# From explanation to synthesis: Compositional program induction for learning from demonstration

arXiv: https://arxiv.org/abs/1902.10657 · HTML: https://mlsanity.com/paper/1902.10657/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- Placed in Robot manipulation imitation learning, one of 13 directions in Offline RL and imitation learning from data.
- Builds on 16 of its 48 references that are in the map.
- First posted on arXiv 27 Feb 2019.

## Placed in

- Offline RL and imitation learning from data / [Robot manipulation imitation learning](https://mlsanity.com/offlinerl/robot-manipulation-imitation-learning/)

## Builds on (most cited first, 16 of 16)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization](https://mlsanity.com/paper/1610.02391/) (2016)
- [Visualizing and Understanding Convolutional Networks](https://mlsanity.com/paper/1311.2901/) (2013)
- [End-to-End Training of Deep Visuomotor Policies](https://mlsanity.com/paper/1504.00702/) (2015)
- [Learning Complex Dexterous Manipulation with Deep Reinforcement Learning and Demonstrations](https://mlsanity.com/paper/1709.10087/) (2017)
- [Supersizing Self-supervision: Learning to Grasp from 50K Tries and 700 Robot Hours](https://mlsanity.com/paper/1509.06825/) (2015)
- [Programmatically Interpretable Reinforcement Learning](https://mlsanity.com/paper/1804.02477/) (2018)
- [Verifiable Reinforcement Learning via Policy Extraction](https://mlsanity.com/paper/1805.08328/) (2018)
- [One-Shot Imitation from Observing Humans via Domain-Adaptive Meta-Learning](https://mlsanity.com/paper/1802.01557/) (2018)
- [Reinforcement and Imitation Learning for Diverse Visuomotor Skills](https://mlsanity.com/paper/1802.09564/) (2018)
- [Building machines that adapt and compute like brains](https://mlsanity.com/paper/1711.04203/) (2017)
- [Deep Drone Racing: Learning Agile Flight in Dynamic Environments](https://mlsanity.com/paper/1806.08548/) (2018)
- [Differentiable Programs with Neural Libraries](https://mlsanity.com/paper/1611.02109/) (2016)
- [Beyond imitation: Zero-shot task transfer on robots by learning concepts as cognitive programs](https://mlsanity.com/paper/1812.02788/) (2018)
- [Specification Patterns for Robotic Missions](https://mlsanity.com/paper/1901.02077/) (2019)
- [Explaining Transition Systems through Program Induction](https://mlsanity.com/paper/1705.08320/) (2017)

## Built on by (most cited first, 9 of 9)

- [A Survey on Interpretable Reinforcement Learning](https://mlsanity.com/paper/2112.13112/) (2021)
- [Learning to Synthesize Programs as Interpretable and Generalizable Policies](https://mlsanity.com/paper/2108.13643/) (2021)
- [SKID RAW: Skill Discovery from Raw Trajectories](https://mlsanity.com/paper/2103.14610/) (2021)
- [Hierarchical Neural Program Synthesis](https://mlsanity.com/paper/2303.06018/) (2023)
- [Explanatory machine learning for sequential human teaching](https://mlsanity.com/paper/2205.10250/) (2022)
- [Imitation Learning of Robot Policies by Combining Language, Vision and Demonstration](https://mlsanity.com/paper/1911.11744/) (2019)
- [Projection: A Mechanism for Human-like Reasoning in Artificial Intelligence](https://mlsanity.com/paper/2103.13512/) (2021)
- [Semi-supervised Learning From Demonstration Through Program Synthesis: An Inspection Robot Case Study](https://mlsanity.com/paper/2007.12500/) (2020)
- [Attainment Regions in Feature-Parameter Space for High-Level Debugging in Autonomous Robots](https://mlsanity.com/paper/2108.03150/) (2021)

## Abstract (arXiv)

Hybrid systems are a compact and natural mechanism with which to address problems in robotics. This work introduces an approach to learning hybrid systems from demonstrations, with an emphasis on extracting models that are explicitly verifiable and easily interpreted by robot operators. We fit a sequence of controllers using sequential importance sampling under a generative switching proportional controller task model. Here, we parameterise controllers using a proportional gain and a visually verifiable joint angle goal. Inference under this model is challenging, but we address this by introducing an attribution prior extracted from a neural end-to-end visuomotor control model. Given the sequence of controllers comprising a task, we simplify the trace using grammar parsing strategies, taking advantage of the sequence compositionality, before grounding the controllers by training perception networks to predict goals given images. Using this approach, we are successfully able to induce a program for a visuomotor reaching task involving loops and conditionals from a single demonstration and a neural end-to-end model. In addition, we are able to discover the program used for a tower building task. We argue that computer program-like control systems are more interpretable than alternative end-to-end learning approaches, and that hybrid systems inherently allow for better generalisation across task configurations.