CVPR 2020 continual learning in computer vision competition: Approaches, results, current challenges and future directions

In the last few years, we have witnessed a renewed and fast-growing interest in continual learning with deep neural networks shared objective of making current AI systems more adaptive, efficient autonomous. However, despite significant undoubted progress field addressing issue catastrophic forgetting, benchmarking different approaches is difficult task by itself. fact, given proliferation settings, training evaluation protocols, metrics nomenclature, it often tricky to properly characterize algorithm, relate other solutions gauge its real-world applicability. The first Continual Learning Computer Vision challenge held at CVPR 2020 has been one opportunities evaluate algorithms on common hardware large set 3 settings based realistic CORe50 video benchmark. this paper, report main results competition, which counted than 79 teams registered 11 finalists. We also summarize winning approaches, challenges future research directions.