Adaptive Shortcut Debiasing for Online Continual Learning

We propose a novel framework DropTop that suppresses the shortcut bias in online continual learning (OCL) while being adaptive to the varying degree of the shortcut bias incurred by continuously changing environment. By the observed high-attention property of the shortcut bias, highly-activated features are considered candidates for debiasing. More importantly, resolving the limitation of the online environment where prior knowledge and auxiliary data are not ready, two novel techniques---feature map fusion and adaptive intensity shifting---enable us to automatically determine the appropriate level and proportion of the candidate shortcut features to be dropped. Extensive experiments on five benchmark datasets demonstrate that, when combined with various OCL algorithms, DropTop increases the average accuracy by up to 10.4% and decreases the forgetting by up to 63.2%.

Authors

• Doyoung Kim KAIST, Daejeon, Republic of Korea
• Dongmin Park KAIST, Daejeon, Republic of Korea
• Yooju Shin Korea Advanced Institute of Science and Technology
• Jihwan Bang KAIST, Daejeon, Republic of Korea
• Hwanjun Song KAIST, Daejeon, Republic of Korea
• Jae-Gil Lee Korea Advanced Institute of Science and Technology

Keywords

• ML: Life-Long and Continual Learning
• ML: Time-Series/Data Streams
• ML: Transfer, Domain Adaptation, Multi-Task Learning