arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2406.14855cs.CVcs.CR
收藏

Six-CD:良性文本到图像扩散模型的概念移除基准测试

Six-CD: Benchmarking Concept Removals for Benign Text-to-image Diffusion Models

Michigan State University(密歇根州立大学);Sony AI(索尼人工智能)

Jie Ren, Kangrui Chen, Yingqian Cui, Shenglai Zeng, Hui Liu, Yue Xing, Jiliang Tang, Lingjuan Lyu

AI总结:

针对文本到图像扩散模型被滥用的风险,提出Six-CD数据集和新评估指标,基准测试概念移除方法,解决现有研究在数据集、提示有效性和评估完整性上的挑战,为该领域提供有价值见解。

摘要:

文本到图像(T2I)扩散模型在生成与文本提示高度对应的图像方面展现出卓越能力。然而,T2I扩散模型的发展带来了重大风险,模型可能被滥用于恶意目的,例如生成包含暴力或裸体的图像,或在不当语境中创建公众人物的未授权肖像。为缓解这些风险,概念移除方法已被提出,旨在修改扩散模型以防止生成恶意和不需要的概念。尽管有这些努力,现有研究仍面临若干挑战:(1)缺乏在综合数据集上的一致比较;(2)针对有害和裸体概念的提示效果不佳;(3)忽视了对包含恶意概念的提示中良性部分生成能力的评估。为解决这些空白,我们提出通过引入新数据集Six-CD和新评估指标来基准测试概念移除方法。在此基准中,我们对概念移除进行了全面评估,实验观察和讨论为该领域提供了有价值的见解。

英文摘要:

Text-to-image (T2I) diffusion models have shown exceptional capabilities in generating images that closely correspond to textual prompts. However, the advancement of T2I diffusion models presents significant risks, as the models could be exploited for malicious purposes, such as generating images with violence or nudity, or creating unauthorized portraits of public figures in inappropriate contexts. To mitigate these risks, concept removal methods have been proposed. These methods aim to modify diffusion models to prevent the generation of malicious and unwanted concepts. Despite these efforts, existing research faces several challenges: (1) a lack of consistent comparisons on a comprehensive dataset, (2) ineffective prompts in harmful and nudity concepts, (3) overlooked evaluation of the ability to generate the benign part within prompts containing malicious concepts. To address these gaps, we propose to benchmark the concept removal methods by introducing a new dataset, Six-CD, along with a novel evaluation metric. In this benchmark, we conduct a thorough evaluation of concept removals, with the experimental observations and discussions offering valuable insights in the field.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑