COMI: Correct and Mitigate Shortcut Learning Behavior in Deep Neural Networks
Association for Computing Machinery (ACM) via YouTube
Overview
Explore a 13-minute conference talk from the Association for Computing Machinery (ACM) that delves into the critical issue of shortcut learning in deep neural networks. Learn about COMI, a novel approach designed to correct and mitigate this problematic behavior. Discover how authors Lili Zhao, Qi Liu, Linan Yue, Wei Chen, Liyi Chen, Ruijun Sun, and Chao Song address the challenges posed by shortcut learning and present their innovative solution. Gain insights into the implications of this research for improving the robustness and reliability of deep learning models across various applications.
Syllabus
SIGIR 2024 M1.4 [fp] COMI: COrrect and MItigate Shortcut Learning Behavior in Deep Neural Networks
Taught by
Association for Computing Machinery (ACM)