Deviation-based multiple coefficient item mixer for heterogeneous set-to-set matching

Hirotaka Hachiya, Kajishiro Yukito
Proceedings of the 17th Asian Conference on Machine Learning, PMLR 304:129-144, 2025.

Abstract

Heterogeneous set-to-set matching tasks such as fashion outfit recommendation, require permutation-invariant and dynamic item-wise transformations to bring compatible sets closer while pushing incompatible ones apart. While attention-based methods satisfy the permutation invariance requirement, they often suffer from convex hull limitations due to their reliance on softmax-based dot-product operations. On the other hand, MLP-based methods like DuMLP-Pin avoid such constraints but tend to lose critical item-wise structure through global aggregation. To address these limitations, we propose DeviMix (Deviation-based multiple coefficient item Mixer), a novel MLP-based architecture that performs item-wise dynamic transformations. Our approach generates multiple item-mixing coefficients by applying MLPs to cross-deviation vectors computed from all possible item pairs in sets. Extensive experiments on fashion outfit and furniture coordination matching tasks demonstrate that DeviMix consistently outperforms attention-based and global pooling-based baselines, validating the effectiveness of our MLP-based item-wise aggregation using cross-deviation for heterogeneous set matching.

Cite this Paper

BibTeX
@InProceedings{pmlr-v304-hachiya25a, title = {Deviation-based multiple coefficient item mixer for heterogeneous set-to-set matching}, author = {Hachiya, Hirotaka and Yukito, Kajishiro}, booktitle = {Proceedings of the 17th Asian Conference on Machine Learning}, pages = {129--144}, year = {2025}, editor = {Lee, Hung-yi and Liu, Tongliang}, volume = {304}, series = {Proceedings of Machine Learning Research}, month = {09--12 Dec}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v304/main/assets/hachiya25a/hachiya25a.pdf}, url = {https://proceedings.mlr.press/v304/hachiya25a.html}, abstract = {Heterogeneous set-to-set matching tasks such as fashion outfit recommendation, require permutation-invariant and dynamic item-wise transformations to bring compatible sets closer while pushing incompatible ones apart. While attention-based methods satisfy the permutation invariance requirement, they often suffer from convex hull limitations due to their reliance on softmax-based dot-product operations. On the other hand, MLP-based methods like DuMLP-Pin avoid such constraints but tend to lose critical item-wise structure through global aggregation. To address these limitations, we propose DeviMix (Deviation-based multiple coefficient item Mixer), a novel MLP-based architecture that performs item-wise dynamic transformations. Our approach generates multiple item-mixing coefficients by applying MLPs to cross-deviation vectors computed from all possible item pairs in sets. Extensive experiments on fashion outfit and furniture coordination matching tasks demonstrate that DeviMix consistently outperforms attention-based and global pooling-based baselines, validating the effectiveness of our MLP-based item-wise aggregation using cross-deviation for heterogeneous set matching.} }
Endnote
%0 Conference Paper %T Deviation-based multiple coefficient item mixer for heterogeneous set-to-set matching %A Hirotaka Hachiya %A Kajishiro Yukito %B Proceedings of the 17th Asian Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Hung-yi Lee %E Tongliang Liu %F pmlr-v304-hachiya25a %I PMLR %P 129--144 %U https://proceedings.mlr.press/v304/hachiya25a.html %V 304 %X Heterogeneous set-to-set matching tasks such as fashion outfit recommendation, require permutation-invariant and dynamic item-wise transformations to bring compatible sets closer while pushing incompatible ones apart. While attention-based methods satisfy the permutation invariance requirement, they often suffer from convex hull limitations due to their reliance on softmax-based dot-product operations. On the other hand, MLP-based methods like DuMLP-Pin avoid such constraints but tend to lose critical item-wise structure through global aggregation. To address these limitations, we propose DeviMix (Deviation-based multiple coefficient item Mixer), a novel MLP-based architecture that performs item-wise dynamic transformations. Our approach generates multiple item-mixing coefficients by applying MLPs to cross-deviation vectors computed from all possible item pairs in sets. Extensive experiments on fashion outfit and furniture coordination matching tasks demonstrate that DeviMix consistently outperforms attention-based and global pooling-based baselines, validating the effectiveness of our MLP-based item-wise aggregation using cross-deviation for heterogeneous set matching.
APA
Hachiya, H. & Yukito, K.. (2025). Deviation-based multiple coefficient item mixer for heterogeneous set-to-set matching. Proceedings of the 17th Asian Conference on Machine Learning, in Proceedings of Machine Learning Research 304:129-144 Available from https://proceedings.mlr.press/v304/hachiya25a.html.

Related Material