Search papers, labs, and topics across Lattice.
This paper tackles the problem of prompt learning with large vision-language models under partial label supervision by proposing Holistic Optimal Label Selection (HopS). HopS combines a local density-based filter to select frequent and plausible labels from nearest neighbors with a global optimal transport objective to map uniform sampling to candidate label distributions. Experiments on eight datasets demonstrate that HopS consistently outperforms baselines, showcasing the benefits of holistic label selection for robust prompt learning in weakly supervised settings.
Overcome the limitations of prompt learning with partial labels: a novel approach that leverages both local feature structure and global distribution alignment to select optimal labels.
Prompt learning has gained significant attention as a parameter-efficient approach for adapting large pre-trained vision-language models to downstream tasks. However, when only partial labels are available, its performance is often limited by label ambiguity and insufficient supervisory information. To address this issue, we propose Holistic Optimal Label Selection (HopS), leveraging the generalization ability of pre-trained feature encoders through two complementary strategies. First, we design a local density-based filter that selects the top frequent labels from the nearest neighbors'candidate sets and uses the softmax scores to identify the most plausible label, capturing structural regularities in the feature space. Second, we introduce a global selection objective based on optimal transport that maps the uniform sampling distribution to the candidate label distributions across a batch. By minimizing the expected transport cost, it can determine the most likely label assignments. These two strategies work together to provide robust label selection from both local and global perspectives. Extensive experiments on eight benchmark datasets show that HopS consistently improves performance under partial supervision and outperforms all baselines. Those results highlight the merit of holistic label selection and offer a practical solution for prompt learning in weakly supervised settings.
