Learning invariant representations via contrastive learning has seen state-of-the-art
performance in domain generalization (DG). Despite such success, in this paper, we find …

Monitoring the training of neural networks is essential for identifying potential data
anomalies, enabling timely interventions and conserving significant computational …

In this work, we propose a general framework called Concept-Monitor to help demystify the
black-box DNN training processes automatically using a novel unified embedding space …

Despite the successes of deep neural networks (DNNs) on a broad range of tasks little has
been understood of why and how they achieve such victories due to their complex …