# Shared Control with Black Box Agents using Oracle Queries

arXiv: https://arxiv.org/abs/2410.19612 · HTML: https://mlsanity.com/paper/2410.19612/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 9 of its 36 references that are in the map.
- First posted on arXiv 25 Oct 2024.

## Builds on (most cited first, 9 of 9)

- [OpenAI Gym](https://mlsanity.com/paper/1606.01540/) (2016)
- [Autonomous Agents Modelling Other Agents: A Comprehensive Survey and Open Problems](https://mlsanity.com/paper/1709.08071/) (2017)
- [Learning Latent Representations to Influence Multi-Agent Interaction](https://mlsanity.com/paper/2011.06619/) (2020)
- [Shared Autonomy via Hindsight Optimization for Teleoperation and Teaming](https://mlsanity.com/paper/1706.00155/) (2017)
- [Active Learning of Dynamics for Data-Driven Control Using Koopman Operators](https://mlsanity.com/paper/1906.05194/) (2019)
- [Policy Iteration for Decentralized Control of Markov Decision Processes](https://mlsanity.com/paper/1401.3460/) (2014)
- [Learning Models for Shared Control of Human-Machine Systems with Unknown Dynamics](https://mlsanity.com/paper/1808.08268/) (2018)
- [A Model-Based Active Testing Approach to Sequential Diagnosis](https://mlsanity.com/paper/1401.3850/) (2014)
- [Sequential Diagnosis by Abstraction](https://mlsanity.com/paper/1401.3892/) (2014)

## Built on by (most cited first, 1 of 1)

- [The Intelligent Disobedience Game: Formulating Disobedience in Stackelberg Games and Markov Decision Processes](https://mlsanity.com/paper/2603.20994/) (2026)

## Abstract (arXiv)

Shared control problems involve a robot learning to collaborate with a human. When learning a shared control policy, short communication between the agents can often significantly reduce running times and improve the system's accuracy. We extend the shared control problem to include the ability to directly query a cooperating agent. We consider two types of potential responses to a query, namely oracles: one that can provide the learner with the best action they should take, even when that action might be myopically wrong, and one with a bounded knowledge limited to its part of the system. Given this additional information channel, this work further presents three heuristics for choosing when to query: reinforcement learning-based, utility-based, and entropy-based. These heuristics aim to reduce a system's overall learning cost. Empirical results on two environments show the benefits of querying to learn a better control policy and the tradeoffs between the proposed heuristics.