Igl, Maximilian (author), Gambardella, Andrew (author), He, J. (author), Nardelli, Nantas (author), Siddharth, N (author), Böhmer, J.W. (author), Whiteson, Shimon (author) We present Multitask Soft Option Learning (MSOL), a hierarchical multitask framework based on Planning as Inference. MSOL extends the concept of options, using separate variational posteriors for each task, regularized by a shared prior. This “soft” version of options avoids several instabilities during training in a multitask setting, and...
conference paper 2020
