# Discovering What You Can Control: Interventional Boundary Discovery for Reinforcement Learning

arXiv: https://arxiv.org/abs/2603.18257 · HTML: https://mlsanity.com/paper/2603.18257/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Deep RL algorithms and representations, one of 15 directions in Deep reinforcement learning methods.
- Builds on 23 of its 34 references that are in the map.
- First posted on arXiv 18 Mar 2026.

## Placed in

- Deep reinforcement learning methods / [Deep RL algorithms and representations](https://mlsanity.com/deeprl/deep-rl-algorithms-and-representations/)

## Builds on (most cited first, 23 of 23)

- [Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor](https://mlsanity.com/paper/1801.01290/) (2018)
- [Addressing Function Approximation Error in Actor-Critic Methods](https://mlsanity.com/paper/1802.09477/) (2018)
- [DeepMind Control Suite](https://mlsanity.com/paper/1801.00690/) (2018)
- [CURL: Contrastive Unsupervised Representations for Reinforcement Learning](https://mlsanity.com/paper/2004.04136/) (2020)
- [Mastering Visual Continuous Control: Improved Data-Augmented Reinforcement Learning](https://mlsanity.com/paper/2107.09645/) (2021)
- [Variational Information Maximisation for Intrinsically Motivated Reinforcement Learning](https://mlsanity.com/paper/1509.08731/) (2015)
- [CITRIS: Causal Identifiability from Temporal Intervened Sequences](https://mlsanity.com/paper/2202.03169/) (2022)
- [Stabilizing Deep Q-Learning with ConvNets and Vision Transformers under Data Augmentation](https://mlsanity.com/paper/2107.00644/) (2021)
- [The Distracting Control Suite -- A Challenging Benchmark for Reinforcement Learning from Pixels](https://mlsanity.com/paper/2101.02722/) (2021)
- [Denoised MDPs: Learning World Models Better Than the World Itself](https://mlsanity.com/paper/2206.15477/) (2022)
- [WoCoCo: Learning Whole-Body Humanoid Control with Sequential Contacts](https://mlsanity.com/paper/2406.06005/) (2024)
- [Causal Representation Learning for Instantaneous and Temporal Effects in Interactive Systems](https://mlsanity.com/paper/2206.06169/) (2022)
- [Unifying Causal Representation Learning with the Invariance Principle](https://mlsanity.com/paper/2409.02772/) (2024)
- [Provable RL with Exogenous Distractors via Multistep Inverse Dynamics](https://mlsanity.com/paper/2110.08847/) (2021)
- [Score-based Causal Representation Learning: Linear and General Transformations](https://mlsanity.com/paper/2402.00849/) (2024)
- [Linear Causal Representation Learning from Unknown Multi-node Interventions](https://mlsanity.com/paper/2406.05937/) (2024)
- [A Forget-and-Grow Strategy for Deep Reinforcement Learning Scaling in Continuous Control](https://mlsanity.com/paper/2507.02712/) (2025)
- [PIGDreamer: Privileged Information Guided World Models for Safe Partially Observable Reinforcement Learning](https://mlsanity.com/paper/2508.02159/) (2025)
- [Confounding Robust Deep Reinforcement Learning: A Causal Approach](https://mlsanity.com/paper/2510.21110/) (2025)
- [Zero-Shot Reinforcement Learning Under Partial Observability](https://mlsanity.com/paper/2506.15446/) (2025)
- [Learning a Fast Mixing Exogenous Block MDP using a Single Trajectory](https://mlsanity.com/paper/2410.03016/) (2024)
- [Disentangled (Un)Controllable Features](https://mlsanity.com/paper/2211.00086/) (2022)
- [Offline Action-Free Learning of Ex-BMDPs by Comparing Diverse Datasets](https://mlsanity.com/paper/2503.21018/) (2025)

## Built on by (most cited first, 2 of 2)

- [RAM: Recover Any 3D Human Motion in-the-Wild](https://mlsanity.com/paper/2603.19929/) (2026)
- [BCL: Bayesian In-Context Learning Framework for Information Extraction](https://mlsanity.com/paper/2606.18620/) (2026)

## Abstract (arXiv)

When an RL agent's observations contain distractors driven by the same confounders as its true state, observational data alone cannot identify which dimensions the agent controls. In our benchmarks, even state-conditioned observational selectors can collapse when distractors mimic controllable state variables. We propose Interventional Boundary Discovery (IBD), which treats the agent's own action channel as a source of randomized interventions: randomizing actions implements an interventional contrast, and per-dimension two-sample tests with FDR correction produce a binary mask over observation dimensions. Across 12 continuous-control settings with up to 100 distractors, IBD matches oracle return in 11 of 12 settings, while observational baselines including mutual information, state-conditioned forward models, and gradient-based sensitivity often underperform simply passing the full observation to SAC. Code is available at https://github.com/jiaxin26/IBD-RL