RoSe22
* *Robustness in Sequential Data
* Analysis and Extensions of Adversarial Training for Video Classification
* CENet: Consolidation-and-Exploration Network for Continuous Domain Adaptation
* Continual Active Adaptation to Evolving Distributional Shifts
* Pose-based Contrastive Learning for Domain Agnostic Activity Representations
* Tragedy Plus Time: Capturing Unintended Human Activities from Weakly-labeled Videos