2022/01/06 · We propose an Information-Theoretic Shortcut Avoidance~(ITSA) approach to automatically restrict shortcut-related information from being encoded into the ...

In this work, we show that avoiding shortcut learning can effectively enhance the robustness of the stereo matching networks and enables a model to generalize ...

This is the official code of the work ITSA: An Information-Theoretic Approach to Automatic Shortcut Avoidance and Domain Generalization in Stereo Matching ...

In this work, we show that avoiding shortcut learning can effectively enhance the robustness of the stereo matching networks and enables a model to generalize ...

For example, Chuah et al. [19] proposed an information theoretical shortcut avoidance method to construct robust features, leading to improved generalization.

This paper proposes an Information-Theoretic Shortcut Avoidance (ITSA) approach to automatically restrict shortcut-related information from being encoded ...

We propose an Information-Theoretic Shortcut Avoidance (ITSA) approach to automatically restrict shortcut-related information from being encoded into the ...

ITSA: An Information-Theoretic Approach to Automatic Shortcut Avoidance and Domain Generalization in Stereo Matching Networks. Supplementary Materials. WeiQin ...

2023/04/20 · We propose an Information-Theoretic Shortcut Avoidance (ITSA) approach to automatically restrict shortcut-related information from being encoded into the ...

Our results show that the proposed method can effectively improve S2R generalization in multiple distinct dense prediction tasks, such as stereo matching, ...