Safeguard Text-to-Image Diffusion Models with Human Feedback Inversion
This addresses societal concerns for AI ethics by improving safety in text-to-image models, though it is incremental as it builds on existing techniques.
The paper tackles the problem of text-to-image diffusion models generating harmful or copyrighted content by proposing Human Feedback Inversion (HFI), which uses human feedback to guide mitigation, resulting in significantly reduced objectionable content while preserving image quality.
This paper addresses the societal concerns arising from large-scale text-to-image diffusion models for generating potentially harmful or copyrighted content. Existing models rely heavily on internet-crawled data, wherein problematic concepts persist due to incomplete filtration processes. While previous approaches somewhat alleviate the issue, they often rely on text-specified concepts, introducing challenges in accurately capturing nuanced concepts and aligning model knowledge with human understandings. In response, we propose a framework named Human Feedback Inversion (HFI), where human feedback on model-generated images is condensed into textual tokens guiding the mitigation or removal of problematic images. The proposed framework can be built upon existing techniques for the same purpose, enhancing their alignment with human judgment. By doing so, we simplify the training objective with a self-distillation-based technique, providing a strong baseline for concept removal. Our experimental results demonstrate our framework significantly reduces objectionable content generation while preserving image quality, contributing to the ethical deployment of AI in the public sphere.