November 02, 2019 - Seoul, Korea
Deep neural networks have shown remarkable success in many computer vision tasks, but current methods typically rely on massive amounts of labeled training data to achieve high performance. Collecting and annotating such large training datasets is costly, time-consuming, and in many cases infeasible, as for certain tasks only a few or no examples at all may be available. In this tutorial, we address the problem of visual learning with limited labeled data. We plan to focus on state-of-the-art techniques to tackle this problem, including meta-learning and metric learning approaches for few-shot classification, hallucination-based techniques for sample synthesis, recent methods for domain adaptation, and applications in computer vision tasks such as video understanding, object detection and instance segmentation. The organizers will share their extensive experience on this topic, and provide links to resources such as relevant datasets and source code.