This paper studies satisfaction of temporal properties on unknown stochastic processes that have continuous state spaces. We show how reinforcement learning (RL) can be applied for computing policies that are finite-memory and deterministic using only the paths of the stochastic process. We address properties expressed in linear temporal logic (LTL) and use their automaton representation to give a path-dependent reward function maximised via the RL algorithm. We develop the required assumptions and theories for the convergence of the learned policy to the optimal policy in the continuous state space. To improve the performance of the learning on the constructed sparse reward function, we propose a sequential learning procedure based on a sequence of labelling functions obtained from the positive normal form of the LTL specification. We use this procedure to guide the RL algorithm towards a policy that converges to an optimal policy under suitable assumptions on the process. We demonstrate the approach on a 4-dim cart-pole system and 6-dim boat driving problem.
Relative Entropy Regularized Reinforcement Learning for Efficient Encrypted Policy Synthesis
School of Aeronautics and Astronautics, Purdue University（航空宇航工程学院，普渡大学）；ASRI, the Department of Electrical and Computer Engineering, Seoul National University（ASRI，电气电子工程系，首尔国立大学）；Japan Society for the Promotion of Science（日本学术振兴会）
SafePILCO is a software tool for safe and data-efficient policy search with reinforcement learning. It extends the known PILCO algorithm, originally written in MATLAB, to support safe learning. We provide a Python implementation and leverage existing libraries that allow the codebase to remain short and modular, which is appropriate for wider use by the verification, reinforcement learning, and control communities.
University of Illinois at Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）
Direct policy search has been widely applied in modern reinforcement learning and continuous control. However, the theoretical properties of direct policy search on nonsmooth robust control synthesis have not been fully understood. The optimal $\mathcal{H}_\infty$ control framework aims at designing a policy to minimize the closed-loop $\mathcal{H}_\infty$ norm, and is arguably the most fundamental robust control paradigm. In this work, we show that direct policy search is guaranteed to find the global solution of the robust $\mathcal{H}_\infty$ state-feedback control design problem. Notice th…
Compositional Policy Learning in Stochastic Control Systems with Formal Guarantees
Institute of Science and Technology Austria (ISTA)（奥地利科学技术研究所（ISTA））；Massachusetts Institute of Technology（麻省理工学院）；The Pennsylvania State University（宾夕法尼亚州立大学）
Learning Policy Representations for Steerable Behavior Synthesis
Department of Electrical and Systems Engineering, University of Pennsylvania, Philadelphia, USA（宾夕法尼亚大学电气与系统工程系）；Department of Signal Theory and Communications, King Juan Carlos University, Madrid, Spain（胡安 Carlos大学信号理论与通信系）
1-2-3-Go! Policy Synthesis for Parameterized Markov Decision Processes via Decision-Tree Learning and Generalization
Technical University of Munich（慕尼黑工业大学）；Vali-e-Asr University of Rafsanjan（拉夫桑贾德瓦利阿斯尔大学）；Masaryk University（马萨里克大学）；Institute of Science and Technology（科学技术研究院）
LCRL is a software tool that implements model-free Reinforcement Learning (RL) algorithms over unknown Markov Decision Processes (MDPs), synthesising policies that satisfy a given linear temporal specification with maximal probability. LCRL leverages partially deterministic finite-state machines known as Limit Deterministic Buchi Automata (LDBA) to express a given linear temporal specification. A reward function for the RL algorithm is shaped on-the-fly, based on the structure of the LDBA. Theoretical guarantees under proper assumptions ensure the convergence of the RL algorithm to an optimal …
University of Michigan（密歇根大学）；University of Illinois at Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）
Motivated by the recent empirical success of policy-based reinforcement learning (RL), there has been a research trend studying the performance of policy-based RL methods on standard control benchmark problems. In this paper, we examine the effectiveness of policy-based RL methods on an important robust control problem, namely $μ$ synthesis. We build a connection between robust adversarial RL and $μ$ synthesis, and develop a model-free version of the well-known $DK$-iteration for solving state-feedback $μ$ synthesis with static $D$-scaling. In the proposed algorithm, the $K$ step mimics the cl…
Deep reinforcement learning has led to several recent breakthroughs, though the learned policies are often based on black-box neural networks. This makes them difficult to interpret and to impose desired specification constraints during learning. We present an iterative framework, MORL, for improving the learned policies using program synthesis. Concretely, we propose to use synthesis techniques to obtain a symbolic representation of the learned policy, which can then be debugged manually or automatically using program repair. After the repair step, we use behavior cloning to obtain the policy…
Google Brain（谷歌大脑）；Tel-Aviv University（特拉维夫大学）；AI2（艾伦人工智能研究所）；SayMosaic Inc.（赛莫赛克公司）
We present Memory Augmented Policy Optimization (MAPO), a simple and novel way to leverage a memory buffer of promising trajectories to reduce the variance of policy gradient estimate. MAPO is applicable to deterministic environments with discrete actions, such as structured prediction and combinatorial optimization tasks. We express the expected return objective as a weighted sum of two terms: an expectation over the high-reward trajectories inside the memory buffer, and a separate expectation over trajectories outside the buffer. To make an efficient algorithm of MAPO, we propose: (1) memory…
Safety is the major consideration in controlling complex dynamical systems using reinforcement learning (RL), where the safety certificate can provide provable safety guarantee. A valid safety certificate is an energy function indicating that safe states are with low energy, and there exists a corresponding safe control policy that allows the energy function to always dissipate. The safety certificate and the safe control policy are closely related to each other and both challenging to synthesize. Therefore, existing learning-based studies treat either of them as prior knowledge to learn the o…
Neural inductive program synthesis is a task generating instructions that can produce desired outputs from given inputs. In this paper, we focus on the generation of a chunk of assembly code that can be executed to match a state change inside the CPU and RAM. We develop a neural program synthesis algorithm, AutoAssemblet, learned via self-learning reinforcement learning that explores the large code space efficiently. Policy networks and value networks are learned to reduce the breadth and depth of the Monte Carlo Tree Search, resulting in better synthesis performance. We also propose an effect…
The automatic synthesis of a policy through reinforcement learning (RL) from a given set of formal requirements depends on the construction of a reward signal and consists of the iterative application of many policy-improvement steps. The synthesis algorithm has to balance target, safety, and comfort requirements in a single objective and to guarantee that the policy improvement does not increase the number of safety-requirements violations, especially for safety-critical applications. In this work, we present a solution to the synthesis problem by solving its two main challenges: reward-shapi…
Vrije Universiteit Brussel（布鲁塞尔自由大学（荷兰语））；Université Libre de Bruxelles（布鲁塞尔自由大学（法语））；Delft University of Technology（代尔夫特理工大学）；Leiden Institute of Advanced Computer Science（莱顿大学高级计算机科学研究所）
Today's advanced Reinforcement Learning algorithms produce black-box policies, that are often difficult to interpret and trust for a person. We introduce a policy distilling algorithm, building on the CN2 rule mining algorithm, that distills the policy into a rule-based decision system. At the core of our approach is the fact that an RL process does not just learn a policy, a mapping from states to actions, but also produces extra meta-information, such as action values indicating the quality of alternative actions. This meta-information can indicate whether more than one action is near-optima…
This paper proposes DeepSynth, a method for effective training of deep Reinforcement Learning (RL) agents when the reward is sparse and non-Markovian, but at the same time progress towards the reward requires achieving an unknown sequence of high-level objectives. Our method employs a novel algorithm for synthesis of compact automata to uncover this sequential structure automatically. We synthesise a human-interpretable automaton from trace data collected by exploring the environment. The state space of the environment is then enriched with the synthesised automaton so that the generation of a…
Quantum Reinforcement Learning-Guided Diffusion Model for Image Synthesis via Hybrid Quantum-Classical Generative Model Architectures
Independent Researcher Cambridge, USA（独立研究者 美国剑桥）；Department of Electrophysics National Yang Ming Chiao Tung University Hsinchu, Taiwan（电子物理系 国立阳明交通大学 台北县）
Reinforcement learning has shown great promise for synthesizing realistic human behaviors by learning humanoid control policies from motion capture data. However, it is still very challenging to reproduce sophisticated human skills like ballet dance, or to stably imitate long-term human behaviors with complex transitions. The main difficulty lies in the dynamics mismatch between the humanoid model and real humans. That is, motions of real humans may not be physically possible for the humanoid model. To overcome the dynamics mismatch, we propose a novel approach, residual force control (RFC), t…
We study the synthesis of mode switching protocols for a class of discrete-time switched linear systems in which the mode jumps are governed by Markov decision processes (MDPs). We call such systems MDP-JLS for brevity. Each state of the MDP corresponds to a mode in the switched system. The probabilistic state transitions in the MDP represent the mode transitions. We focus on finding a policy that selects the switching actions at each mode such that the switched system that follows these actions is guaranteed to be stable. Given a policy in the MDP, the considered MDP-JLS reduces to a Markov j…
Recent employments of SMT solvers within the Lyapunov function synthesis provided effective tools for automated construction of Lyapunov functions alongside with sound computer-assisted certificates. The main benefit of the suggested approach is the formal correctness and elimination of the numerical uncertainty. In the present work, we extend the SMT-based synthesis approach for wider classes of continuous and discrete-time systems. Additionally, we address constructions of Lyapunov functions for state-dependent switching systems. We illustrate our approach by means of various examples from t…
Air Force Research Laboratory（空军研究实验室）；University of Central Florida（中佛罗里达大学）；University of Colorado Boulder（科罗拉多大学博尔德分校）
Given a Markov decision process (MDP) and a linear-time ($ω$-regular or LTL) specification, the controller synthesis problem aims to compute the optimal policy that satisfies the specification. More recently, problems that reason over the asymptotic behavior of systems have been proposed through the lens of steady-state planning. This entails finding a control policy for an MDP such that the Markov chain induced by the solution policy satisfies a given set of constraints on its steady-state distribution. This paper studies a generalization of the controller synthesis problem for a linear-time …
In decision-making problems, the actions of an agent may reveal sensitive information that drives its decisions. For instance, a corporation's investment decisions may reveal its sensitive knowledge about market dynamics. To prevent this type of information leakage, we introduce a policy synthesis algorithm that protects the privacy of the transition probabilities in a Markov decision process. We use differential privacy as the mathematical definition of privacy. The algorithm first perturbs the transition probabilities using a mechanism that provides differential privacy. Then, based on the p…
In this paper, we propose a formal controller synthesis approach for integrating a population of plug-in electric vehicles in frequency regulation of power systems. This approach is the first application of formal methods to the smart grids in particular to the frequency regulation of power systems. We design and simulate a novel symbolic controller for the Great Britain power system. The proposed controller enhances the frequency response behaviour of the system when encountered with a large outage event. The symbolic controller guarantees the settlement of the after-event's frequency in the …
The dramatic increase of autonomous systems subject to variable environments has given rise to the pressing need to consider risk in both the synthesis and verification of policies for these systems. This paper aims to address a few problems regarding risk-aware verification and policy synthesis, by first developing a sample-based method to bound the risk measure evaluation of a random variable whose distribution is unknown. These bounds permit us to generate high-confidence verification statements for a large class of robotic systems. Second, we develop a sample-based method to determine solu…
In this paper, we study formal synthesis of control policies for partially observed jump-diffusion systems against complex logic specifications. Given a state estimator, we utilize a discretization-free approach for formal synthesis of control policies by using a notation of control barrier functions without requiring any knowledge of the estimation accuracy. Our goal is to synthesize an offline control policy providing (potentially maximizing) a lower bound on the probability that the trajectories of the partially observed jump-diffusion system satisfy some complex specifications expressed by…
Decision-making policies for agents are often synthesized with the constraint that a formal specification of behaviour is satisfied. Here we focus on infinite-horizon properties. On the one hand, Linear Temporal Logic (LTL) is a popular example of a formalism for qualitative specifications. On the other hand, Steady-State Policy Synthesis (SSPS) has recently received considerable attention as it provides a more quantitative and more behavioural perspective on specifications, in terms of the frequency with which states are visited. Finally, rewards provide a classic framework for quantitative p…
We propose an automatic and formally sound method for synthesising Lyapunov functions for the asymptotic stability of autonomous non-linear systems. Traditional methods are either analytical and require manual effort or are numerical but lack of formal soundness. Symbolic computational methods for Lyapunov functions, which are in between, give formal guarantees but are typically semi-automatic because they rely on the user to provide appropriate function templates. We propose a method that finds Lyapunov functions fully automatically$-$using machine learning$-$while also providing formal guara…
This paper proposes a framework for automatic formal controller synthesis for general hybrid systems with a subset of safety and reachability specifications. The framework uses genetic programming to automatically co-synthesize controllers and candidate Lyapunov-like functions. These candidate Lyapunov-like functions are used to formally verify the control specification, and their correctness is proven using a Satisfiability Modulo Theories solver. The advantages of this approach are: no restriction is made to polynomial systems, the synthesized controllers are expressed as compact expressions…
We introduce an automated, formal, counterexample-based approach to synthesise Barrier Certificates (BC) for the safety verification of continuous and hybrid dynamical models. The approach is underpinned by an inductive framework: this is structured as a sequential loop between a learner, which manipulates a candidate BC structured as a neural network, and a sound verifier, which either certifies the candidate's validity or generates counter-examples to further guide the learner. We compare the approach against state-of-the-art techniques, over polynomial and non-polynomial dynamical models: t…
We propose a counterexample-guided inductive synthesis framework for the formal synthesis of closed-form sampled-data controllers for nonlinear systems to meet STL specifications over finite-time trajectories. Rather than stating the STL specification for a single initial condition, we consider an (infinite and bounded) set of initial conditions. Candidate solutions are proposed using genetic programming, which evolves controllers based on a finite number of simulations. Subsequently, the best candidate is verified using reachability analysis; if the candidate solution does not satisfy the spe…
This letter proposes a learning-based bounded synthesis for a semi-Markov decision process (SMDP) with a linear temporal logic (LTL) specification. In the product of the SMDP and the deterministic $K$-co-Büchi automaton (d$K$cBA) converted from the LTL specification, we learn both the winning region of satisfying the LTL specification and the dynamics therein based on reinforcement learning and Bayesian inference. Then, we synthesize an optimal policy satisfying the following two conditions. (1) It maximizes the probability of reaching the wining region. (2) It minimizes a long-term risk for t…
PerFACT: Motion Policy with LLM-Powered Dataset Synthesis and Fusion Action-Chunking Transformers
J. Mike Walker ’66 Department of Mechanical Engineering, Texas A&M University（J. Mike Walker ’66机械工程系，德克萨斯A&M大学）；Zachry Department of Civil and Environmental Engineering, Texas A&M University（Zachry土木与环境工程系，德克萨斯A&M大学）
We propose a novel controller synthesis involving feedback from pixels, whereby the measurement is a high dimensional signal representing a pixelated image with Red-Green-Blue (RGB) values. The approach neither requires feature extraction, nor object detection, nor visual correspondence. The control policy does not involve the estimation of states or similar latent representations. Instead, tracking is achieved directly in image space, with a model of the reference signal embedded as required by the internal model principle. The reference signal is generated by a neural network with learning-b…
