ASPS-L2: resource-efficient and generative AI systems
Oral
Wed, 6 May, 14:00 - 16:00
Location: Room 116
Session Type: Oral
Track: Applied Signal Processing Systems [AS]
Click the to view the manuscript on IEEE Xplore Open Preview
Wed, 6 May, 14:00 - 14:20

ASPS-L2.1: Alternating Balancing Sums for Accurate Low-Power Dot Products
Vikas Natesh, H.T. Kung, Harvard University, United States of America
Wed, 6 May, 14:20 - 14:40

ASPS-L2.2: ADAPTING WHISPER FOR PADDING-FREE INFERENCE USING AN ENCODER ATTENTION MASK AND KNOWLEDGE DISTILLATION
Amin Abdaoui, Abdelmoumene Boumadane, Trideba PADHI, Nancy BERTIN, Nathan SOUVIRAA LABASTIE, Serge LE HUITOUZE, Kyu Jeong Han, Oracle AI, France
Wed, 6 May, 14:40 - 15:00

ASPS-L2.3: EXPERIENCE-DRIVEN DYNAMIC EXITS FOR LLMS WITH REINFORCEMENT LEARNING
Yanyu Zhu, Hoilam Pao, Hai-Tao Zheng, Tsinghua Shenzhen International Graduate School, China; Niu Hu, Wei Guo, Huawei, China; Shaoxiong Zhan, Tsinghua Shenzhen International Graduate School, China; Boyu Lai, Northwestern Polytechnical University, China; Zitai Wang, Yongqin Zeng, Tsinghua Shenzhen International Graduate School, China
Wed, 6 May, 15:00 - 15:20

ASPS-L2.4: SEMANTICACHE: EFFICIENT KV CACHE COMPRESSION VIA SEMANTIC CHUNKING AND CLUSTERED MERGING
Shunlong Wu, Hai Lin, Shaoshen Chen, Tingwei Lu, Yongqin Zeng, Shaoxiong Zhan, Hai-Tao Zheng, Tsinghua University, China; Hong-Gee Kim, Seoul National University, Korea, Republic of
Wed, 6 May, 15:20 - 15:40

ASPS-L2.5: ITERATIVE REDUNDANCY-BASED HEAD PRUNING FOR EFFICIENT SELF-SUPERVISED SPEECH RECOGNITION MODELS
Sang-Heon Jeong, Dong-Hyun Kim, Joon-Hyuk Chang, Hanyang University, Korea, Republic of
Wed, 6 May, 15:40 - 16:00

ASPS-L2.6: COMNET: A COMPLEMENTARY PROTOTYPES-GUIDED RECONSTRUCTION FRAMEWORK FOR MULTI-CLASS ANOMALY DETECTION
Xiaodong Wang, Xiaolong Huang, Fei Yan, Hua Guo, Zhiqiang Zeng, Junwen Lu, Xiamen University of Technology, China