Jean-François Raskin

Lorenzo Clemente, Jean-François Raskin

The beyond worst-case threshold problem (BWC), recently introduced by Bruyère et al., asks given a quantitative game graph for the synthesis of a strategy that i) enforces some minimal level of performance against any adversary, and ii) achieves a good expectation against a stochastic model of the adversary. They solved the BWC problem for finite-memory strategies and unidimensional mean-payoff objectives and they showed membership of the problem in NP$\cap$coNP. They also noted that infinite-memory strategies are more powerful than finite-memory ones, but the respective threshold problem was left open. We extend these results in several directions. First, we consider multidimensional mean-payoff objectives. Second, we study both finite-memory and infinite-memory strategies. We show that the multidimensional BWC problem is coNP-complete in both cases. Third, in the special case when the worst-case objective is unidimensional (but the expectation objective is still multidimensional) we show that the complexity decreases to NP$\cap$coNP. This solves the infinite-memory threshold problem left open by Bruyère et al., and this complexity cannot be improved without improving the currently known complexity of classical mean-payoff games. Finally, we introduce a natural relaxation of the BWC problem, the beyond almost-sure threshold problem (BAS), which asks for the synthesis of a strategy that ensures some minimal level of performance with probability one and a good expectation against the stochastic model of the adversary. We show that the multidimensional BAS threshold problem is solvable in P.

Peter Bulychev, Franck Cassez, Alexandre David et al.

We consider the problem of controller synthesis under imperfect information in a setting where there is a set of available observable predicates equipped with a cost function. The problem that we address is the computation of a subset of predicates sufficient for control and whose cost is minimal. Our solution avoids a full exploration of all possible subsets of predicates and reuses some information between different iterations. We apply our approach to timed systems. We have developed a tool prototype and analyze the performance of our optimization algorithm on two case studies.

Gavin Rens, Wen-Chi Yang, Jean-François Raskin et al.

We propose a framework for learning a fragment of probabilistic computation tree logic (pCTL) formulae from a set of states that are labeled as safe or unsafe. We work in a relational setting and combine ideas from relational Markov Decision Processes with pCTL model-checking. More specifically, we assume that there is an unknown relational pCTL target formula that is satisfied by only safe states, and has a horizon of maximum $k$ steps and a threshold probability $α$. The task then consists of learning this unknown formula from states that are labeled as safe or unsafe by a domain expert. We apply principles of relational learning to induce a pCTL formula that is satisfied by all safe states and none of the unsafe ones. This formula can then be used as a safety specification for this domain, so that the system can avoid getting into dangerous situations in future. Following relational learning principles, we introduce a candidate formula generation process, as well as a method for deciding which candidate formula is a satisfactory specification for the given labeled states. The cases where the expert knows and does not know the system policy are treated, however, much of the learning process is the same for both cases. We evaluate our approach on a synthetic relational domain.

Debraj Chakraborty, Damien Busatto-Gaston, Jean-François Raskin et al.

We study how to efficiently combine formal methods, Monte Carlo Tree Search (MCTS), and deep learning in order to produce high-quality receding horizon policies in large Markov Decision processes (MDPs). In particular, we use model-checking techniques to guide the MCTS algorithm in order to generate offline samples of high-quality decisions on a representative set of states of the MDP. Those samples can then be used to train a neural network that imitates the policy used to generate them. This neural network can either be used as a guide on a lower-latency MCTS online search, or alternatively be used as a full-fledged policy when minimal latency is required. We use statistical model checking to detect when additional samples are needed and to focus those additional samples on configurations where the learnt neural network policy differs from the (computationally-expensive) offline policy. We illustrate the use of our method on MDPs that model the Frozen Lake and Pac-Man environments -- two popular benchmarks to evaluate reinforcement-learning algorithms.

Wen-Chi Yang, Jean-François Raskin, Luc De Raedt

Probabilistic model checking has been developed for verifying systems that have stochastic and nondeterministic behavior. Given a probabilistic system, a probabilistic model checker takes a property and checks whether or not the property holds in that system. For this reason, probabilistic model checking provide rigorous guarantees. So far, however, probabilistic model checking has focused on propositional models where a state is represented by a symbol. On the other hand, it is commonly required to make relational abstractions in planning and reinforcement learning. Various frameworks handle relational domains, for instance, STRIPS planning and relational Markov Decision Processes. Using propositional model checking in relational settings requires one to ground the model, which leads to the well known state explosion problem and intractability. We present pCTL-REBEL, a lifted model checking approach for verifying pCTL properties of relational MDPs. It extends REBEL, a relational model-based reinforcement learning technique, toward relational pCTL model checking. PCTL-REBEL is lifted, which means that rather than grounding, the model exploits symmetries to reason about a group of objects as a whole at the relational level. Theoretically, we show that pCTL model checking is decidable for relational MDPs that have a possibly infinite domain, provided that the states have a bounded size. Practically, we contribute algorithms and an implementation of lifted relational model checking, and we show that the lifted approach improves the scalability of the model checking approach.

Raphaël Berthon, Adrien Boiret, Guillermo A. Perez et al.

Active learning is a setting in which a student queries a teacher, through membership and equivalence queries, in order to learn a language. Performance on these algorithms is often measured in the number of queries required to learn a target, with an emphasis on costly equivalence queries. In graybox learning, the learning process is accelerated by foreknowledge of some information on the target. Here, we consider graybox active learning of subsequential string transducers, where a regular overapproximation of the domain is known by the student. We show that there exists an algorithm using string equation solvers that uses this knowledge to learn subsequential string transducers with a better guarantee on the required number of equivalence queries than classical active learning.

Gavin Rens, Jean-François Raskin, Raphaël Reynouad et al.

There are situations in which an agent should receive rewards only after having accomplished a series of previous tasks, that is, rewards are non-Markovian. One natural and quite general way to represent history-dependent rewards is via a Mealy machine, a finite state automaton that produces output sequences from input sequences. In our formal setting, we consider a Markov decision process (MDP) that models the dynamics of the environment in which the agent evolves and a Mealy machine synchronized with this MDP to formalize the non-Markovian reward function. While the MDP is known by the agent, the reward function is unknown to the agent and must be learned. Our approach to overcome this challenge is to use Angluin's $L^*$ active learning algorithm to learn a Mealy machine representing the underlying non-Markovian reward machine (MRM). Formal methods are used to determine the optimal strategy for answering so-called membership queries posed by $L^*$. Moreover, we prove that the expected reward achieved will eventually be at least as much as a given, reasonable value provided by a domain expert. We evaluate our framework on three problems. The results show that using $L^*$ to learn an MRM in a non-Markovian reward decision process is effective.

Damien Busatto-Gaston, Debraj Chakraborty, Shibashis Guha et al.

In this paper, we investigate the combination of synthesis, model-based learning, and online sampling techniques to obtain safe and near-optimal schedulers for a preemptible task scheduling problem. Our algorithms can handle Markov decision processes (MDPs) that have 1020 states and beyond which cannot be handled with state-of-the art probabilistic model-checkers. We provide probably approximately correct (PAC) guarantees for learning the model. Additionally, we extend Monte-Carlo tree search with advice, computed using safety games or obtained using the earliest-deadline-first scheduler, to safely explore the learned model online. Finally, we implemented and compared our algorithms empirically against shielded deep Q-learning on large task systems.

Raphaël Berthon, Shibashis Guha, Jean-François Raskin

In this paper, we consider algorithms to decide the existence of strategies in MDPs for Boolean combinations of objectives. These objectives are omega-regular properties that need to be enforced either surely, almost surely, existentially, or with non-zero probability. In this setting, relevant strategies are randomized infinite memory strategies: both infinite memory and randomization may be needed to play optimally. We provide algorithms to solve the general case of Boolean combinations and we also investigate relevant subcases. We further report on complexity bounds for these problems.

Gavin Rens, Jean-François Raskin

There are situations in which an agent should receive rewards only after having accomplished a series of previous tasks. In other words, the reward that the agent receives is non-Markovian. One natural and quite general way to represent history-dependent rewards is via a Mealy machine; a finite state automaton that produces output sequences (rewards in our case) from input sequences (state/action observations in our case). In our formal setting, we consider a Markov decision process (MDP) that models the dynamic of the environment in which the agent evolves and a Mealy machine synchronised with this MDP to formalise the non-Markovian reward function. While the MDP is known by the agent, the reward function is unknown from the agent and must be learnt. Learning non-Markov reward functions is a challenge. Our approach to overcome this challenging problem is a careful combination of the Angluin's L* active learning algorithm to learn finite automata, testing techniques for establishing conformance of finite model hypothesis and optimisation techniques for computing optimal strategies in Markovian (immediate) reward MDPs. We also show how our framework can be combined with classical heuristics such as Monte Carlo Tree Search. We illustrate our algorithms and a preliminary implementation on two typical examples for AI.

Jan Křetínský, Guillermo A. Pérez, Jean-François Raskin

We formalize the problem of maximizing the mean-payoff value with high probability while satisfying a parity objective in a Markov decision process (MDP) with unknown probabilistic transition function and unknown reward function. Assuming the support of the unknown transition function and a lower bound on the minimal transition probability are known in advance, we show that in MDPs consisting of a single end component, two combinations of guarantees on the parity and mean-payoff objectives can be achieved depending on how much memory one is willing to use. (i) For all $ε$ and $γ$ we can construct an online-learning finite-memory strategy that almost-surely satisfies the parity objective and which achieves an $ε$-optimal mean payoff with probability at least $1 - γ$. (ii) Alternatively, for all $ε$ and $γ$ there exists an online-learning infinite-memory strategy that satisfies the parity objective surely and which achieves an $ε$-optimal mean payoff with probability at least $1 - γ$. We extend the above results to MDPs consisting of more than one end component in a natural way. Finally, we show that the aforementioned guarantees are tight, i.e. there are MDPs for which stronger combinations of the guarantees cannot be ensured.

Raphaël Berthon, Mickael Randour, Jean-François Raskin

The beyond worst-case synthesis problem was introduced recently by Bruyère et al. [BFRR14]: it aims at building system controllers that provide strict worst-case performance guarantees against an antagonistic environment while ensuring higher expected performance against a stochastic model of the environment. Our work extends the framework of [BFRR14] and follow-up papers, which focused on quantitative objectives, by addressing the case of $ω$-regular conditions encoded as parity objectives, a natural way to represent functional requirements of systems. We build strategies that satisfy a main parity objective on all plays, while ensuring a secondary one with sufficient probability. This setting raises new challenges in comparison to quantitative objectives, as one cannot easily mix different strategies without endangering the functional properties of the system. We establish that, for all variants of this problem, deciding the existence of a strategy lies in ${\sf NP} \cap {\sf coNP}$, the same complexity class as classical parity games. Hence, our framework provides additional modeling power while staying in the same complexity class. [BFRR14] Véronique Bruyère, Emmanuel Filiot, Mickael Randour, and Jean-François Raskin. Meet your expectations with guarantees: Beyond worst-case synthesis in quantitative games. In Ernst W. Mayr and Natacha Portier, editors, 31st International Symposium on Theoretical Aspects of Computer Science, STACS 2014, March 5-8, 2014, Lyon, France, volume 25 of LIPIcs, pages 199-213. Schloss Dagstuhl - Leibniz - Zentrum fuer Informatik, 2014.

Krishnendu Chatterjee, Petr Novotný, Guillermo A. Pérez et al.

A standard objective in partially-observable Markov decision processes (POMDPs) is to find a policy that maximizes the expected discounted-sum payoff. However, such policies may still permit unlikely but highly undesirable outcomes, which is problematic especially in safety-critical applications. Recently, there has been a surge of interest in POMDPs where the goal is to maximize the probability to ensure that the payoff is at least a given threshold, but these approaches do not consider any optimization beyond satisfying this threshold constraint. In this work we go beyond both the "expectation" and "threshold" approaches and consider a "guaranteed payoff optimization (GPO)" problem for POMDPs, where we are given a threshold $t$ and the objective is to find a policy $σ$ such that a) each possible outcome of $σ$ yields a discounted-sum payoff of at least $t$, and b) the expected discounted-sum payoff of $σ$ is optimal (or near-optimal) among all policies satisfying a). We present a practical approach to tackle the GPO problem and evaluate it on standard POMDP benchmarks.

Jean-François Raskin, Ocan Sankur

We introduce Multi-Environment Markov Decision Processes (MEMDPs) which are MDPs with a set of probabilistic transition functions. The goal in a MEMDP is to synthesize a single controller with guaranteed performances against all environments even though the environment is unknown a priori. While MEMDPs can be seen as a special class of partially observable MDPs, we show that several verification problems that are undecidable for partially observable MDPs, are decidable for MEMDPs and sometimes have even efficient solutions.