https://scholars.lib.ntu.edu.tw/handle/123456789/642021
標題: | Towards Large Certified Radius in Randomized Smoothing Using Quasiconcave Optimization | 作者: | Kung, Bo Han SHANG-TSE CHEN |
公開日期: | 25-三月-2024 | 卷: | 38 | 期: | 19 | 來源出版物: | Proceedings of the AAAI Conference on Artificial Intelligence | 摘要: | Randomized smoothing is currently the state-of-the-art method that provides certified robustness for deep neural networks. However, due to its excessively conservative nature, this method of incomplete verification often cannot achieve an adequate certified radius on real-world datasets. One way to obtain a larger certified radius is to use an input-specific algorithm instead of using a fixed Gaussian filter for all data points. Several methods based on this idea have been proposed, but they either suffer from high computational costs or gain marginal improvement in certified radius. In this work, we show that by exploiting the quasiconvex problem structure, we can find the optimal certified radii for most data points with slight computational overhead. This observation leads to an efficient and effective input-specific randomized smoothing algorithm. We conduct extensive experiments and empirical analysis on CIFAR-10 and ImageNet. The results show that the proposed method significantly enhances the certified radii with low computational overhead. |
URI: | https://scholars.lib.ntu.edu.tw/handle/123456789/642021 | ISSN: | 21595399 | DOI: | 10.1609/aaai.v38i19.30123 |
顯示於: | 資訊工程學系 |
在 IR 系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。