Mid-Level Visual Representations for Improving Generalization and Sample Efficiency of Visuomotor Policies
Description: Taskonomy: Disentangling Task Transfer Learning, CVPR 2018 (Best Paper). Stanford, UC Berkeley. We propose a fully computational approach for modeling the structure in the space of visual tasks.
vision (2957) computer vision (853) berkeley (643) stanford (273) tasks (270) cvpr (20) transfer learning (14) meta-learning (8) metalearning (4) taskonomy (1)
Visit the live API to find a supervision-efficient transfer strategy for your specified arguments.
Upload an image and see live results for 20 distinct vision tasks in the Task Bank, all output from our networks.
The paper and supplementary material describing the methodology and evaluation.
Mid-Level Visual Representations for Improving Generalization and Sample Efficiency of Visuomotor Policies