ProtoDiffusion: Classifier-Free Diffusion Guidance with Prototype Learning

Gulcin Baykal, Halil Faruk Karagoz, Taha Binhuraib, Gozde Unal
Proceedings of the 15th Asian Conference on Machine Learning, PMLR 222:106-120, 2024.

Abstract

Diffusion models are generative models that have shown significant advantages compared to other generative models in terms of higher generation quality and more stable training. However, the computational need for training diffusion models is considerably increased. In this work, we incorporate prototype learning into diffusion models to achieve high generation quality faster than the original diffusion model. Instead of randomly initialized class embeddings, we use separately learned class prototypes as the conditioning information to guide the diffusion process. We observe that our method, called ProtoDiffusion, achieves better performance in the early stages of training compared to the baseline method, signifying that using the learned prototypes shortens the training time. We demonstrate the performance of ProtoDiffusion using various datasets and experimental settings, achieving the best performance in shorter times across all settings.

Cite this Paper

BibTeX
@InProceedings{pmlr-v222-baykal24a, title = {{ProtoDiffusion}: {C}lassifier-Free Diffusion Guidance with Prototype Learning}, author = {Baykal, Gulcin and Karagoz, Halil Faruk and Binhuraib, Taha and Unal, Gozde}, booktitle = {Proceedings of the 15th Asian Conference on Machine Learning}, pages = {106--120}, year = {2024}, editor = {Yanıkoğlu, Berrin and Buntine, Wray}, volume = {222}, series = {Proceedings of Machine Learning Research}, month = {11--14 Nov}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v222/baykal24a/baykal24a.pdf}, url = {https://proceedings.mlr.press/v222/baykal24a.html}, abstract = {Diffusion models are generative models that have shown significant advantages compared to other generative models in terms of higher generation quality and more stable training. However, the computational need for training diffusion models is considerably increased. In this work, we incorporate prototype learning into diffusion models to achieve high generation quality faster than the original diffusion model. Instead of randomly initialized class embeddings, we use separately learned class prototypes as the conditioning information to guide the diffusion process. We observe that our method, called ProtoDiffusion, achieves better performance in the early stages of training compared to the baseline method, signifying that using the learned prototypes shortens the training time. We demonstrate the performance of ProtoDiffusion using various datasets and experimental settings, achieving the best performance in shorter times across all settings.} }
Endnote
%0 Conference Paper %T ProtoDiffusion: Classifier-Free Diffusion Guidance with Prototype Learning %A Gulcin Baykal %A Halil Faruk Karagoz %A Taha Binhuraib %A Gozde Unal %B Proceedings of the 15th Asian Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Berrin Yanıkoğlu %E Wray Buntine %F pmlr-v222-baykal24a %I PMLR %P 106--120 %U https://proceedings.mlr.press/v222/baykal24a.html %V 222 %X Diffusion models are generative models that have shown significant advantages compared to other generative models in terms of higher generation quality and more stable training. However, the computational need for training diffusion models is considerably increased. In this work, we incorporate prototype learning into diffusion models to achieve high generation quality faster than the original diffusion model. Instead of randomly initialized class embeddings, we use separately learned class prototypes as the conditioning information to guide the diffusion process. We observe that our method, called ProtoDiffusion, achieves better performance in the early stages of training compared to the baseline method, signifying that using the learned prototypes shortens the training time. We demonstrate the performance of ProtoDiffusion using various datasets and experimental settings, achieving the best performance in shorter times across all settings.
APA
Baykal, G., Karagoz, H.F., Binhuraib, T. & Unal, G.. (2024). ProtoDiffusion: Classifier-Free Diffusion Guidance with Prototype Learning. Proceedings of the 15th Asian Conference on Machine Learning, in Proceedings of Machine Learning Research 222:106-120 Available from https://proceedings.mlr.press/v222/baykal24a.html.

Related Material