cs.LGApr 6, 2026

Selecting Decision-Relevant Concepts in Reinforcement Learning

Authors: Naveen Raman, Stephanie Milani, Fei Fang

Organizations: Carnegie Mellon University · New York University

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Training interpretable concept-based policies requires practitioners to manually select which human-understandable concepts an agent should reason with when making sequential decisions. This selection demands domain expertise, is time-consuming and costly, scales poorly with the number of candidates, and provides no performance guarantees. To overcome this limitation, we propose the first algorithms for principled automatic concept selection in sequential decision-making. Our key insight is that concept selection can be viewed through the lens of state abstraction: intuitively, a concept is decision-relevant if removing it would cause the agent to confuse states that require different actions. As a result, agents should rely on decision-relevant concepts; states with the same concept representation should share the same optimal action, which preserves the optimal decision structure of the original state space. This perspective leads to the Decision-Relevant Selection (DRS) algorithm, which selects a subset of concepts from a candidate set, along with performance bounds relating the selected concepts to the performance of the resulting policy. Empirically, DRS automatically recovers manually curated concept sets while matching or exceeding their performance, and improves the effectiveness of test-time concept interventions across reinforcement learning benchmarks and real-world healthcare environments.

Figures & tables

Figure 1: Standard pipeline for training concept-based policies. Practitioners select concepts for decision-making through a labor-intensive process of iteratively selecting candidate concepts, training concept-based policies, and evaluating their performance.
Figure 2: Concept-based models rely on a set of decision-relevant concepts that help distinguish between different states, yet currently these concepts are manually selected. In this work, we study how to identify and select decision-relevant concepts. Our key insight is that decision-relevant concepts best separate “different” states, where difference is defined by their decision consequences. We use this insight to develop algorithms for concept selection with performance guarantees.
Figure 3: Normalized reward of concept selection algorithms with perfect (top) and imperfect (bottom) concept predictors. Our algorithm, DRS, improves performance compared to the random, variance, and greedy baselines for four out of five environments in the perfect setting. DRS and DRS-log improve performance or are optimal in all environments in the imperfect setting.
Figure 4: We vary the number of timesteps that we train policies for in MiniGrid, while also varying the accuracy of concept predictors (left) or the number of concepts selected (right). Increasing the accuracy of concept predictors speeds up training, while increasing the number of concepts increases the maximum performance.
Figure 5: Impact of the number of concepts selected against the accuracy of the underlying concepts. Increasing either number or accuracy of concepts has a similar impact on performance, and that sufficiently accurate and many concepts are needed to ensure good performance.
Figure 6: We assess the impact of concept selection upon test-time intervention across four environments. Across all four environments, the DRS algorithm has the highest reward both before and after intervention, showing that well-selected concepts improve intervention.
Figure 7: The DRS algorithm can mimic manual concept selection in CUB. It matches performance (left) while selecting a subset of the manually selected concepts (right).
Appendix figures & tables7 assets

Supplementary material from the paper’s appendix.

Appendix

Figure 8: We assess the performance of a) not relaxing constraint P1c for DRS and b) varying ρ\rho for CartPole and MiniGrid. We find that ρ=0.75\rho=0.75 performs well across perfect and imperfect settings, while introducing P1c leads to lower performance in CartPole.
Figure 9: We compare concept selection algorithms while varying the number of concepts selected kk with perfect (top) and imperfect (bottom) concept predictors. We find that for CartPole and MiniGrid, DRS outperforms baselines across values of kk , while for environments such as Boxing and Glucose, a threshold number of concepts is needed before DRS outperforms baselines.
Concept GroupAvg. SelectedTotal
Agent
Position X15
Position Y25
Direction34
Action: Right02
Action: Left12
Appendix
Table 1: Average number of concepts selected per group by DRS in MiniGrid, with total concepts per group.
Figure 10: We compare the time to compute across methods and datasets, and find that all algorithms run in under ten minutes. While DRS and DRS-log take longer than variance and greedy, they still run quickly across environments.
Figure 11: We vary the number of timesteps that we train policies for in CartPole, while also varying the accuracy of concept predictors (left) or the number of concepts selected (right). When the number of concepts is small or the accuracy of the concepts is low, policies fail to train, with the efficiency of training increasing with both parameters.
Figure 12: We assess test-time intervention performance when varying the level of intervention α\alpha in CartPole. Across all α\alpha , the DRS algorithm performs best, with steadily improving performance as α\alpha is increased.
Figure 13: We vary the number of timesteps we use to train the policy π\pi without concepts in MiniGrid. Typically, we have π=π∗\pi=\pi^{*} , but here, we demonstrate that the DRS and DRS-log algorithms perform well even when π\pi is trained for only 100k timesteps.
More in this topic

Reinforcement Learning

Reinforcement learning studies how agents learn sequential decisions from rewards and interaction, balancing exploration, performance, safety, and adaptation.

See all 1,661 papers

Explore similar work

CardsList

  1. Interpretable reinforcement learning with decision-tree pruning

    Aug 7, 2026Mark Leon Ringer, Michel TokicInterpretability in RL

    Faculty of Mathematics, Informatics and Statistics, Ludwig-Maximilians-University Munich, Munich, Germany · Siemens AG, Data & Artificial Intelligence, Otto-Hahn-Ring 6, 81739 Munich, Germany

  2. CB-Agent: Concept-Based In-Context Learning for Sequential Action Selection

    Oct 10, 2026Qingyue Zhang, Huaze Tang, Minyu Zhang +3LLM AgentsIn-Context Learning

    Tsinghua Shenzhen International Graduate School, Tsinghua University, Shenzhen, China

  3. GRAIL: Autonomous Concept Grounding for Neuro-Symbolic Reinforcement Learning

    Apr 18, 2026Hikaru Shindo, Henri Rößler, Quentin Delfosse +1Reinforcement LearningLLM-Guided RL

    Technical University of Darmstadt · Google Intrinsic · German Research Center for Artificial Intelligence (DFKI) +1