Understanding Retrieval-Augmented Task Adaptation for Vision-Language Models

Yifei Ming, Yixuan Li
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:35719-35743, 2024.

Abstract

Pre-trained contrastive vision-language models have demonstrated remarkable performance across a wide range of tasks. However, they often struggle on fine-trained datasets with categories not adequately represented during pre-training, which makes adaptation necessary. Recent works have shown promising results by utilizing samples from web-scale databases for retrieval-augmented adaptation, especially in low-data regimes. Despite the empirical success, understanding how retrieval impacts the adaptation of vision-language models remains an open research question. In this work, we adopt a reflective perspective by presenting a systematic study to understand the roles of key components in retrieval-augmented adaptation. We unveil new insights on uni-modal and cross-modal retrieval and highlight the critical role of logit ensemble for effective adaptation. We further present theoretical underpinnings that directly support our empirical observations.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-ming24a, title = {Understanding Retrieval-Augmented Task Adaptation for Vision-Language Models}, author = {Ming, Yifei and Li, Yixuan}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {35719--35743}, year = {2024}, editor = {Salakhutdinov, Ruslan and Kolter, Zico and Heller, Katherine and Weller, Adrian and Oliver, Nuria and Scarlett, Jonathan and Berkenkamp, Felix}, volume = {235}, series = {Proceedings of Machine Learning Research}, month = {21--27 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v235/main/assets/ming24a/ming24a.pdf}, url = {https://proceedings.mlr.press/v235/ming24a.html}, abstract = {Pre-trained contrastive vision-language models have demonstrated remarkable performance across a wide range of tasks. However, they often struggle on fine-trained datasets with categories not adequately represented during pre-training, which makes adaptation necessary. Recent works have shown promising results by utilizing samples from web-scale databases for retrieval-augmented adaptation, especially in low-data regimes. Despite the empirical success, understanding how retrieval impacts the adaptation of vision-language models remains an open research question. In this work, we adopt a reflective perspective by presenting a systematic study to understand the roles of key components in retrieval-augmented adaptation. We unveil new insights on uni-modal and cross-modal retrieval and highlight the critical role of logit ensemble for effective adaptation. We further present theoretical underpinnings that directly support our empirical observations.} }
Endnote
%0 Conference Paper %T Understanding Retrieval-Augmented Task Adaptation for Vision-Language Models %A Yifei Ming %A Yixuan Li %B Proceedings of the 41st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Ruslan Salakhutdinov %E Zico Kolter %E Katherine Heller %E Adrian Weller %E Nuria Oliver %E Jonathan Scarlett %E Felix Berkenkamp %F pmlr-v235-ming24a %I PMLR %P 35719--35743 %U https://proceedings.mlr.press/v235/ming24a.html %V 235 %X Pre-trained contrastive vision-language models have demonstrated remarkable performance across a wide range of tasks. However, they often struggle on fine-trained datasets with categories not adequately represented during pre-training, which makes adaptation necessary. Recent works have shown promising results by utilizing samples from web-scale databases for retrieval-augmented adaptation, especially in low-data regimes. Despite the empirical success, understanding how retrieval impacts the adaptation of vision-language models remains an open research question. In this work, we adopt a reflective perspective by presenting a systematic study to understand the roles of key components in retrieval-augmented adaptation. We unveil new insights on uni-modal and cross-modal retrieval and highlight the critical role of logit ensemble for effective adaptation. We further present theoretical underpinnings that directly support our empirical observations.
APA
Ming, Y. & Li, Y.. (2024). Understanding Retrieval-Augmented Task Adaptation for Vision-Language Models. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:35719-35743 Available from https://proceedings.mlr.press/v235/ming24a.html.

Related Material