Summary of Safegen: Mitigating Sexually Explicit Content Generation in Text-to-image Models, by Xinfeng Li et al.
SafeGen: Mitigating Sexually Explicit Content Generation in Text-to-Image Models
by Xinfeng Li, Yuchen Yang, Jiangyi Deng, Chen Yan, Yanjiao Chen, Xiaoyu Ji, Wenyuan Xu
First submitted to arxiv on: 10 Apr 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Cryptography and Security (cs.CR)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary Medium Difficulty summary: This paper presents SafeGen, a framework to mitigate sexually explicit content generation by text-to-image models in a text-agnostic manner. The existing countermeasures focus on filtering inappropriate inputs and outputs or suppressing improper text embeddings, which can be vulnerable to adversarial prompts. In contrast, SafeGen eliminates explicit visual representations from the model regardless of the text input, making it resistant to such prompts. The framework is evaluated on four datasets and large-scale user studies, outperforming eight state-of-the-art baseline methods with a 99.4% sexual content removal performance. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Low Difficulty summary: This research paper talks about how artificial intelligence models can create images from text descriptions. Some of these models can be tricked into creating inappropriate or explicit content. The authors of this paper want to fix this problem by creating a new way to prevent AI models from generating such content. They call it SafeGen and tested it on many different datasets and with real users. The results show that SafeGen is very effective in preventing the creation of unwanted images while still allowing for high-quality, appropriate pictures. |
