MLSP-L22: Safety and Robustness in Machine Learning
Oral
Thu, 7 May, 14:00 - 16:00
Location: Room 112
Session Type: Oral
Track: Machine Learning for Signal Processing [ML]
Click the to view the manuscript on IEEE Xplore Open Preview
Thu, 7 May, 14:00 - 14:20

MLSP-L22.1: Planning-oriented Adversarial Attack against End-to-End Autonomous Driving Systems
Hao Tan, Harbin Institute of Technology (Shenzhen), China; Ruonan Li, Peng Cheng Laboratory, China; Junjian Zhang, National University of Defense Technology, China; Huan Zhang, Harbin Institute of Technology (Shenzhen), China; Di Shao, Peng Cheng Laboratory, China; Zhaoquan Gu, Harbin Institute of Technology (Shenzhen), China
Thu, 7 May, 14:20 - 14:40

MLSP-L22.2: SAFEGEN: SCULPTING REPRESENTATION SPACE FOR SAFER AND SMARTER LLMS
Heng Zhang, South China Normal University, China; Weihao Yu, Research Institute of China Telecom Corporate Ltd, China; Yilei Yuan, University of Michigan, United States of America; Yan Gong, Peking University, China; Zumeng Zhang, Yunnan University, China; Jin Huang, South China Normal University, China
Thu, 7 May, 14:40 - 15:00

MLSP-L22.3: Safety Alignment Should Be Made More Than Just A Few Attention Heads
Chao Huang, Zefeng Zhang, Juwei Yue, Jiawei Sheng, Quangang Li, Chuang Zhang, Tingwen Liu, Institute of Information Engineering, Chinese Academy of Sciences, Beijing China, China
Thu, 7 May, 15:00 - 15:20

MLSP-L22.4: ROBUST UNCERTAINTY ESTIMATION UNDER DISTRIBUTION SHIFT VIA DIFFERENCE RECONSTRUCTION
Xinran Xu, Li Rong Wang, Xiuyi Fan, Nanyang Technological University, Singapore, Singapore
Thu, 7 May, 15:20 - 15:40

MLSP-L22.5: INTACT: INDUCING NOISE TOLERANCE THROUGH ADVERSARIAL CURRICULUM TRAINING FOR LIDAR-BASED SAFETY-CRITICAL PERCEPTION AND AUTONOMY
Nastaran Darabi, Sina Tayebati, Divake Kumar, Theja Tulabandhula, Amit Trivedi, University of Illinois Chicago, United States of America
Thu, 7 May, 15:40 - 16:00

MLSP-L22.6: RoCo: Robust Code for Fast and Effective Proactive Defense against Voice Cloning Attack
Seungmin Kim, Dain Kim, Sohee Park, Daeseon Choi, Soongsil University, Korea, Republic of