no code implementations • 7 Dec 2023 • Yanrui Du, Sendong Zhao, Ming Ma, Yuhan Chen, Bing Qin
The jailbreak idea of our method is "Inherent Response Tendency Analysis" which identifies real-world instructions that can inherently induce LLMs to generate affirmation responses and the corresponding jailbreak strategy is "Real-World Instructions-Driven Jailbreak" which involves strategically splicing real-world instructions identified through the above analysis around the malicious instruction.
no code implementations • 7 Oct 2023 • Zexu Sun, Bowei He, Ming Ma, Jiakai Tang, Yuchen Wang, Chen Ma, Dugang Liu
Specifically, our RUAD can more effectively alleviate the feature sensitivity of the uplift model through two customized modules, including a feature selection module with joint multi-label modeling to identify a key subset from the input features and an adversarial feature desensitization module using adversarial training and soft interpolation operations to enhance the robustness of the model against this selected subset of features.
1 code implementation • 8 Sep 2023 • Yanrui Du, Sendong Zhao, MuZhen Cai, Ming Ma, Danyang Zhao, Jiawei Cao, Bing Qin
We conduct several experiments to analyze the dual logic ability of LLMs by examining the consistency of the stance in responses to paired questions about the same fact.
no code implementations • 8 Sep 2022 • Jiapeng Wang, Ming Ma, Zhenhua Yu
In this paper, we propose a simple yet effective differentiable network pruning method CWP based on instance complexity weighted filter importance scores.
no code implementations • 20 Dec 2021 • Bin Zhang, Yang Wu, Xiaojing Zhang, Ming Ma
In the current salient object detection network, the most popular method is using U-shape structure.
no code implementations • 14 Jul 2021 • Zongxia Liang, Yang Liu, Ming Ma, Rahul Pothi Vinoth
We derive a unified closed-form formula of the optimal portfolio, which is a four-term division.
no code implementations • 23 Nov 2019 • Charles Huang, Masoud Badiei, Hyunseok Seo, Ming Ma, Xiaokun Liang, Dante Capaldi, Michael Gensheimer, Lei Xing
Whereas supervised segmentation methods only automate the segmentation process for a select few number of OARs, we demonstrate that our methods can achieve similar performance for OARs of interest, while also providing segmentations for every other OAR on the provided atlas.
no code implementations • 25 Sep 2019 • Ming Ma, Stephan Chalup, Fayeem Aziz, Yang Liu, Defu Cheng, Zhijian Zhou
In this paper, we generalize WSL to counting machines that apply convolutional neural networks (CNN) and density maps for counting.
no code implementations • ICCV 2017 • Xiaopeng Zheng, Chengfeng Wen, Na lei, Ming Ma, Xianfeng GU
This work introduces a novel surface registration method based on foliation.