Self-supervised models trained with a contrastive loss such as CLIP have shown to be very powerful in zero-shot classification settings. However, to be used as a zero-shot classifier these models require the user to provide new captions over a fixed set of labels at test time. In many settings, it is hard or impossible to know if a new query caption is compatible with the source captions used to train the model. We address these limitations by framing the zero-shot classification task as an outlier detection problem and develop a conformal prediction procedure to assess when a given test caption may be reliably used. On a real-world medical example, we show that our proposed conformal procedure improves the reliability of CLIP-style models in the zero-shot classification setting, and we provide an empirical analysis of the factors that may affect its performance.
@article{arxiv.2210.15805,
title = {Towards Reliable Zero Shot Classification in Self-Supervised Models with Conformal Prediction},
author = {Bhawesh Kumar and Anil Palepu and Rudraksh Tuwani and Andrew Beam},
journal= {arXiv preprint arXiv:2210.15805},
year = {2022}
}