explainability (LBEE) and propose a set of metrics to evaluate and compare different methods for this task. We propose solutions that operate in a joint vision-and-language embedding space, and can characterize through language descriptions model failures caused, e.g., by objects unseen during training or adverse visual conditions. We experiment with different tasks, such as classification under the presence of dataset bias and semantic segmentation in unseen environments, and show that the proposed methodology isolates nontrivial sentences associated with specific error causes. We hope our work will help practitioners better understand the behavior of models, increasing their overall safety and interpretability.