Logo des Repositoriums
 

Analyzing Code Corpora to Improve the Correctness and Reliability of Programs

dc.contributor.authorPatra, Jibesh
dc.contributor.editorHerrmann, Andrea
dc.date.accessioned2024-04-08T08:43:36Z
dc.date.available2024-04-08T08:43:36Z
dc.date.issued2022
dc.description.abstractThe goal of the dissertation summarized here is to use program analysis and novel learning-based techniques to alleviate some of the challenges faced by developers while ensuring the correctness and reliability of programs. We focus on dynamically typed languages such as JavaScript and Python for their popularity and present six approaches that leverages analysis of code corpora in aiding to solve software engineering problems. We use static analysis to generate new programs, to seed bugs in programs, and to obtain data for training neural models. We present an effective technique called Generalized Tree Reduction algorithm (GTR), to reduce arbitrary test inputs that can be represented as a tree, such as program code, PDF files, and XML documents. The efficiency of input reduction is increased by learning transformations from a corpus of example data.en
dc.identifier.issn0720-8928
dc.identifier.urihttps://dl.gi.de/handle/20.500.12116/43851
dc.language.isoen
dc.pubPlaceBonn
dc.publisherGesellschaft für Informatik e.V.
dc.relation.ispartofSoftwaretechnik-Trends Band 42, Heft 2
dc.relation.ispartofseriesSoftwaretechnik-Trends
dc.subjectcode analysis
dc.subjectprogram analysis
dc.subjectlearning-based techniques
dc.subjectcorrectness
dc.subjectreliability
dc.subjectJavaScript
dc.subjectPython
dc.titleAnalyzing Code Corpora to Improve the Correctness and Reliability of Programsen
dc.typeText/Journal Article
mci.reference.pages18-19

Dateien

Originalbündel
1 - 1 von 1
Vorschaubild nicht verfügbar
Name:
KurzfassungPatra.pdf
Größe:
140.13 KB
Format:
Adobe Portable Document Format