Page d'accueil // SnT // News & E... // SnT Scientist Wins University Thesis Award

SnT Scientist Wins University Thesis Award

twitter linkedin facebook google+ email this page
Publié le mardi, 11 avril 2017

Dr. Chetan Arora of the University of Luxembourg's Interdisciplinary Centre for Security Reliability and Trust has been awarded the University's 'Best Computer Science and Communications PhD Thesis 2016', for his thesis titled 'Automated Analysis of Natural-Language Requirements Using Natural Language Processing'.

Requirements Engineering is the systematic process of identifying and defining the characteristics and capabilities of a proposed software system. The resulting requirements, put together with input from multiple stakeholders and in alignment with stringent regulations, are crucial documents for any development project; they typically constitute the basis for contractual agreements and are intricately linked with issues of cost and responsibility.

As with any complex technical documents written by multiple people in 'natural' (or human) language, requirements present a high degree of risk in terms of the potential for ambiguity, incompleteness and inconsistency. These issues, if not detected early enough, may ripple thorough later stages of software development, leading to (in the best case) multiple-fold cost and effort to fix these issues, or (in the worst case) project failures. Dr Arora's thesis uses Natural Language Processing (NLP) to propose a range of techniques to automate the complex and laborious quality assurance and analysis tasks necessary to resolve these issues at early stages.

Dr. Arora’s PhD research developed a number of Requirements Engineering solutions, including the ability to automatically identify and extract the key semantic information from requirements documents; to identify the impact of changes to individual requirements on the requirements as a whole; to automate the process of ensuring that requirements conform to specified templates; and to allow for the extraction of models from a set of requirements, providing the basis for sound and high quality requirements.

Dr. Arora, wrote his thesis under the supervision of Prof. Dr. Lionel Briand and Dr. Mehrdad Sabetzadeh of SnT’s Software Verification and Validation Research Group, where he now works as a Research Associate. The solutions presented were developed and tested in close collaboration with the leading satellite solutions provider, SES.

'Automated Analysis of Natural-Language Requirements Using Natural Language Processing' is available as an Open Access PDF here.