TY - GEN
T1 - Joint Input and Output Coordination for Class-Incremental Learning
AU - Wang, Shuai
AU - Zhan, Yibing
AU - Luo, Yong
AU - Hu, Han
AU - Yu, Wei
AU - Wen, Yonggang
AU - Tao, Dacheng
N1 - Publisher Copyright:
© 2024 International Joint Conferences on Artificial Intelligence. All rights reserved.
PY - 2024
Y1 - 2024
N2 - Incremental learning is nontrivial due to severe catastrophic forgetting. Although storing a small amount of data on old tasks during incremental learning is a feasible solution, current strategies still do not 1) adequately address the class bias problem, and 2) alleviate the mutual interference between new and old tasks, and 3) consider the problem of class bias within tasks. This motivates us to propose a joint input and output coordination (JIOC) mechanism to address these issues. This mechanism assigns different weights to different categories of data according to the gradient of the output score, and uses knowledge distillation (KD) to reduce the mutual interference between the outputs of old and new tasks. The proposed mechanism is general and flexible, and can be incorporated into different incremental learning approaches that use memory storage. Extensive experiments show that our mechanism can significantly improve their performance.
AB - Incremental learning is nontrivial due to severe catastrophic forgetting. Although storing a small amount of data on old tasks during incremental learning is a feasible solution, current strategies still do not 1) adequately address the class bias problem, and 2) alleviate the mutual interference between new and old tasks, and 3) consider the problem of class bias within tasks. This motivates us to propose a joint input and output coordination (JIOC) mechanism to address these issues. This mechanism assigns different weights to different categories of data according to the gradient of the output score, and uses knowledge distillation (KD) to reduce the mutual interference between the outputs of old and new tasks. The proposed mechanism is general and flexible, and can be incorporated into different incremental learning approaches that use memory storage. Extensive experiments show that our mechanism can significantly improve their performance.
UR - http://www.scopus.com/inward/record.url?scp=85204295190&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:85204295190
T3 - IJCAI International Joint Conference on Artificial Intelligence
SP - 5108
EP - 5116
BT - Proceedings of the 33rd International Joint Conference on Artificial Intelligence, IJCAI 2024
A2 - Larson, Kate
PB - International Joint Conferences on Artificial Intelligence
T2 - 33rd International Joint Conference on Artificial Intelligence, IJCAI 2024
Y2 - 3 August 2024 through 9 August 2024
ER -