Learning from Synthetic Data via Provenance-Based Input Gradient Guidance

TL;DR

This paper introduces a provenance-guided input gradient method that enhances model robustness by focusing on target regions during training with synthetic data, reducing reliance on spurious correlations.

Contribution

It proposes a novel framework leveraging provenance information to explicitly guide models toward target regions, improving discrimination and robustness across various tasks.

Findings

Effective in weakly supervised object localization

Improves spatio-temporal action localization accuracy

Enhances image classification robustness

Abstract

Learning methods using synthetic data have attracted attention as an effective approach for increasing the diversity of training data while reducing collection costs, thereby improving the robustness of model discrimination. However, many existing methods improve robustness only indirectly through the diversification of training samples and do not explicitly teach the model which regions in the input space truly contribute to discrimination; consequently, the model may learn spurious correlations caused by synthesis biases and artifacts. Motivated by this limitation, this paper proposes a learning framework that uses provenance information obtained during the training data synthesis process, indicating whether each region in the input space originates from the target object, as an auxiliary supervisory signal to promote the acquisition of representations focused on target regions. Specifically, input gradients are decomposed based on information about target and non-target regions during synthesis, and input gradient guidance is introduced to suppress gradients over non-target regions. This suppresses the model's reliance on non-target regions and directly promotes the learning of discriminative representations for target regions. Experiments demonstrate the effectiveness and generality of the proposed method across multiple tasks and modalities, including weakly supervised object localization, spatio-temporal action localization, and image classification.