Testing probabilistic and hybrid black-box systems (Delivered in English)
- LecturerDr. Kohei Suenaga (Kyoto University)
Host: Shin-Cheng Mu - Time2024-06-26 (Wed.) 10:00 ~ 12:00
- LocationAuditorium 101 at IIS new Building
Abstract
We introduce a novel methodology for testing stochastic black-box systems. Our approach enhances the established black-box checking (BBC) technique to address stochastic behavior.
Traditional BBC primarily involves iteratively identifying an input that breaches the system’s specifications by executing the following three phases: the learning phase to construct an automaton approximating the black box’s behavior, the synthesis phase to identify a candidate counterexample from the learned automaton, and the validation phase to validate the obtained candidate counterexample and the learned automaton against the original black-box system.
Our method, ProbBBC, refines the conventional BBC approach by (1) employing an active Markov Decision Process (MDP) learning method during the learning phase, (2) incorporating probabilistic model checking in the synthesis phase, and (3) applying statistical hypothesis testing in the validation phase.
ProbBBC uniquely integrates these techniques rather than merely substituting each method in the traditional BBC; for instance, the statistical hypothesis testing and the MDP learning procedure exchange information regarding the black-box system’s observation with one another. The experiment results suggest that ProbBBC outperforms an existing method, especially for systems with limited observation.
If time allows, we also introduce our recent work on hybrid automata learning. We present an algorithm to learn a nonlinear hybrid automaton (HA) that approximates a black-box hybrid system (HS) from a set of input--output traces generated by the HS. Our method is novel in handling (1) both exogenous and endogenous HS and (2) HA with reset associated with each transition. We applied our algorithm to various benchmarks and confirmed its effectiveness.
Traditional BBC primarily involves iteratively identifying an input that breaches the system’s specifications by executing the following three phases: the learning phase to construct an automaton approximating the black box’s behavior, the synthesis phase to identify a candidate counterexample from the learned automaton, and the validation phase to validate the obtained candidate counterexample and the learned automaton against the original black-box system.
Our method, ProbBBC, refines the conventional BBC approach by (1) employing an active Markov Decision Process (MDP) learning method during the learning phase, (2) incorporating probabilistic model checking in the synthesis phase, and (3) applying statistical hypothesis testing in the validation phase.
ProbBBC uniquely integrates these techniques rather than merely substituting each method in the traditional BBC; for instance, the statistical hypothesis testing and the MDP learning procedure exchange information regarding the black-box system’s observation with one another. The experiment results suggest that ProbBBC outperforms an existing method, especially for systems with limited observation.
If time allows, we also introduce our recent work on hybrid automata learning. We present an algorithm to learn a nonlinear hybrid automaton (HA) that approximates a black-box hybrid system (HS) from a set of input--output traces generated by the HS. Our method is novel in handling (1) both exogenous and endogenous HS and (2) HA with reset associated with each transition. We applied our algorithm to various benchmarks and confirmed its effectiveness.