Scalable Synthesis of Verified Controllers in Deep Reinforcement Learning

• URL: http://arxiv.org/abs/2104.10219v1
• Date: Tue, 20 Apr 2021 19:30:29 GMT
• Title: Scalable Synthesis of Verified Controllers in Deep Reinforcement Learning
• Authors: Zikang Xiong and Suresh Jagannathan
• Abstract summary: We propose an automated verification pipeline capable of synthesizing high-quality safety shields. Our key insight involves separating safety verification from neural controller, using pre-computed verified safety shields to constrain neural controller training. Experimental results over a range of realistic high-dimensional deep RL benchmarks demonstrate the effectiveness of our approach.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: There has been significant recent interest in devising verification techniques for learning-enabled controllers (LECs) that manage safety-critical systems. Given the opacity and lack of interpretability of the neural policies that govern the behavior of such controllers, many existing approaches enforce safety properties through the use of shields, a dynamic monitoring and repair mechanism that ensures a LEC does not emit actions that would violate desired safety conditions. These methods, however, have shown to have significant scalability limitations because verification costs grow as problem dimensionality and objective complexity increase. In this paper, we propose a new automated verification pipeline capable of synthesizing high-quality safety shields even when the problem domain involves hundreds of dimensions, or when the desired objective involves stochastic perturbations, liveness considerations, and other complex non-functional properties. Our key insight involves separating safety verification from neural controller, using pre-computed verified safety shields to constrain neural controller training which does not only focus on safety. Experimental results over a range of realistic high-dimensional deep RL benchmarks demonstrate the effectiveness of our approach.

Related papers

• BarrierSteer: LLM Safety via Learning Barrier Steering [83.12893815611052]
  BarrierSteer is a novel framework that formalizes safety by embedding learned non-linear safety constraints directly into the model's latent representation space.<n>We show that BarrierSteer substantially reduces adversarial success rates, decreases unsafe generations, and outperforms existing methods.
  arXiv  Detail & Related papers  (2026-02-23T18:19:46Z)
• Safe Reinforcement Learning via Recovery-based Shielding with Gaussian Process Dynamics Models [57.006252510102506]
  Reinforcement learning (RL) is a powerful framework for optimal decision-making and control but often lacks provable guarantees for safety-critical applications.<n>We introduce a novel recovery-based shielding framework that enables safe RL with a provable safety lower bound for unknown and non-linear continuous dynamical systems.
  arXiv  Detail & Related papers  (2026-02-12T22:03:35Z)
• Kernel-Based Learning of Safety Barriers [0.9367224590861915]
  Rapid integration of AI algorithms in safety-critical applications is raising concerns about the ability to meet stringent safety standards.<n>Traditional tools for formal safety verification struggle with the black-box nature of AI-driven systems.<n>We present a data-driven approach for safety verification and synthesis of black-box systems with discrete-time dynamics.
  arXiv  Detail & Related papers  (2026-01-17T10:42:35Z)
• A Safety-Constrained Reinforcement Learning Framework for Reliable Wireless Autonomy [1.5469452301122173]
  We propose a proactive safety-constrained RL framework that integrates proof-carrying control with empowerment-budgeted (EB) enforcement.<n>Our method achieves provable safety guarantees with minimal performance degradation.<n>Results highlight the potential of proactive safety constrained RL to enable trustworthy wireless autonomy in future 6G networks.
  arXiv  Detail & Related papers  (2026-01-12T02:02:52Z)
• LatentGuard: Controllable Latent Steering for Robust Refusal of Attacks and Reliable Response Generation [4.29885665563186]
  LATENTGUARD is a framework that combines behavioral alignment with supervised latent space control for interpretable and precise safety steering.<n>Our results show significant improvements in both safety controllability and response interpretability without compromising utility.
  arXiv  Detail & Related papers  (2025-09-24T07:31:54Z)
• CARE: Decoding Time Safety Alignment via Rollback and Introspection Intervention [68.95008546581339]
  Existing decoding-time interventions, such as Contrastive Decoding, often force a severe trade-off between safety and response quality.<n>We propose CARE, a novel framework for decoding-time safety alignment that integrates three key components.<n>The framework achieves a superior balance of safety, quality, and efficiency, attaining a low harmful response rate and minimal disruption to the user experience.
  arXiv  Detail & Related papers  (2025-09-01T04:50:02Z)
• Of Good Demons and Bad Angels: Guaranteeing Safe Control under Finite Precision [0.716879432974126]
  This paper bridges the gap between theoretical guarantees and real-world implementations by incorporating robustness under finite-precision perturbations.<n>We employ state-of-the-art mixed-precision fixed-point tuners to synthesize sound and efficient implementations, thus providing a complete end-to-end solution.<n>We evaluate our approach on case studies from the automotive and aeronautics domains, producing efficient NN implementations with rigorous infinite-time horizon safety guarantees.
  arXiv  Detail & Related papers  (2025-07-30T15:21:22Z)
• Verification of Visual Controllers via Compositional Geometric Transformations [49.81690518952909]
  We introduce a novel verification framework for perception-based controllers that can generate outer-approximations of reachable sets.<n>We provide theoretical guarantees on the soundness of our method and demonstrate its effectiveness across benchmark control environments.
  arXiv  Detail & Related papers  (2025-07-06T20:22:58Z)
• Safely Learning Controlled Stochastic Dynamics [61.82896036131116]
  We introduce a method that ensures safe exploration and efficient estimation of system dynamics.<n>After training, the learned model enables predictions of the system's dynamics and permits safety verification of any given control.<n>We provide theoretical guarantees for safety and derive adaptive learning rates that improve with increasing Sobolev regularity of the true dynamics.
  arXiv  Detail & Related papers  (2025-06-03T11:17:07Z)
• SafeSteer: Interpretable Safety Steering with Refusal-Evasion in LLMs [7.120986296945107]
  This paper investigates an approach called SafeSteer for guiding the outputs of large language models (LLMs)<n>We employ a simple, gradient-free unsupervised method to enhance safety steering while preserving text quality, topic relevance, and without explicit refusal.
  arXiv  Detail & Related papers  (2025-06-01T01:19:37Z)
• Learning Vision-Based Neural Network Controllers with Semi-Probabilistic Safety Guarantees [24.650302053973142]
  We introduce a novel semi-probabilistic verification framework that integrates reachability analysis with conditional generative adversarial networks. Next, we develop a gradient-based training approach that employs a novel safety loss function, safety-aware data-sampling strategy, and curriculum learning. Empirical evaluations in X-Plane 11 airplane landing simulation, CARLA-simulated autonomous lane following, and F1Tenth lane following in a visually-rich miniature environment demonstrate the effectiveness of our method in achieving formal safety guarantees.
  arXiv  Detail & Related papers  (2025-02-28T21:16:42Z)
• Nothing in Excess: Mitigating the Exaggerated Safety for LLMs via Safety-Conscious Activation Steering [56.92068213969036]
  Safety alignment is indispensable for Large language models (LLMs) to defend threats from malicious instructions. Recent researches reveal safety-aligned LLMs prone to reject benign queries due to the exaggerated safety issue. We propose a Safety-Conscious Activation Steering (SCANS) method to mitigate the exaggerated safety concerns.
  arXiv  Detail & Related papers  (2024-08-21T10:01:34Z)
• Unifying Qualitative and Quantitative Safety Verification of DNN-Controlled Systems [18.049286149364075]
  The rapid advance of deep reinforcement learning techniques enables the oversight of safety-critical systems through the utilization of Deep Neural Networks (DNNs) Most of the existing verification approaches rely on qualitative approaches, predominantly employing reachability analysis. We propose a novel framework for unifying both qualitative and quantitative safety verification problems.
  arXiv  Detail & Related papers  (2024-04-02T09:31:51Z)
• Sampling-based Safe Reinforcement Learning for Nonlinear Dynamical Systems [15.863561935347692]
  We develop provably safe and convergent reinforcement learning algorithms for control of nonlinear dynamical systems. Recent advances at the intersection of control and RL follow a two-stage, safety filter approach to enforcing hard safety constraints. We develop a single-stage, sampling-based approach to hard constraint satisfaction that learns RL controllers enjoying classical convergence guarantees.
  arXiv  Detail & Related papers  (2024-03-06T19:39:20Z)
• Scaling #DNN-Verification Tools with Efficient Bound Propagation and Parallel Computing [57.49021927832259]
  Deep Neural Networks (DNNs) are powerful tools that have shown extraordinary results in many scenarios. However, their intricate designs and lack of transparency raise safety concerns when applied in real-world applications. Formal Verification (FV) of DNNs has emerged as a valuable solution to provide provable guarantees on the safety aspect.
  arXiv  Detail & Related papers  (2023-12-10T13:51:25Z)
• Online Safety Property Collection and Refinement for Safe Deep Reinforcement Learning in Mapless Navigation [79.89605349842569]
  We introduce the Collection and Refinement of Online Properties (CROP) framework to design properties at training time. CROP employs a cost signal to identify unsafe interactions and use them to shape safety properties. We evaluate our approach in several robotic mapless navigation tasks and demonstrate that the violation metric computed with CROP allows higher returns and lower violations over previous Safe DRL approaches.
  arXiv  Detail & Related papers  (2023-02-13T21:19:36Z)
• Evaluating Model-free Reinforcement Learning toward Safety-critical Tasks [70.76757529955577]
  This paper revisits prior work in this scope from the perspective of state-wise safe RL. We propose Unrolling Safety Layer (USL), a joint method that combines safety optimization and safety projection. To facilitate further research in this area, we reproduce related algorithms in a unified pipeline and incorporate them into SafeRL-Kit.
  arXiv  Detail & Related papers  (2022-12-12T06:30:17Z)
• Recursively Feasible Probabilistic Safe Online Learning with Control Barrier Functions [60.26921219698514]
  We introduce a model-uncertainty-aware reformulation of CBF-based safety-critical controllers. We then present the pointwise feasibility conditions of the resulting safety controller. We use these conditions to devise an event-triggered online data collection strategy.
  arXiv  Detail & Related papers  (2022-08-23T05:02:09Z)
• Lyapunov-based uncertainty-aware safe reinforcement learning [0.0]
  InReinforcement learning (RL) has shown a promising performance in learning optimal policies for a variety of sequential decision-making tasks. In many real-world RL problems, besides optimizing the main objectives, the agent is expected to satisfy a certain level of safety. We propose a Lyapunov-based uncertainty-aware safe RL model to address these limitations.
  arXiv  Detail & Related papers  (2021-07-29T13:08:15Z)
• Enforcing robust control guarantees within neural network policies [76.00287474159973]
  We propose a generic nonlinear control policy class, parameterized by neural networks, that enforces the same provable robustness criteria as robust control. We demonstrate the power of this approach on several domains, improving in average-case performance over existing robust control methods and in worst-case stability over (non-robust) deep RL methods.
  arXiv  Detail & Related papers  (2020-11-16T17:14:59Z)
• Neural Lyapunov Redesign [36.2939747271983]
  Learning controllers must guarantee some notion of safety to ensure that it does not harm either the agent or the environment. Lyapunov functions are effective tools to assess stability in nonlinear dynamical systems. We propose a two-player collaborative algorithm that alternates between estimating a Lyapunov function and deriving a controller that gradually enlarges the stability region.
  arXiv  Detail & Related papers  (2020-06-06T19:22:20Z)

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.