MLSP-P60.14
From Parameters to Prompts: Understanding and Mitigating the Factuality Gap between Fine-Tuned LLMs
Xuan Gong, Hanbo Huang, Yiran Zhang, Shiyu Liang, Shanghai Jiao Tong University, China
Session:
MLSP-P60: Training and Optimization Strategies for Machine Learning Models Poster
Track:
Machine Learning for Signal Processing [ML]
Location:
Poster Area 41
Presentation Time:
Thu, 7 May, 14:00 - 16:00
Presentation
Discussion
Resources
No resources available.
Session MLSP-P60
MLSP-P60.1: Orthogonal Weight Modification Enhances Learning Scalability and Convergence Efficiency without Gradient Backpropagation
Guoqing Ma, Shan Yu, Chinese Academy of Sciences, Institute of Automation, China
MLSP-P60.2: GUIDING EFFICIENT LLM INSTRUCTION-TUNING VIA GRADIENT FLOW MATCHING
Heng Zhang, South China Normal University, China; Weihao Yu, Research Institute of China Telecom Corporate Ltd, China; Yan Gong, Zhejiang University, China; Wenjun Huang, Sun Yat-sen University, China; Hao Zhang, University of Chinese Academy of Sciences, China; Jin Huang, South China Normal University, China
MLSP-P60.3: MORE THAN A SHORTCUT: A HYPERBOLIC APPROACH TO EARLY-EXIT NETWORKS
Swapnil Bhosale, University of Surrey, UK, United Kingdom of Great Britain and Northern Ireland; Cosmin Frateanu, Camilla Clark, Arnoldas Jasonas, Chris Mitchell, Meta, United Kingdom of Great Britain and Northern Ireland; Xiatian Zhu, University of Surrey, United Kingdom of Great Britain and Northern Ireland; Vamsi Krishna Ithapu, Giacomo Ferroni, Cagdas Bilen, Sanjeel Parekh, Meta, United States of America
MLSP-P60.4: MTSearch-R1: Reinforcement Learning for Flexible Multi-Tool Search with Large Language Models
Yiqing Shen, JHU, United States of America; Jing Ke, SJTU, China
MLSP-P60.5: SHARED-WEIGHTS EXTENDER AND GRADIENT VOTING FOR NEURAL NETWORK EXPANSION
NIKOLAS CHATZIS, IOANNIS KORDONIS, NTUA, Greece; EMMANOUIL THEODOSIS, HARVARD UNIVERSITY CAMBRIDGE, Greece; PETROS MARAGOS, NTUA, Greece
MLSP-P60.6: SAM-GT: SAM AS A GENERAL TEACHER ENHANCES MEDICAL IMAGE SEGMENTATION BY DISTILLING ONLY WHAT MATTERS
Zhuolin Li, Xing Wu, Chongqing University, China; Qiuju Deng, Chongqing College of Mobile Communication, China; Peng Wang, Hongqian Wang, The First Affiliated Hospital of Army Medical University, China
MLSP-P60.7: WHY DELETE? JUST MAKE IT NATURAL. MAXIMUM ENTROPY DISTRIBUTION DISTILLATION FOR LARGE LANGUAGE MODELS UNLEARNING
Ruyun Wang, Fuqing Zhu, Xiaodan Zhang, iie.ac.cn, China
MLSP-P60.8: A State-Dependent Markov Diffusion Process for Generative Speech Enhancement
Yasir Iqbal, Tao Zhang, Tianjin University, China; Anjum Iqbal, Dalian University of Technology, China; Xin Zhao, Yanzhang Geng, Tianjin University, China
MLSP-P60.9: FLEXI-LORA WITH INPUT-ADAPTIVE RANKS: EFFICIENT FINETUNING FOR SPEECH AND REASONING TASKS
Zongqian Li, Yixuan Su, Han Zhou, Zihao Fu, Nigel Collier, University of Cambridge, China
MLSP-P60.10: 3D-AWARE SEMANTIC ALIGNMENT: JOINT GLOBAL AND LOCAL MODELING FOR 3D FEW-SHOT ANOMALY DETECTION
Min Huang, Jinxia Zhang, Shixiong Fang, Shenghao Dong, Ziai Zhou, Chaoyang Song, Yang Hu, Southeast University, China
MLSP-P60.11: Deepfake-HMDE: Hierarchical Mixture of Deepfake Experts for Deepfake Detection
Zhifei Ren, Southeast University, China; Jiaming Zhang, Xiaohua Feng, Zhejiang University, China; Yuyuan Li, Hangzhou Dianzi University, China; Chaochao Chen, Zhejiang University, China
MLSP-P60.12: Sparse Gradient Compression for Fine-Tuning Large Language Models
David H. Yang, Mohammad Mohammadi Amiri, Rensselaer Polytechnic Institute, United States of America; Tejaswini Pedapati, Subhajit Chaudhury, Pin-Yu Chen, IBM Research, United States of America
MLSP-P60.13: Investigating Batch Inference in a Sequential Monte Carlo Framework for Neural Networks
Andrew Millard, Joshua Murphy, Peter Green, Simon Maskell, University of Liverpool, United Kingdom of Great Britain and Northern Ireland
MLSP-P60.14: From Parameters to Prompts: Understanding and Mitigating the Factuality Gap between Fine-Tuned LLMs
Xuan Gong, Hanbo Huang, Yiran Zhang, Shiyu Liang, Shanghai Jiao Tong University, China
MLSP-P60.15: STRUCTURED PRUNING VIA MULTI-OBSERVATION ITERATIVE HARD THRESHOLDING
Huoxiang Yang, Bohuai Xiao, Shenzhen University, China; Shuangyan Yi, Shenzhen Institute of Information Technology, China; Binqiang Liu, Shenzhen University, China; Fanyang Meng, Pengcheng Laboratory, China; Wei Liu, Shenzhen University of Information Technology, China; Yongsheng Liang, Shenzhen University, China
MLSP-P60.16: Hierarchical Channel Aggregation with Entropy-Driven Distillation for Federated Segmentation
Shuchang Wang, Yuxuan Zhang, Wei Yang, University of Science and Technology of China, China
Contacts