Inferring Mealy Machines

Shahbaz, Muzammil; Groz, Roland

doi:10.1007/978-3-642-05089-3_14

Cited by 146 publications

(127 citation statements)

References 11 publications

Supporting

Mentioning

119

Contrasting

Unclassified

Order By: Relevance

“…Applied research into linking testing and inference has generally arisen from work that takes a state machine-centric view of software [13], [17], [20], [19], [11]. There are numerous well-established state machine inference algorithms and state machine testing algorithms, which makes the two areas particularly complementary in this respect.…”

Section: Related Workmentioning

confidence: 99%

“…However, efforts to investigate this idea have been overwhelmingly geared towards the relatively narrow class of sequential systems that can be modelled as state machines [13], [17], [20], [19], [11] (possibly because such models are associated with an extensive range of state-machine testing techniques). Similar work on data-driven programs, of the sort that can be executed by a single command on the command-line, remains limited, and the authors are not aware of any openly available implementations (that do not require the source-code of the program under test) to facilitate this experimentation.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Black-Box Test Generation from Inferred Models

Papadopoulos

Walkinshaw

2015

2015 IEEE/ACM 4th International Workshop on Realizing Artificial Intelligence Synergies in Software Engineering

View full text Add to dashboard Cite

Abstract-Automatically generating test inputs for components without source code (are 'black-box') and specification is challenging. One particularly interesting solution to this problem is to use Machine Learning algorithms to infer testable models from program executions in an iterative cycle. Although the idea has been around for over 30 years, there is little empirical information to inform the choice of suitable learning algorithms, or to show how good the resulting test sets are. This paper presents an openly available framework to facilitate experimentation in this area, and provides a proof-of-concept inference-driven testing framework, along with evidence of the efficacy of its test sets on three programs.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Black-Box Test Generation from Inferred Models

Papadopoulos

Walkinshaw

2015

2015 IEEE/ACM 4th International Workshop on Realizing Artificial Intelligence Synergies in Software Engineering

View full text Add to dashboard Cite

show abstract

“…Various methods have been proposed to enhance Angluin's algorithm to infer a Finite State Machine (FSM) with inputs and outputs [2]. In [3], the model was enhanced again to consider input/output parameters with finite (or infinite) domains; this is called Parameterized Finite State Machine (PFSM).…”

Section: A Extended Modelsmentioning

confidence: 99%

Automatic Generation of Test Drivers for Model Inference of Web Applications

Hossen¹,

Groz²,

Oriat³

et al. 2013

2013 IEEE Sixth International Conference on Software Testing, Verification and Validation Workshops

Self Cite

View full text Add to dashboard Cite

Abstract-In the "Internet of Services" (IoS) vision of the Internet, applications are developed as services using the web standards. Model-based testing combined with active model inference is one of the methods to test the applications pretty automatically, in particular to look for vulnerabilities. But one part still needs to be written manually, the test driver. It contains an abstraction of the real application and the methods to interact with the system at abstract and concrete level. We propose a generic abstraction of the web applications and an approach to generate the corresponding test driver automatically using a crawler to identify the needed information. Keywords: model inference, test driver generation, crawling method, web application I. MODEL INFERENCEIn the IoS (Internet of services) vision, information is provided to the user as a service. Applications which were developed as a single block are now built by assembling components provided by various service providers. The complexity of this kind of application, its architecture and the fact that most of the time a formal model is not available, make model-based testing and verification difficult and timeconsuming.However, in a testing context, partial models can actually be retrieved, even for black box components, from observation gathered in the testing process. Model inference -that is, methods to automatically derive models from observation of the behaviors -can seriously help to test services by providing models of the components automatically.Model inference is commonly divided into two categories: active and passive inference. Passive learning learns only from a given set of observations whereas active learning is free to interact with the tested system. In the case of web applications, we are free to interact with it through its web interface. As active learning converges faster, has better results than passive learning and is usable for website, it is the best choice. Angluin's well-known algorithm (L*) [1] is in the active learning category and can infer a deterministic finite automaton (DFA). This algorithm has often been used as a base for developing enhanced model inference algorithms. A. Extended ModelsWeb-based applications use a lot of structured messages to communicate. This makes application behavior mainly dependent on the inputs and outputs (I/O) and not only on the states. Basing the alphabet on the I/O value pairs would produce a large number of transitions and states and it would be highly inefficient. In order to use adapted models, we must have an efficient model which considers inputs and outputs, which may be structured.

show abstract

“…Thus, the number of membership queries for treating counterexamples can be estimated by m · (|Q| + |Σ 2 ) , cf. [24,25]. Accordingly, the total number of required membership queries is at most |Σ…”

Section: Sketch Of Proofmentioning

confidence: 99%

Automata Learning with Automated Alphabet Abstraction Refinement

Howar

Steffen

Merten

2011

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstraction is the key when learning behavioral models of realistic systems, but also the cause of a major problem: the introduction of non-determinism. In this paper, we introduce a method for refining a given abstraction to automatically regain a deterministic behavior onthe-fly during the learning process. Thus the control over abstraction becomes part of the learning process, with the effect that detected nondeterminism does not lead to failure, but to a dynamic alphabet abstraction refinement. Like automata learning itself, this method in general is neither sound nor complete, but it also enjoys similar convergence properties even for infinite systems as long as the concrete system itself behaves deterministically, as illustrated along a concrete example.

show abstract

Inferring Mealy Machines

Cited by 146 publications

References 11 publications

Black-Box Test Generation from Inferred Models

Black-Box Test Generation from Inferred Models

Automatic Generation of Test Drivers for Model Inference of Web Applications

Automata Learning with Automated Alphabet Abstraction Refinement

Contact Info

Product

Resources

About