From Black Box to Executable Logic: Explainable Reinforcement Learning through Prolog Expert Systems

• URL: http://arxiv.org/abs/2607.15459v2
• Date: Mon, 20 Jul 2026 09:36:34 GMT
• Title: From Black Box to Executable Logic: Explainable Reinforcement Learning through Prolog Expert Systems
• Abstract summary: A trained deep reinforcement learning policy is a black box, and we ask whether it can be made explainable by rewriting it as an executable logic program.<n>We present a three-stage post-hoc transformation that extracts a frozen policy optimization teacher, induces an ordered rule list from its decisions, and emits the result as a Prolog program.<n>A return-loss bound makes the distilled program a machine-checkable certificate in a finite decision process, and the expansion loop improves monotonically and terminates.
• Score: 1.5736899098702974
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: A trained deep reinforcement learning policy is a black box, and we ask whether it can be made explainable by rewriting it as an executable logic program that reproduces its behaviour and that a person can read, a logic engine can run, and an optimizer can edit. We present a three-stage post-hoc transformation that extracts a frozen proximal policy optimization teacher, induces an ordered rule list from its decisions in the manner of classical relational learning, and emits the result as a Prolog program whose every decision is executed by an off-the-shelf logic engine; a subsequent expansion stage edits the rule base and accepts an edit only when policy evaluation certifies a return increase. We prove four guarantees. A return-loss bound makes the distilled program a machine-checkable certificate in a finite Markov decision process, and the expansion loop improves monotonically and terminates. For the continuous-observation setting we answer whether the conversion is possible at all: the propositional threshold instantiation converts the network to arbitrary fidelity as the resolution B grows, with disagreement O(1/B) and a return gap that closes at the same rate, and a matching lower bound shows the cost is exponential in the observation dimension for an oblique decision boundary. Empirically, on a two-room key-and-door task with 16,944 reachable states the expanded Prolog program attains exact optimal return in every seed and, in a budget-capped regime, exceeds the stochastic teacher on exact return in ten of ten seeds. On three continuous-control tasks the emitted program substitutes the network, matching the neural teacher within noise on Acrobot with eleven clauses and recovering about 97% of its return on CartPole, while on the finer-control LunarLander it recovers only partially, exactly the ceiling the exponential lower bound predicts.

Related papers

• Knowing When to Ask for Help: Bayesian Self-Escalation in Hierarchical LLM Agents [0.0]
  Current agent systems decide delegation before reasoning begins.<n>We study a third regime: an agent that recognises, during its own reasoning, that it is unlikely to succeed.
  arXiv  Detail & Related papers  (2026-08-25T05:35:07Z)
• Search as Computation Allocation [52.080489413778]
  Many algorithms spend an internal resource before returning a decision and are evaluated only by the quality of that terminal output.<n>We formalize such procedures as terminal computation-allocation problems.<n>Bellman equations characterize optimal allocation under fixed budgets, priced computation, and exact certification.
  arXiv  Detail & Related papers  (2026-07-30T08:45:14Z)
• RecRec: Latent Interests Recursive Reasoning for Sequential Recommendation [52.23179023639867]
  Sequential recommender systems rely on a single forward pass to encode user interaction and predict the next item.<n>How to structure the reasoning process for sequential recommendation remains an open question.<n>Existing approaches couple reasoning and prediction in a single $d$-dimensional state, limiting reasoning depth.<n>We propose RecRec (Recursive Reasoning for Recommendation), an RL-free framework that overcomes fixed $d$-dimensional state of methods.
  arXiv  Detail & Related papers  (2026-07-14T16:28:28Z)
• Stabilizing Extrapolation in Looped Transformers via Learned Stochastic Stopping [56.14767235650558]
  Looped Transformers, which repeatedly apply a shared transformer block, are an architecturally natural fit for variable-length computation tasks.<n>We trace this variance to the spurious correlation in simple algorithmic tasks between sequence length and number of loops.<n>Our work suggests that "when to stop" should be treated as a training design choice, not merely an inference-time-allocation rule.
  arXiv  Detail & Related papers  (2026-06-29T08:58:09Z)
• Rethinking Reward Supervision: Rubric-Conditioned Self-Distillation [60.55792673956761]
  We propose textbfRubric-Conditioned Self-Distillation, a framework that incorporates rubrics as structured, fine-grained feedback for on-policy self-distillation.<n>We evaluate on a diverse suite of science reasoning benchmarks and results show that rubric-conditioned self-distillation effectively converts rubric-level criteria into token-level guidance.
  arXiv  Detail & Related papers  (2026-06-17T17:54:04Z)
• Repair Before Veto, When Repair Is Hidden: Quantum-Accessible Features for Repair-Augmented Constraint Learning [6.016090674751934]
  We introduce Q-RACL (Quantum Repair-Augmented Constraint Learning), a repair-before-veto framework.<n>It accepts a candidate when a sequential repair plan restores feasibility and preference; otherwise it returns structured rejection credit.<n>Across six primes and ten seeds, bounded raw-input classical policies and a wrong raw-Fourier encoding remain near chance.<n>Q-DLP policy keeps false-veto rate below 1.1%, wins all paired seeds, and yields QNI_cond = 0.9777 to 0.9972.
  arXiv  Detail & Related papers  (2026-06-06T07:17:38Z)
• The Deterministic Horizon: Impossibility Results as Design Specifications for Trustworthy AI Systems [0.0]
  This thesis turns impossibility results from curiosities into design rules.<n>Its flagship result proves an accuracy ceiling set by architecture alone.<n>The same argument recasts across subfields.
  arXiv  Detail & Related papers  (2026-05-21T20:48:35Z)
• Restarted contractive operators to learn at equilibrium [0.0]
  We introduce an algorithm that combines a restart strategy with JFB computed by AD and we show that the learned steps can be made arbitrarily close to the optimal DEQ framework.<n>We show that this method is effective for training weights in weighted norms; stepsizes and regularization levels of Plug-and-Play schemes; and a DRUNet denoiser embedded in Forward-Backward iterates.
  arXiv  Detail & Related papers  (2025-06-16T08:38:56Z)
• Maximum-Likelihood Inverse Reinforcement Learning with Finite-Time Guarantees [56.848265937921354]
  Inverse reinforcement learning (IRL) aims to recover the reward function and the associated optimal policy. Many algorithms for IRL have an inherently nested structure. We develop a novel single-loop algorithm for IRL that does not compromise reward estimation accuracy.
  arXiv  Detail & Related papers  (2022-10-04T17:13:45Z)
• Improving the Efficiency of Off-Policy Reinforcement Learning by Accounting for Past Decisions [20.531576904743282]
  Off-policy estimation bias is corrected in a per-decision manner. Off-policy algorithms such as Tree Backup and Retrace rely on this mechanism. We propose a multistep operator that permits arbitrary past-dependent traces.
  arXiv  Detail & Related papers  (2021-12-23T00:07:28Z)
• Improper Learning with Gradient-based Policy Optimization [62.50997487685586]
  We consider an improper reinforcement learning setting where the learner is given M base controllers for an unknown Markov Decision Process. We propose a gradient-based approach that operates over a class of improper mixtures of the controllers.
  arXiv  Detail & Related papers  (2021-02-16T14:53:55Z)
• Upper Confidence Primal-Dual Reinforcement Learning for CMDP with Adversarial Loss [145.54544979467872]
  We consider online learning for episodically constrained Markov decision processes (CMDPs) We propose a new emphupper confidence primal-dual algorithm, which only requires the trajectories sampled from the transition model. Our analysis incorporates a new high-probability drift analysis of Lagrange multiplier processes into the celebrated regret analysis of upper confidence reinforcement learning.
  arXiv  Detail & Related papers  (2020-03-02T05:02:23Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.