Learning Programmatically Structured Representations with Perceptor Gradients

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.

Cite

Text

Penkov and Ramamoorthy. "Learning Programmatically Structured Representations with Perceptor Gradients." International Conference on Learning Representations, 2019.

Markdown

[Penkov and Ramamoorthy. "Learning Programmatically Structured Representations with Perceptor Gradients." International Conference on Learning Representations, 2019.](https://mlanthology.org/iclr/2019/penkov2019iclr-learning/)

BibTeX

@inproceedings{penkov2019iclr-learning,
  title     = {{Learning Programmatically Structured Representations with Perceptor Gradients}},
  author    = {Penkov, Svetlin and Ramamoorthy, Subramanian},
  booktitle = {International Conference on Learning Representations},
  year      = {2019},
  url       = {https://mlanthology.org/iclr/2019/penkov2019iclr-learning/}
}