Timezone: »

 
The Journey, Not the Destination: How Data Guides Diffusion Models
Kristian Georgiev · Joshua Vendrow · Hadi Salman · Sung Min (Sam) Park · Aleksander Madry

@
in Challenges in Deployable Generative AI »
Event URL: https://openreview.net/forum?id=9hK9NbUAex »

Diffusion-based generative models can synthesize photo-realistic images of unprecedented quality and diversity. However, attributing these images back to the training data---that is, identifying specific training examples which caused the images to be generated---remains challenging. In this paper, we propose a framework that: i) formalizes data attribution in the context of diffusion models, and ii) provides a method for computing attributions efficiently. By applying our framework to CIFAR-10 and MS COCO, we uncover visually compelling attributions, which we validate through counterfactual analysis.

Keywords: Privacy · Data Valuation · memorization · Diffusion Models · influence estimation · data attribution

Author Information

Kristian Georgiev (MIT)
Joshua Vendrow (Massachusetts Institute of Technology)
Hadi Salman (OpenAI / MIT)
Sung Min (Sam) Park (MIT)
Aleksander Madry (MIT)

More from the Same Authors