CVJul 29, 2024

Advancing Prompt Learning through an External Layer

arXiv:2407.19674v65 citationsh-index: 23
Originality Incremental advance
AI Analysis

This work addresses generalization issues in prompt learning for vision-language models, offering an incremental improvement for tasks like few-shot learning and domain shifts.

The paper tackles the poor generalization of prompt learning in vision-language models by introducing an external layer with learnable visual embeddings and a two-pronged alignment approach, achieving superior performance in four experiments across 15 datasets compared to existing methods.

Prompt learning represents a promising method for adapting pre-trained vision-language models (VLMs) to various downstream tasks by learning a set of text embeddings. One challenge inherent to these methods is the poor generalization performance due to the invalidity of the learned text embeddings for unseen tasks. A straightforward approach to bridge this gap is to freeze the text embeddings in prompts, which results in a lack of capacity to adapt VLMs for downstream tasks. To address this dilemma, we propose a paradigm called EnPrompt with a novel External Layer (EnLa). Specifically, we propose a textual external layer and learnable visual embeddings for adapting VLMs to downstream tasks. The learnable external layer is built upon valid embeddings of pre-trained CLIP. This design considers the balance of learning capabilities between the two branches. To align the textual and visual features, we propose a novel two-pronged approach: i) we introduce the optimal transport as the discrepancy metric to align the vision and text modalities, and ii) we introduce a novel strengthening feature to enhance the interaction between these two modalities. Four representative experiments (i.e., base-to-novel generalization, few-shot learning, cross-dataset generalization, domain shifts generalization) across 15 datasets demonstrate that our method outperforms the existing prompt learning method.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes