Label Propagation for Zero-shot Classification with Vision-Language Models
It addresses zero-shot classification for vision-language models in scenarios with unlabeled data, representing an incremental improvement.
The paper tackles zero-shot classification with unlabeled data by introducing ZLaP, a label propagation method using geodesic distances on graphs with text and image features, and shows it outperforms latest works on 14 datasets.
Vision-Language Models (VLMs) have demonstrated impressive performance on zero-shot classification, i.e. classification when provided merely with a list of class names. In this paper, we tackle the case of zero-shot classification in the presence of unlabeled data. We leverage the graph structure of the unlabeled data and introduce ZLaP, a method based on label propagation (LP) that utilizes geodesic distances for classification. We tailor LP to graphs containing both text and image features and further propose an efficient method for performing inductive inference based on a dual solution and a sparsification step. We perform extensive experiments to evaluate the effectiveness of our method on 14 common datasets and show that ZLaP outperforms the latest related works. Code: https://github.com/vladan-stojnic/ZLaP