Low-Rank Similarity Mining for Multimodal Dataset Distillation

Yue Xu, Zhilin Lin, Yusong Qiu, Cewu Lu, Yong-Lu Li
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:55144-55161, 2024.

Abstract

Though dataset distillation has witnessed rapid development in recent years, the distillation of multimodal data, e.g., image-text pairs, poses unique and under-explored challenges. Unlike unimodal data, image-text contrastive learning (ITC) data lack inherent categorization and should instead place greater emphasis on modality correspondence. In this work, we propose Low-Rank Similarity Mining (LoRS) for multimodal dataset distillation, that concurrently distills a ground truth similarity matrix with image-text pairs, and leverages low-rank factorization for efficiency and scalability. The proposed approach brings significant improvement to the existing algorithms, marking a significant contribution to the field of visual-language dataset distillation. We advocate adopting LoRS as a foundational synthetic data setup for image-text dataset distillation. Our code is available at https://github.com/silicx/LoRS_Distill.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-xu24q, title = {Low-Rank Similarity Mining for Multimodal Dataset Distillation}, author = {Xu, Yue and Lin, Zhilin and Qiu, Yusong and Lu, Cewu and Li, Yong-Lu}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {55144--55161}, 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/xu24q/xu24q.pdf}, url = {https://proceedings.mlr.press/v235/xu24q.html}, abstract = {Though dataset distillation has witnessed rapid development in recent years, the distillation of multimodal data, e.g., image-text pairs, poses unique and under-explored challenges. Unlike unimodal data, image-text contrastive learning (ITC) data lack inherent categorization and should instead place greater emphasis on modality correspondence. In this work, we propose Low-Rank Similarity Mining (LoRS) for multimodal dataset distillation, that concurrently distills a ground truth similarity matrix with image-text pairs, and leverages low-rank factorization for efficiency and scalability. The proposed approach brings significant improvement to the existing algorithms, marking a significant contribution to the field of visual-language dataset distillation. We advocate adopting LoRS as a foundational synthetic data setup for image-text dataset distillation. Our code is available at https://github.com/silicx/LoRS_Distill.} }
Endnote
%0 Conference Paper %T Low-Rank Similarity Mining for Multimodal Dataset Distillation %A Yue Xu %A Zhilin Lin %A Yusong Qiu %A Cewu Lu %A Yong-Lu 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-xu24q %I PMLR %P 55144--55161 %U https://proceedings.mlr.press/v235/xu24q.html %V 235 %X Though dataset distillation has witnessed rapid development in recent years, the distillation of multimodal data, e.g., image-text pairs, poses unique and under-explored challenges. Unlike unimodal data, image-text contrastive learning (ITC) data lack inherent categorization and should instead place greater emphasis on modality correspondence. In this work, we propose Low-Rank Similarity Mining (LoRS) for multimodal dataset distillation, that concurrently distills a ground truth similarity matrix with image-text pairs, and leverages low-rank factorization for efficiency and scalability. The proposed approach brings significant improvement to the existing algorithms, marking a significant contribution to the field of visual-language dataset distillation. We advocate adopting LoRS as a foundational synthetic data setup for image-text dataset distillation. Our code is available at https://github.com/silicx/LoRS_Distill.
APA
Xu, Y., Lin, Z., Qiu, Y., Lu, C. & Li, Y.. (2024). Low-Rank Similarity Mining for Multimodal Dataset Distillation. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:55144-55161 Available from https://proceedings.mlr.press/v235/xu24q.html.

Related Material