Enhancing Cross-Modal Fine-Tuning with Gradually Intermediate Modality Generation

Lincan Cai, Shuang Li, Wenxuan Ma, Jingxuan Kang, Binhui Xie, Zixun Sun, Chengwei Zhu
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:5236-5257, 2024.

Abstract

Large-scale pretrained models have proven immensely valuable in handling data-intensive modalities like text and image. However, fine-tuning these models for certain specialized modalities, such as protein sequence and cosmic ray, poses challenges due to the significant modality discrepancy and scarcity of labeled data. In this paper, we propose an end-to-end method, PaRe, to enhance cross-modal fine-tuning, aiming to transfer a large-scale pretrained model to various target modalities. PaRe employs a gating mechanism to select key patches from both source and target data. Through a modality-agnostic Patch Replacement scheme, these patches are preserved and combined to construct data-rich intermediate modalities ranging from easy to hard. By gradually intermediate modality generation, we can not only effectively bridge the modality gap to enhance stability and transferability of cross-modal fine-tuning, but also address the challenge of limited data in the target modality by leveraging enriched intermediate modality data. Compared with hand-designed, general-purpose, task-specific, and state-of-the-art cross-modal fine-tuning approaches, PaRe demonstrates superior performance across three challenging benchmarks, encompassing more than ten modalities.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-cai24c, title = {Enhancing Cross-Modal Fine-Tuning with Gradually Intermediate Modality Generation}, author = {Cai, Lincan and Li, Shuang and Ma, Wenxuan and Kang, Jingxuan and Xie, Binhui and Sun, Zixun and Zhu, Chengwei}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {5236--5257}, 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/cai24c/cai24c.pdf}, url = {https://proceedings.mlr.press/v235/cai24c.html}, abstract = {Large-scale pretrained models have proven immensely valuable in handling data-intensive modalities like text and image. However, fine-tuning these models for certain specialized modalities, such as protein sequence and cosmic ray, poses challenges due to the significant modality discrepancy and scarcity of labeled data. In this paper, we propose an end-to-end method, PaRe, to enhance cross-modal fine-tuning, aiming to transfer a large-scale pretrained model to various target modalities. PaRe employs a gating mechanism to select key patches from both source and target data. Through a modality-agnostic Patch Replacement scheme, these patches are preserved and combined to construct data-rich intermediate modalities ranging from easy to hard. By gradually intermediate modality generation, we can not only effectively bridge the modality gap to enhance stability and transferability of cross-modal fine-tuning, but also address the challenge of limited data in the target modality by leveraging enriched intermediate modality data. Compared with hand-designed, general-purpose, task-specific, and state-of-the-art cross-modal fine-tuning approaches, PaRe demonstrates superior performance across three challenging benchmarks, encompassing more than ten modalities.} }
Endnote
%0 Conference Paper %T Enhancing Cross-Modal Fine-Tuning with Gradually Intermediate Modality Generation %A Lincan Cai %A Shuang Li %A Wenxuan Ma %A Jingxuan Kang %A Binhui Xie %A Zixun Sun %A Chengwei Zhu %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-cai24c %I PMLR %P 5236--5257 %U https://proceedings.mlr.press/v235/cai24c.html %V 235 %X Large-scale pretrained models have proven immensely valuable in handling data-intensive modalities like text and image. However, fine-tuning these models for certain specialized modalities, such as protein sequence and cosmic ray, poses challenges due to the significant modality discrepancy and scarcity of labeled data. In this paper, we propose an end-to-end method, PaRe, to enhance cross-modal fine-tuning, aiming to transfer a large-scale pretrained model to various target modalities. PaRe employs a gating mechanism to select key patches from both source and target data. Through a modality-agnostic Patch Replacement scheme, these patches are preserved and combined to construct data-rich intermediate modalities ranging from easy to hard. By gradually intermediate modality generation, we can not only effectively bridge the modality gap to enhance stability and transferability of cross-modal fine-tuning, but also address the challenge of limited data in the target modality by leveraging enriched intermediate modality data. Compared with hand-designed, general-purpose, task-specific, and state-of-the-art cross-modal fine-tuning approaches, PaRe demonstrates superior performance across three challenging benchmarks, encompassing more than ten modalities.
APA
Cai, L., Li, S., Ma, W., Kang, J., Xie, B., Sun, Z. & Zhu, C.. (2024). Enhancing Cross-Modal Fine-Tuning with Gradually Intermediate Modality Generation. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:5236-5257 Available from https://proceedings.mlr.press/v235/cai24c.html.

Related Material