Learning Programmatically Structured Representations with Perceptor GradientsDownload PDF

27 Sep 2018 (modified: 09 Feb 2019)ICLR 2019 Conference Blind SubmissionReaders: Everyone
  • Abstract: We present the perceptor gradients algorithm -- a novel approach to learning symbolic representations based on the idea of decomposing an agent's policy into i) a perceptor network extracting symbols from raw observation data and ii) a task encoding program which maps the input symbols to output actions. We show that the proposed algorithm is able to learn representations that can be directly fed into a Linear-Quadratic Regulator (LQR) or a general purpose A* planner. Our experimental results confirm that the perceptor gradients algorithm is able to efficiently learn transferable symbolic representations as well as generate new observations according to a semantically meaningful specification.
  • Keywords: representation learning, structured representations, symbols, programs
12 Replies