Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2505.03338 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909967768879104 |
|---|---|
| author | Reissinger, Lena Li, Yuanyuan Haensch, Anna-Carolina Sarna, Neeraj |
| author_facet | Reissinger, Lena Li, Yuanyuan Haensch, Anna-Carolina Sarna, Neeraj |
| contents | Visual Generative AI models have demonstrated remarkable capability in generating high-quality images from user inputs like text prompts. However, because these models have billions of parameters, they risk memorizing certain parts of the training data and reproducing the memorized content. Memorization often raises concerns about safety of such models -- usually involving intellectual property (IP) infringement risk -- and deters their large scale adoption. In this paper, we evaluate the effectiveness of prompt engineering techniques in reducing memorization risk in image generation. Our findings demonstrate the effectiveness of prompt engineering in reducing the similarity between generated images and the training data of diffusion models, while maintaining relevance and aestheticity of the generated output. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_03338 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Safer Prompts: Reducing Risks from Memorization in Visual Generative AI Reissinger, Lena Li, Yuanyuan Haensch, Anna-Carolina Sarna, Neeraj Computer Vision and Pattern Recognition Artificial Intelligence Computers and Society Visual Generative AI models have demonstrated remarkable capability in generating high-quality images from user inputs like text prompts. However, because these models have billions of parameters, they risk memorizing certain parts of the training data and reproducing the memorized content. Memorization often raises concerns about safety of such models -- usually involving intellectual property (IP) infringement risk -- and deters their large scale adoption. In this paper, we evaluate the effectiveness of prompt engineering techniques in reducing memorization risk in image generation. Our findings demonstrate the effectiveness of prompt engineering in reducing the similarity between generated images and the training data of diffusion models, while maintaining relevance and aestheticity of the generated output. |
| title | Safer Prompts: Reducing Risks from Memorization in Visual Generative AI |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Computers and Society |
| url | https://arxiv.org/abs/2505.03338 |