libalf: The Automata Learning Framework

Bollig, Benedikt; Katoen, Joost-Pieter; Kern, Carsten; Leucker, Martin; Neider, Daniel; Piegdon, David R.

doi:10.1007/978-3-642-14295-6_32

Cited by 75 publications

(45 citation statements)

References 9 publications

(8 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…NL* works in a similar fashion to L*, making it straightforward to substitute into the learning loop shown in Figure 3. We added NL* to our existing implementation from [20], which is based on an extension of PRISM [26] and the libalf [6] learning library. We then compared the performance of the two learning algorithms on a set of five case studies.…”

Section: A Comparison Of Learning Methods: L* Versus Nl*mentioning

confidence: 99%

Automated Learning of Probabilistic Assumptions for Compositional Reasoning

Kwiatkowska

Parker

2011

Fundamental Approaches to Software Engineering

View full text Add to dashboard Cite

Abstract. Probabilistic verification techniques have been applied to the formal modelling and analysis of a wide range of systems, from communication protocols such as Bluetooth, to nanoscale computing devices, to biological cellular processes. In order to tackle the inherent challenge of scalability, compositional approaches to verification are sorely needed. An example is assume-guarantee reasoning, where each component of a system is analysed independently, using assumptions about the other components that it interacts with. We discuss recent developments in the area of automated compositional verification techniques for probabilistic systems. In particular, we describe techniques to automatically generate probabilistic assumptions that can be used as the basis for compositional reasoning. We do so using algorithmic learning techniques, which have already proved to be successful for the generation of assumptions for compositional verification of non-probabilistic systems. We also present recent improvements and extensions to this work and survey some of the promising potential directions for further research in this area.

show abstract

Section: A Comparison Of Learning Methods: L* Versus Nl*mentioning

confidence: 99%

Automated Learning of Probabilistic Assumptions for Compositional Reasoning

Kwiatkowska

Parker

2011

Fundamental Approaches to Software Engineering

View full text Add to dashboard Cite

show abstract

“…Experiments: We built a prototype tool implementing the set-based ICE-learning algorithm for EQDAs, consisting of both a learner and a teacher. The ICE-learner is implemented by extending the classical RPNI algorithm from the libALF library [65]. Given an EQDA conjectured by the learner, the teacher we build converts it to a quantified formula in the APF [61] or decidable Strand for lists [62], and uses a constraint solver to check adequacy of invariants.…”

Section: Programmentioning

confidence: 99%

ICE: A Robust Framework for Learning Invariants

Garg

Löding

Madhusudan

et al. 2014

Computer Aided Verification

Self Cite

165

204

View full text Add to dashboard Cite

Abstract. We introduce ICE, a robust learning paradigm for synthesizing invariants, that learns using examples, counter-examples, and implications, and show that it admits honest teachers and strongly convergent mechanisms for invariant synthesis. We observe that existing algorithms for black-box abstract interpretation can be interpreted as ICE-learning algorithms. We develop new strongly convergent ICE-learning algorithms for two domains, one for learning Boolean combinations of numerical invariants for scalar variables and one for quantified invariants for arrays and dynamic lists. We implement these ICE-learning algorithms in a verification tool and show they are robust, practical, and efficient.

show abstract

“…The queries executed during the learning process are performed by either PRISM [16] or the extension of PRISM developed for multiobjective model checking in [5]. For the implementation of the L* algorithm, we use the libalf [17] learning library. To generate counterexamples, we build adversaries using PRISM and then apply the techniques of [14] using CARMEL which implements Eppstein's algorithm.…”

Section: A Implementation and Case Studiesmentioning

confidence: 99%

Compositional Verification of Probabilistic Systems Using Learning

Kwiatkowska

Parker

2010

2010 Seventh International Conference on the Quantitative Evaluation of Systems

View full text Add to dashboard Cite

Abstract-We present a fully automated technique for compositional verification of probabilistic systems. Our approach builds upon a recently proposed assume-guarantee framework for probabilistic automata, in which assumptions and guarantees are probabilistic safety properties, represented using finite automata. A limitation of this work is that the assumptions need to be created manually. To overcome this, we propose a novel learning technique based on the L* algorithm, which automatically generates probabilistic assumptions using the results of queries executed by a probabilistic model checker. Learnt assumptions either establish satisfaction of the verification problem or are used to generate a probabilistic counterexample that refutes it. In the case where an assumption cannot be generated, lower and upper bounds on the probability of satisfaction are produced. We illustrate the applicability of the approach on a range of case studies.

show abstract

libalf: The Automata Learning Framework

Cited by 75 publications

References 9 publications

Automated Learning of Probabilistic Assumptions for Compositional Reasoning

Automated Learning of Probabilistic Assumptions for Compositional Reasoning

ICE: A Robust Framework for Learning Invariants

Compositional Verification of Probabilistic Systems Using Learning

Contact Info

Product

Resources

About