Safe Latent Diffusion: Mitigating Inappropriate Degeneration in Diffusion Models

11/09/2022
by   Patrick Schramowski, et al.
1

Text-conditioned image generation models have recently achieved astonishing results in image quality and text alignment and are consequently employed in a fast-growing number of applications. Since they are highly data-driven, relying on billion-sized datasets randomly scraped from the internet, they also suffer, as we demonstrate, from degenerated and biased human behavior. In turn, they may even reinforce such biases. To help combat these undesired side effects, we present safe latent diffusion (SLD). Specifically, to measure the inappropriate degeneration due to unfiltered and imbalanced training sets, we establish a novel image generation test bed-inappropriate image prompts (I2P)-containing dedicated, real-world image-to-text prompts covering concepts such as nudity and violence. As our exhaustive empirical evaluation demonstrates, the introduced SLD removes and suppresses inappropriate image parts during the diffusion process, with no additional training required and no adverse effect on overall image quality or text alignment.

READ FULL TEXT

page 16

page 17

page 18

page 19

page 21

page 22

page 24

page 25

research
02/07/2023

Fair Diffusion: Instructing Text-to-Image Generation Models on Fairness

Generative AI models have recently achieved astonishing results in quali...
research
05/28/2023

Mitigating Inappropriateness in Image Generation: Can there be Value in Reflecting the World's Ugliness?

Text-conditioned image generation models have recently achieved astonish...
research
07/12/2023

Towards Safe Self-Distillation of Internet-Scale Text-to-Image Diffusion Models

Large-scale image generation models, with impressive quality made possib...
research
09/20/2023

Distilling Adversarial Prompts from Safety Benchmarks: Report for the Adversarial Nibbler Challenge

Text-conditioned image generation models have recently achieved astonish...
research
07/11/2023

TIAM – A Metric for Evaluating Alignment in Text-to-Image Generation

The progress in the generation of synthetic images has made it crucial t...
research
09/22/2022

Implementing and Experimenting with Diffusion Models for Text-to-Image Generation

Taking advantage of the many recent advances in deep learning, text-to-i...
research
01/05/2023

Accuracy and Fidelity Comparison of Luna and DALL-E 2 Diffusion-Based Image Generation Systems

We qualitatively examine the accuracy and fideltiy between two diffusion...

Please sign up or login with your details

Forgot password? Click here to reset