Inferring Computational State Machine Models from Program Executions

Abstract

Abstract-The challenge of inferring state machines from log data or execution traces is well-established, and has led to the development of several powerful techniques. Current approaches tend to focus on the inference of conventional finite state machines or, in few cases, state machines with guards. However, these machines are ultimately only partial, because they fail to model how any underlying variables are computed during the course of an execution; they are not computational. In this paper we introduce a technique based upon Genetic Programming to infer these data transformation functions, which in turn render inferred automata fully computational. Instead of merely determining whether or not a sequence is possible, they can be simulated, and be used to compute the variable values throughout the course of an execution. We demonstrate the approach by using a CrossValidation study to reverse-engineer complete (computational) EFSMs from traces of established implementations.

Learning Concise Models from Long Execution Traces

“…Realizing abstract and understandable models requires the user to know what abstract conditions are significant in the system evolution, for example 'the FIFO became more than half-full', and to instrument the system to record such conditions. Extensions of these traditional algortihms [6], [7] generate Extended Finite-State Machines (EFSMs) with syntactically-expressed predicates on transitions edges, but require a substantial amount of additional trace samples from simulation of the learned model for predicate inference.…”

Active Learning of Abstract System Models from Traces using Model Checking [Extended]

“…We performed a set of experiments to check if random sampling is sufficient to learn abstractions that admit all behaviours. A million randomly sampled inputs are used to execute each 2 The dataset includes another implementation of this benchmark with similar results. Due to space constraints, we present the results for only one of them.…”

Data and Analysis Code for GP EFSM Inference

“…This artifact captures the workflow that we adopted for our experimental evaluation in our ICSME paper on inferring state transition functions durfing EFSM inference [1]. To summarise, the paper uses Genetic Programming to infer data transformations, to enable the inference of fully 'computational' extended finite state machine models.…”

“…• Samples of the data generated by MINT The artifact is contained in a Git repository available online 1 . To obtain a copy, clone it using git clone.…”

“…The Makefile automates the process of building MINT, including fetching its dependencies. It also includes pre-scripted goals (makefile targets) that will generate results for the case studies 1 https://bitbucket.org/mathew_hall/icsme2016-data included in the paper. When run, the makefile will produce a JAR file for MINT (EFSMTool.jar) in the experiment directory.…”