MANGA: Method Agnostic Neural-policy Generalization and Adaptation | IEEE Conference Publication | IEEE Xplore