# Belief-Grounded Networks for Accelerated Robot Learning under Partial Observability

arXiv: https://arxiv.org/abs/2010.09170 · HTML: https://mlsanity.com/paper/2010.09170/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 19 Oct 2020.

## Built on by (most cited first, 10 of 10)

- [Partially Observable Markov Decision Processes in Robotics: A Survey](https://mlsanity.com/paper/2209.10342/) (2022)
- [Unbiased Asymmetric Reinforcement Learning under Partial Observability](https://mlsanity.com/paper/2105.11674/) (2021)
- [Recurrent Off-policy Baselines for Memory-based Continuous Control](https://mlsanity.com/paper/2110.12628/) (2021)
- [Leveraging Fully Observable Policies for Learning under Partial Observability](https://mlsanity.com/paper/2211.01991/) (2022)
- [Equivariant Reinforcement Learning under Partial Observability](https://mlsanity.com/paper/2408.14336/) (2024)
- [Informed POMDP: Leveraging Additional Information in Model-Based RL](https://mlsanity.com/paper/2306.11488/) (2023)
- [Symmetry-aware Reinforcement Learning for Robotic Assembly under Partial Observability with a Soft Wrist](https://mlsanity.com/paper/2402.18002/) (2024)
- [Seq2Seq Imitation Learning for Tactile Feedback-based Manipulation](https://mlsanity.com/paper/2303.02646/) (2023)
- [DAVIS: Planning Agent with Knowledge Graph-Powered Inner Monologue](https://mlsanity.com/paper/2410.09252/) (2024)
- [Reinforcement Learning from Real-Robot Interaction without High-Quality State Estimates](https://mlsanity.com/paper/2610.14042/) (2026)

## Abstract (arXiv)

Many important robotics problems are partially observable in the sense that a single visual or force-feedback measurement is insufficient to reconstruct the state. Standard approaches involve learning a policy over beliefs or observation-action histories. However, both of these have drawbacks; it is expensive to track the belief online, and it is hard to learn policies directly over histories. We propose a method for policy learning under partial observability called the Belief-Grounded Network (BGN) in which an auxiliary belief-reconstruction loss incentivizes a neural network to concisely summarize its input history. Since the resulting policy is a function of the history rather than the belief, it can be executed easily at runtime. We compare BGN against several baselines on classic benchmark tasks as well as three novel robotic touch-sensing tasks. BGN outperforms all other tested methods and its learned policies work well when transferred onto a physical robot.