DeCoOp: Robust Prompt Tuning with Out-of-Distribution Detection

TL;DR

This paper introduces DeCoOp, a robust prompt tuning method for vision-language models that incorporates out-of-distribution detection to improve performance on open-world tasks involving both base and new classes.

Contribution

The paper proposes DeCoOp, a novel prompt tuning approach that enhances discriminability between base and new classes by integrating out-of-distribution detection within the tuning process.

Findings

DeCoOp outperforms state-of-the-art methods by 2% average accuracy.

Theoretical analysis demonstrates the effectiveness of out-of-distribution detection in open-world prompt tuning.

Experimental validation on 11 datasets confirms the robustness of DeCoOp.

Abstract

Vision-language models (VLMs), such as CLIP, have demonstrated impressive zero-shot capabilities for various downstream tasks. Their performance can be further enhanced through few-shot prompt tuning methods. However, current studies evaluate the performance of learned prompts separately on base and new classes. This evaluation lacks practicality for real-world applications since downstream tasks cannot determine whether the data belongs to base or new classes in advance. In this paper, we explore a problem setting called Open-world Prompt Tuning (OPT), which involves tuning prompts on base classes and evaluating on a combination of base and new classes. By introducing Decomposed Prompt Tuning framework (DePT), we theoretically demonstrate that OPT can be solved by incorporating out-of-distribution detection into prompt tuning, thereby enhancing the base-to-new discriminability. Based on DePT, we present a novel prompt tuning approach, namely, Decomposed Context Optimization (DeCoOp), which introduces new-class detectors and sub-classifiers to further enhance the base-class and new-class discriminability. Experimental results on 11 benchmark datasets validate the effectiveness of DePT and demonstrate that DeCoOp outperforms current state-of-the-art methods, providing a significant 2% average accuracy improvement.