Modelling program verification tools for software engineers

Lathouwers, Sophie; Zaytsev, Vadim

doi:10.1145/3550355.3552426

Cited by 4 publications

(5 citation statements)

References 51 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Direct browsing in the Internet as well as perusal of notable reviews, such as the report of J. Hutchinson, considered above, survey of E. Jakumeit et al [40] and investigation of S. Lathouwers and V. Zaytsev [41], covering two decades of MDE development let hardly find a tool dedicated to teaching of MDE. E. Jakumeit et al [22] have thoroughly compared 13 transformation tools selected to represent the state of the art in the worlds of model transformation and graph rewriting.…”

Section: Discussionmentioning

confidence: 99%

“…S. Lathouwers and V. Zaytsev [41] have investigated the domain of program verification tools, and presented a concise megamodel splitting the domain into 7 levels. No one of the tools attributed to the domain was explicitly dedicated for teaching.…”

Section: Discussionmentioning

confidence: 99%

“…Due to its checking capabilities Procrust can be attributed to the level PV1 of the megamodel proposed in [41], playing a role of the "model solver".…”

Section: Ar Tool Functionalitymentioning

confidence: 99%

See 2 more Smart Citations

Initial training in design of computer programs using UML

Andreyanov,

Gagarin

2023

P Sci Edu

View full text Add to dashboard Cite

Introduction. Design is commonly acknowledged as a key process in the life cycle of computer programs and softwareintensive systems. The process efficaciously reveals creative capabilities of the author programmer and predetermines the basic advantages as well as shortcomings of the resulting program product, service or system. Design automation and other kinds of its support have been emerged simultaneously with onset of the software industry. From the beginning of the century the MDA (Model Driven Architecture) approach grew famous, offering to comprehend a design of a program as a process of sequential transformation of the program’s models, represented in UML or in other languages related to UML. Without becoming dominant, the MDA approach successfully evolves inside of a broader approach called MDE (Model Driven Engineering). Mastering UML and understanding of its role in MDE appears to be a prerequisite for a seamless entering of a professional programmer into domain of engineering of program systems. This provision justifies an increased attention to methods of UML teaching and to peculiarities of UML learning, and, moreover, to the professional education in the area of programs and program systems design technologies. This paper is aimed to draw attentions to challenges of the initial training in UML and tackle the challenges contributing a methodology and tools for sake of fostering the quality of education in program design. Materials and methods. The research is based on the university syllabi, the author’s lecture notes on software engineering, reports on laboratory works and course papers, protocols of experimental and sample runs of the developed programs, scientific publications, including pedagogical and engineering periodicals, conference materials, Internet resources, such as sites of the ACM, the OMG and others. The following research methods were used: system analysis of the own experience in program design as well as experience remarkable in student’s works; theoretical conceptual modelling of the program design including modelling of design scenarios; a trial implementation of the developed algorithms. Research results. A conceptual cognitive model of design as of a process during program development is proposed, highlighting the designer’s behavior when dealing with project artefacts. Feasible scenarios of analytic and advisory intervention into the design process are specified, using the model. A sample architecture of a program tool suitable for such an intervention is proposed as well as algorithms that enable online critical evaluation of the project’s current state, prompting feasible tracks of the project’s evolvement, revealing designer’s ideas and pointing some UML misuse. The algorithms are built in a program tool named Procrust that implements the proposed architecture. Trials of Procrust show that it basically fits to primary training of software design using UML and, in particular, to a training in the MDA/MDE methodology. Conclusion. The research results enable to set up an environment for comfortable and efficient program design using UML due to employing the proposed analytic and advisory support of the design process. Algorithms of the support and supporting program tools are targeted at initial design training on curricula of higher school and secondary vocational education. The algorithms foster in-depth study of UML and acquiring of advanced software design skills.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Initial training in design of computer programs using UML

Andreyanov,

Gagarin

2023

P Sci Edu

View full text Add to dashboard Cite

show abstract

“…Program synthesis is the task of finding a program for a given specification, such as a natural language description or input-output examples (Gulwani, Polozov, and Singh 2017). By distilling neural network policies into programmatic policies, we are able to verify the program for correctness and use traditional formal verification tools (Lathouwers and Zaytsev 2022) to analyze the behavior and edge cases of synthesized programs. Another benefit of distilling policies into programs is that software developers can adapt the policy to their own needs, which makes it easy to further improve the programmatic policies or adapt them to other scenarios (Trivedi et al 2021).…”

Section: Introductionmentioning

confidence: 99%

Learning of Generalizable and Interpretable Knowledge in Grid-Based Reinforcement Learning Environments

Eberhardinger,

Maucher,

Maghsudi

2023

AIIDE

View full text Add to dashboard Cite

Understanding the interactions of agents trained with deep reinforcement learning is crucial for deploying agents in games or the real world. In the former, unreasonable actions confuse players. In the latter, that effect is even more significant, as unexpected behavior cause accidents with potentially grave and long-lasting consequences for the involved individuals. In this work, we propose using program synthesis to imitate reinforcement learning policies after seeing a trajectory of the action sequence. Programs have the advantage that they are inherently interpretable and verifiable for correctness. We adapt the state-of-the-art program synthesis system DreamCoder for learning concepts in grid-based environments, specifically, a navigation task and two miniature versions of Atari games, Space Invaders and Asterix. By inspecting the generated libraries, we can make inferences about the concepts the black-box agent has learned and better understand the agent's behavior. We achieve the same by visualizing the agent's decision-making process for the imitated sequences. We evaluate our approach with different types of program synthesizers based on a search-only method, a neural-guided search, and a language model fine-tuned on code.

show abstract

“…This chapter is based on the following publication: The text is mostly based on [LZ22] but the numbers and future work have been updated to reflect the current situation of the project, as in [LLZ].…”

Section: Contributionsmentioning

confidence: 99%

Exploring annotations for deductive verification

Lathouwers

View full text Add to dashboard Cite

Ten slotte, wil ik Joran bedanken. Je hebt me door de zwaarste momenten gesteund en dat zal ik nooit vergeten. Je herinnert me eraan dat het goed is om af en toe (of eigenlijk altijd) (een beetje) gek te doen en om vol te gaan voor wat ik wil. Je blijft altijd kritisch, en ook al zit ik daar niet altijd op te wachten, heb je stiekem vaak wel gelijk. Dank je wel voor al je steun de afgelopen jaren. Sophie

show abstract

Modelling program verification tools for software engineers

Cited by 4 publications

References 51 publications

Initial training in design of computer programs using UML

Initial training in design of computer programs using UML

Learning of Generalizable and Interpretable Knowledge in Grid-Based Reinforcement Learning Environments

Exploring annotations for deductive verification

Contact Info

Product

Resources

About