Task-specific Compression for Multi-task Language Models using Attribution-based Pruning

被引:0
|
作者
Yang, Nakyeong [1 ]
Jang, Yunah [1 ]
Lee, Hwanhee [2 ]
Jung, Seohyeong [3 ]
Jung, Kyomin [1 ]
机构
[1] Seoul Natl Univ, Seoul, South Korea
[2] Chung Ang Univ, Seoul, South Korea
[3] Hyundai Motor Grp & 42dot Inc, Seoul, South Korea
关键词
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Multi-task language models show outstanding performance for various natural language understanding tasks with only a single model. However, these language models utilize an unnecessarily large number of model parameters, even when used only for a specific task. This paper proposes a novel training-free compression method for multi-task language models using a pruning method. Specifically, we use an attribution method to determine which neurons are essential for performing a specific task. We task-specifically prune unimportant neurons and leave only task-specific parameters. Furthermore, we extend our method to be applicable in low-resource and unsupervised settings. Since our compression method is training-free, it uses few computing resources and does not destroy the pre-trained knowledge of language models. Experimental results on the six widely-used datasets show that our proposed pruning method significantly outperforms baseline pruning methods. In addition, we demonstrate that our method preserves performance even in an unseen domain setting.
引用
收藏
页码:594 / 604
页数:11
相关论文
共 50 条
  • [31] A Compression-Based Dissimilarity Measure for Multi-task Clustering
    Nguyen Huy Thach
    Shao, Hao
    Tong, Bin
    Suzuki, Einoshin
    FOUNDATIONS OF INTELLIGENT SYSTEMS, 2011, 6804 : 123 - 132
  • [32] Scalable Privacy in Multi-Task Image Compression
    Alvar, Saeed Ranjbar
    Bajic, Ivan, V
    2021 INTERNATIONAL CONFERENCE ON VISUAL COMMUNICATIONS AND IMAGE PROCESSING (VCIP), 2021,
  • [33] MULTIPLE AND TASK-SPECIFIC MODELS IN ASSESSMENT OF TEACHING
    SHARPLES, D
    WOODMAN, PF
    DURHAM RESEARCH REVIEW, 1972, 6 (28): : 627 - 634
  • [34] Task-specific internal models for kinematic transformations
    Tong, C
    Flanagan, JR
    JOURNAL OF NEUROPHYSIOLOGY, 2003, 90 (02) : 578 - 585
  • [35] Task-specific adaptation of speech recognition models
    Sankar, A
    Kannan, A
    Shahshahani, B
    Jackson, E
    ASRU 2001: IEEE WORKSHOP ON AUTOMATIC SPEECH RECOGNITION AND UNDERSTANDING, CONFERENCE PROCEEDINGS, 2001, : 433 - 436
  • [36] VLUE: A Multi-Task Benchmark for Evaluating Vision-Language Models
    Zhou, Wangchunshu
    Zeng, Yan
    Diao, Shizhe
    Zhang, Xinsong
    INTERNATIONAL CONFERENCE ON MACHINE LEARNING, VOL 162, 2022,
  • [37] Deep Reinforcement Learning Based Multi-Task Automated Channel Pruning for DNNs
    Ma, Xiaodong
    Fang, Weiwei
    2023 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, IJCNN, 2023,
  • [38] Increasing Transaction Fraud Prediction Ability by Using Multi-Task Learning and Pruning
    Ozcan, Faik
    Genc, Yakup
    2022 30TH SIGNAL PROCESSING AND COMMUNICATIONS APPLICATIONS CONFERENCE, SIU, 2022,
  • [39] Object detection network pruning with multi-task information fusion
    Shengming Li
    Linsong Xue
    Lin Feng
    Yifan Wang
    Dong Wang
    World Wide Web, 2022, 25 : 1667 - 1683
  • [40] A Multi-Task Neural Approach for Emotion Attribution, Classification, and Summarization
    Tu, Guoyun
    Fu, Yanwei
    Li, Boyang
    Gao, Jiarui
    Jiang, Yu-Gang
    Xue, Xiangyang
    IEEE TRANSACTIONS ON MULTIMEDIA, 2020, 22 (01) : 148 - 159