Useful Confidence Measures: Beyond the Max Score
This work addresses the need for reliable confidence estimation in safety-critical ML applications, particularly for NLP tasks with distribution shifts, though it is incremental as it builds on existing confidence measure concepts.
The paper tackled the problem of unreliable confidence measures in machine learning classifiers by proposing and evaluating alternative measures beyond the maximum score, showing that entropy-based confidence is highly effective, especially for out-of-distribution data where using only the maximum score is suboptimal.
An important component in deploying machine learning (ML) in safety-critic applications is having a reliable measure of confidence in the ML model's predictions. For a classifier $f$ producing a probability vector $f(x)$ over the candidate classes, the confidence is typically taken to be $\max_i f(x)_i$. This approach is potentially limited, as it disregards the rest of the probability vector. In this work, we derive several confidence measures that depend on information beyond the maximum score, such as margin-based and entropy-based measures, and empirically evaluate their usefulness, focusing on NLP tasks with distribution shifts and Transformer-based models. We show that when models are evaluated on the out-of-distribution data ``out of the box'', using only the maximum score to inform the confidence measure is highly suboptimal. In the post-processing regime (where the scores of $f$ can be improved using additional in-distribution held-out data), this remains true, albeit less significant. Overall, our results suggest that entropy-based confidence is a surprisingly useful measure.