Mitigate Replication and Copying in Diffusion Models with Generalized Caption and Dual Fusion Enhancement

September 13, 2023 ยท Entered Twilight ยท ๐Ÿ› IEEE International Conference on Acoustics, Speech, and Signal Processing

๐Ÿ’ค TWILIGHT: Eternal Rest
Repo abandoned since publication

Repo contents: .idea, README.md, cal_gen_score.py, datasets.py, diff_inference.py, diff_retrieval.py, diff_train.py, dino_vits.py, figure_mitigate_samples.png, generality.py, openai_api_sample.py, overview_dual_fusion.png, utils_ret.py

Authors Chenghao Li, Dake Chen, Yuke Zhang, Peter A. Beerel arXiv ID 2309.07254 Category cs.CV: Computer Vision Cross-listed cs.CR Citations 12 Venue IEEE International Conference on Acoustics, Speech, and Signal Processing Repository https://github.com/HowardLi0816/dual-fusion-diffusion โญ 1 Last Checked 1 month ago
Abstract
While diffusion models demonstrate a remarkable capability for generating high-quality images, their tendency to `replicate' training data raises privacy concerns. Although recent research suggests that this replication may stem from the insufficient generalization of training data captions and duplication of training images, effective mitigation strategies remain elusive. To address this gap, our paper first introduces a generality score that measures the caption generality and employ large language model (LLM) to generalize training captions. Subsequently, we leverage generalized captions and propose a novel dual fusion enhancement approach to mitigate the replication of diffusion models. Our empirical results demonstrate that our proposed methods can significantly reduce replication by 43.5% compared to the original diffusion model while maintaining the diversity and quality of generations. Code is available at https://github.com/HowardLi0816/dual-fusion-diffusion.
Community shame:
Not yet rated
Community Contributions

Found the code? Know the venue? Think something is wrong? Let us know!

๐Ÿ“œ Similar Papers

In the same crypt โ€” Computer Vision