Learning to Synthesize Programs as Interpretable and Generalizable Policies

by   Dweep Trivedi, et al.

Recently, deep reinforcement learning (DRL) methods have achieved impressive performance on tasks in a variety of domains. However, neural network policies produced with DRL methods are not human-interpretable and often have difficulty generalizing to novel scenarios. To address these issues, prior works explore learning programmatic policies that are more interpretable and structured for generalization. Yet, these works either employ limited policy representations (e.g. decision trees, state machines, or predefined program templates) or require stronger supervision (e.g. input/output state pairs or expert demonstrations). We present a framework that instead learns to synthesize a program, which details the procedure to solve a task in a flexible and expressive manner, solely from reward signals. To alleviate the difficulty of learning to compose programs to induce the desired agent behavior from scratch, we propose to first learn a program embedding space that continuously parameterizes diverse behaviors in an unsupervised manner and then search over the learned program embedding space to yield a program that maximizes the return for a given task. Experimental results demonstrate that the proposed framework not only learns to reliably synthesize task-solving programs but also outperforms DRL and program synthesis baselines while producing interpretable and more generalizable policies. We also justify the necessity of the proposed two-stage learning scheme as well as analyze various methods for learning the program embedding.



There are no comments yet.


page 6

page 40

page 42


MoËT: Interpretable and Verifiable Reinforcement Learning via Mixture of Expert Trees

Deep Reinforcement Learning (DRL) has led to many recent breakthroughs o...

Programmatically Interpretable Reinforcement Learning

We study the problem of generating interpretable and verifiable policies...

Neural Logic Reinforcement Learning

Deep reinforcement learning (DRL) has achieved significant breakthroughs...

Neural Program Synthesis with Priority Queue Training

We consider the task of program synthesis in the presence of a reward fu...

Continuous Control for High-Dimensional State Spaces: An Interactive Learning Approach

Deep Reinforcement Learning (DRL) has become a powerful methodology to s...

Semantic embeddings for program behavior patterns

In this paper, we propose a new feature extraction technique for program...

Explaining Transition Systems through Program Induction

Explaining and reasoning about processes which underlie observed black-b...
This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.