中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
首页 机构 成果 学者
  
  1. CAS IR Grid
  2. 计算技术研究所
  3. 中国科学院计算技术研究所
  4. 中国科学院计算技术研究所期刊论文
  5. 英文
General Greedy De-Bias Learning

文献类型：期刊论文

作者Han, Xinzhe3,4; Wang, Shuhui1,4; Su, Chi2; Huang, Qingming3,4; Tian, Qi5
刊名IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
出版日期2023-08-01
卷号45期号:8页码:9789-9805
ISSN号0162-8828
关键词Task analysis Correlation Training Data models Question answering (information retrieval) Visualization Image classification Curriculum learning dataset biases greedy strategy robust learning
DOI10.1109/TPAMI.2023.3240337
英文摘要Neural networks often make predictions relying on the spurious correlations from the datasets rather than the intrinsic properties of the task of interest, facing with sharp degradation on out-of-distribution (OOD) test data. Existing de-bias learning frameworks try to capture specific dataset bias by annotations but they fail to handle complicated OOD scenarios. Others implicitly identify the dataset bias by special design low capability biased models or losses, but they degrade when the training and testing data are from the same distribution. In this paper, we propose a General Greedy De-bias learning framework (GGD), which greedily trains the biased models and base model. The base model is encouraged to focus on examples that are hard to solve with biased models, thus remaining robust against spurious correlations in the test stage. GGD largely improves models' OOD generalization ability on various tasks, but sometimes over-estimates the bias level and degrades on the in-distribution test. We further re-analyze the ensemble process of GGD and introduce the Curriculum Regularization inspired by curriculum learning, which achieves a good trade-off between in-distribution (ID) and out-of-distribution performance. Extensive experiments on image classification, adversarial question answering, and visual question answering demonstrate the effectiveness of our method. GGD can learn a more robust base model under the settings of both task-specific biased models with prior knowledge and self-ensemble biased model without prior knowledge. Codes are available at https://github.com/GeraldHan/GGD.
资助项目National Key R&D Program of China[2018AAA0102000] ; National Natural Science Foundation of China[62022083] ; National Natural Science Foundation of China[62236008] ; National Natural Science Foundation of China[61931008] ; National Natural Science Foundation of China[U21B2038] ; Beijing Nova Program[Z201100006820023]
WOS研究方向Computer Science ; Engineering
语种英语
出版者IEEE COMPUTER SOC
WOS记录号WOS:001022958600034
源URL[http://119.78.100.204/handle/2XEOYT63/21319]  
专题中国科学院计算技术研究所期刊论文_英文
通讯作者Wang, Shuhui
作者单位1.Peng Cheng Lab, Shenzhen 518066, Peoples R China
2.SmartMore, Beijing 100085, Peoples R China
3.Univ Chinese Acad Sci, Sch Comp Sci & Technol, Beijing 101408, Peoples R China
4.Chinese Acad Sci, Inst Comp Technol, Key Lab Intelligent Informat Proc, Beijing 100190, Peoples R China
5.Huawei Cloud & AI, Shenzhen 518129, Guangdong, Peoples R China
推荐引用方式
GB/T 7714		 Han, Xinzhe,Wang, Shuhui,Su, Chi,et al. General Greedy De-Bias Learning[J]. IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE,2023,45(8):9789-9805.
APA Han, Xinzhe,Wang, Shuhui,Su, Chi,Huang, Qingming,&Tian, Qi.(2023).General Greedy De-Bias Learning.IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE,45(8),9789-9805.
MLA Han, Xinzhe,et al."General Greedy De-Bias Learning".IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE 45.8(2023):9789-9805.

入库方式： OAI收割

来源：计算技术研究所

浏览0
下载0
收藏0
其他版本

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。