KEST: Kernel Distance Based Efficient Self-Training for Improving Controllable Text Generation

被引:0
|
作者
Feng, Yuxi [1 ]
Yi, Xiaoyuan [2 ]
Lakshmanan, Laks V. S. [1 ]
Xie, Xing [2 ]
机构
[1] Univ British Columbia, Vancouver, BC, Canada
[2] Microsoft Res Asia, Beijing, Peoples R China
来源
PROCEEDINGS OF THE THIRTY-SECOND INTERNATIONAL JOINT CONFERENCE ON ARTIFICIAL INTELLIGENCE, IJCAI 2023 | 2023年
基金
加拿大自然科学与工程研究理事会;
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
Self-training (ST) has come to fruition in language understanding tasks by producing pseudo labels, which reduces the labeling bottleneck of language model fine-tuning. Nevertheless, in facilitating semi-supervised controllable language generation, ST faces two key challenges. First, augmented by self-generated pseudo text, generation models tend to over-exploit the previously learned text distribution, suffering from mode collapse and poor generation diversity. Second, generating pseudo text in each iteration is time-consuming, severely decelerating the training process. In this work, we propose KEST, a novel and efficient self-training framework to handle these problems. KEST utilizes a kernel-based loss, rather than standard cross entropy, to learn from the soft pseudo text produced by a shared non-autoregressive generator. We demonstrate both theoretically and empirically that KEST can benefit from more diverse pseudo text in an efficient manner, which allows not only refining and exploiting the previously fitted distribution but also enhanced exploration towards a larger potential text space, providing a guarantee of improved performance. Experiments on three controllable generation tasks demonstrate that KEST significantly improves control accuracy while maintaining comparable text fluency and generation diversity against several strong baselines.
引用
收藏
页码:5049 / 5057
页数:9
相关论文
共 50 条
  • [41] Controllable Video Generation With Text-Based Instructions
    Koksal, Ali
    Ak, Kenan E.
    Sun, Ying
    Rajan, Deepu
    Lim, Joo Hwee
    IEEE TRANSACTIONS ON MULTIMEDIA, 2024, 26 : 190 - 201
  • [42] The Dialog Must Go On: Improving Visual Dialog via Generative Self-Training
    Kang, Gi-Cheon
    Kim, Sungdong
    Kim, Jin-Hwa
    Kwak, Donghyun
    Zhang, Byoung-Tak
    2023 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR, 2023, : 6746 - 6756
  • [43] Improving Machine Reading Comprehension with Multi-Task Learning and Self-Training
    Ouyang, Jianquan
    Fu, Mengen
    MATHEMATICS, 2022, 10 (03)
  • [44] Improving self-training with density peaks of data and cut edge weight statistic
    Danni Wei
    Youlong Yang
    Haiquan Qiu
    Soft Computing, 2020, 24 : 15595 - 15610
  • [45] Improving emotion recognition in schizophrenia with "VOICES": An on-line prosodic self-training
    Lado-Codesido, Maria
    Mendez Perez, Cristina
    Mateos, Raimundo
    Manuel Olivares, Jose
    Garcia Caballero, Alejandro
    PLOS ONE, 2019, 14 (01):
  • [46] Improving self-training with density peaks of data and cut edge weight statistic
    Wei, Danni
    Yang, Youlong
    Qiu, Haiquan
    SOFT COMPUTING, 2020, 24 (20) : 15595 - 15610
  • [47] Text Classification Using Label Names Only: A Language Model Self-Training Approach
    Meng, Yu
    Zhang, Yunyi
    Huang, Jiaxin
    Xiong, Chenyan
    Ji, Heng
    Zhang, Chao
    Han, Jiawei
    PROCEEDINGS OF THE 2020 CONFERENCE ON EMPIRICAL METHODS IN NATURAL LANGUAGE PROCESSING (EMNLP), 2020, : 9006 - 9017
  • [48] SelfHAR: Improving Human Activity Recognition through Self-training with Unlabeled Data
    Tang C.I.
    Perez-Pozuelo I.
    Spathis D.
    Brage S.
    Wareham N.
    Mascolo C.
    Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies, 2021, 5 (01)
  • [49] SelfHAR: Improving Human Activity Recognition through Self-training with Unlabeled Data
    Tang, Chi Ian
    Perez-Pozuelo, Ignacio
    Spathis, Dimitris
    Brage, Soren
    Wareham, Nick
    Mascolo, Cecilia
    PROCEEDINGS OF THE ACM ON INTERACTIVE MOBILE WEARABLE AND UBIQUITOUS TECHNOLOGIES-IMWUT, 2021, 5 (01):
  • [50] Rank-based self-training for graph convolutional networks
    Guimaraes Pedronette, Daniel Carlos
    Latecki, Longin Jan
    INFORMATION PROCESSING & MANAGEMENT, 2021, 58 (02)