Don’t be so Negative! Score-based Generative Modeling with Oracle-assisted Guidance

Saeid Naderiparizi, Xiaoxuan Liang, Setareh Cohan, Berend Zwartsenberg, Frank Wood
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:37164-37187, 2024.

Abstract

Score-based diffusion models are a powerful class of generative models, widely utilized across diverse domains. Despite significant advancements in large-scale tasks such as text-to-image generation, their application to constrained domains has received considerably less attention. This work addresses model learning in a setting where, in addition to the training dataset, there further exists side-information in the form of an oracle that can label samples as being outside the support of the true data generating distribution. Specifically we develop a new denoising diffusion probabilistic modeling methodology, Gen-neG, that leverages this additional side-information. Gen-neG builds on classifier guidance in diffusion models to guide the generation process towards the positive support region indicated by the oracle. We empirically establish the utility of Gen-neG in applications including collision avoidance in self-driving simulators and safety-guarded human motion generation.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-naderiparizi24a, title = {Don’t be so Negative! {S}core-based Generative Modeling with Oracle-assisted Guidance}, author = {Naderiparizi, Saeid and Liang, Xiaoxuan and Cohan, Setareh and Zwartsenberg, Berend and Wood, Frank}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {37164--37187}, 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/naderiparizi24a/naderiparizi24a.pdf}, url = {https://proceedings.mlr.press/v235/naderiparizi24a.html}, abstract = {Score-based diffusion models are a powerful class of generative models, widely utilized across diverse domains. Despite significant advancements in large-scale tasks such as text-to-image generation, their application to constrained domains has received considerably less attention. This work addresses model learning in a setting where, in addition to the training dataset, there further exists side-information in the form of an oracle that can label samples as being outside the support of the true data generating distribution. Specifically we develop a new denoising diffusion probabilistic modeling methodology, Gen-neG, that leverages this additional side-information. Gen-neG builds on classifier guidance in diffusion models to guide the generation process towards the positive support region indicated by the oracle. We empirically establish the utility of Gen-neG in applications including collision avoidance in self-driving simulators and safety-guarded human motion generation.} }
Endnote
%0 Conference Paper %T Don’t be so Negative! Score-based Generative Modeling with Oracle-assisted Guidance %A Saeid Naderiparizi %A Xiaoxuan Liang %A Setareh Cohan %A Berend Zwartsenberg %A Frank Wood %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-naderiparizi24a %I PMLR %P 37164--37187 %U https://proceedings.mlr.press/v235/naderiparizi24a.html %V 235 %X Score-based diffusion models are a powerful class of generative models, widely utilized across diverse domains. Despite significant advancements in large-scale tasks such as text-to-image generation, their application to constrained domains has received considerably less attention. This work addresses model learning in a setting where, in addition to the training dataset, there further exists side-information in the form of an oracle that can label samples as being outside the support of the true data generating distribution. Specifically we develop a new denoising diffusion probabilistic modeling methodology, Gen-neG, that leverages this additional side-information. Gen-neG builds on classifier guidance in diffusion models to guide the generation process towards the positive support region indicated by the oracle. We empirically establish the utility of Gen-neG in applications including collision avoidance in self-driving simulators and safety-guarded human motion generation.
APA
Naderiparizi, S., Liang, X., Cohan, S., Zwartsenberg, B. & Wood, F.. (2024). Don’t be so Negative! Score-based Generative Modeling with Oracle-assisted Guidance. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:37164-37187 Available from https://proceedings.mlr.press/v235/naderiparizi24a.html.

Related Material