Embedding an Ethical Mind: Aligning Text-to-Image Synthesis via Lightweight Value Optimization
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866913548950568960 |
|---|---|
| author | Wang, Xingqi Yi, Xiaoyuan Xie, Xing Jia, Jia |
| author_facet | Wang, Xingqi Yi, Xiaoyuan Xie, Xing Jia, Jia |
| contents | Recent advancements in diffusion models trained on large-scale data have enabled the generation of indistinguishable human-level images, yet they often produce harmful content misaligned with human values, e.g., social bias, and offensive content. Despite extensive research on Large Language Models (LLMs), the challenge of Text-to-Image (T2I) model alignment remains largely unexplored. Addressing this problem, we propose LiVO (Lightweight Value Optimization), a novel lightweight method for aligning T2I models with human values. LiVO only optimizes a plug-and-play value encoder to integrate a specified value principle with the input prompt, allowing the control of generated images over both semantics and values. Specifically, we design a diffusion model-tailored preference optimization loss, which theoretically approximates the Bradley-Terry model used in LLM alignment but provides a more flexible trade-off between image quality and value conformity. To optimize the value encoder, we also develop a framework to automatically construct a text-image preference dataset of 86k (prompt, aligned image, violating image, value principle) samples. Without updating most model parameters and through adaptive value selection from the input prompt, LiVO significantly reduces harmful outputs and achieves faster convergence, surpassing several strong baselines and taking an initial step towards ethically aligned T2I models. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_12700 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Embedding an Ethical Mind: Aligning Text-to-Image Synthesis via Lightweight Value Optimization Wang, Xingqi Yi, Xiaoyuan Xie, Xing Jia, Jia Computer Vision and Pattern Recognition Artificial Intelligence Computers and Society Machine Learning Multimedia Recent advancements in diffusion models trained on large-scale data have enabled the generation of indistinguishable human-level images, yet they often produce harmful content misaligned with human values, e.g., social bias, and offensive content. Despite extensive research on Large Language Models (LLMs), the challenge of Text-to-Image (T2I) model alignment remains largely unexplored. Addressing this problem, we propose LiVO (Lightweight Value Optimization), a novel lightweight method for aligning T2I models with human values. LiVO only optimizes a plug-and-play value encoder to integrate a specified value principle with the input prompt, allowing the control of generated images over both semantics and values. Specifically, we design a diffusion model-tailored preference optimization loss, which theoretically approximates the Bradley-Terry model used in LLM alignment but provides a more flexible trade-off between image quality and value conformity. To optimize the value encoder, we also develop a framework to automatically construct a text-image preference dataset of 86k (prompt, aligned image, violating image, value principle) samples. Without updating most model parameters and through adaptive value selection from the input prompt, LiVO significantly reduces harmful outputs and achieves faster convergence, surpassing several strong baselines and taking an initial step towards ethically aligned T2I models. |
| title | Embedding an Ethical Mind: Aligning Text-to-Image Synthesis via Lightweight Value Optimization |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Computers and Society Machine Learning Multimedia |
| url | https://arxiv.org/abs/2410.12700 |