Improving the trustworthiness of image classification models by utilizing bounding-box annotations

Dharma KC, Chicheng Zhang

We study utilizing auxiliary information in training data to improve the trustworthiness of machine learning models. Specifically, in the context of image classification, we propos…