Monday, May 6, 2024, 11am

Companies leverage personalization techniques to tailor user experiences. Personalization ap- pears in search engines and online stores, which include salutations and statistically learned correlations over search-, browsing- and purchase-histories. However, users have a wider variety of substantive, domain-specific preferences that influence their choices when they use directory services, and these have largely been overlooked or ignored. Specifically, users have preferences about what they are looking for, and are using services with varying levels of personalization to aid in discovering their things of interest. In the realm of requirements engineering, requirement analysts endeavor to gather, comprehend, and prioritize requirements, with an important fo- cus on stakeholder preferences and needs, employing diverse requirement elicitation techniques. Advances in Machine Learning (ML) and Natural Language Processing (NLP) have revolutionized the way people understand and interact with natural language, and opened up new opportunities to enhance and automate various facets of requirements engineering, including stakeholder preference elicitation.

The thesis aims to explore the potential of NLP techniques to enhance and automate stake- holder preference elicitation practices. Specifically, we demonstrate the following NLP-assisted preference elicitation methods: 1) we study the efficacy of extracting domain knowledge from user-authored scenarios using typed dependency parsing techniques, and from word embed- dings using a BERT-based Masked Language Model (MLM); 2) we research on how stakeholder preferences are expressed in text scenarios, whether elicited preferences represent missing re- quirements in existing systems, and how we may use named entity recognition techniques to build classifiers to label preference words in scenarios and link them to form preference phrases; 3) we build a tool to support and improve preference elicitation practices in interviews by using various NLP techniques, including MLM, transformers, speech-to-text transcription, and part- of-speech tagging. We also propose to build on and improve existing research findings with Large Language Models (LLM). The expected outcome of the thesis is to shed light on how NLP can be integrated into existing requirement elicitation practices to enhance and enrich them, while advancing our understanding about how stakeholder preference may be elicited more effectively and comprehensively.

Thesis Committee
Travis Breaux (Chair)
Christian Kästner
Bogdan Vasilescu
Fabiano Dalpiaz (Utrecht University)

Additional Information

In Person and Zoom Participation. See announcement.

Event Type: Thesis Proposals
Room Number: In Person and Virtual - ET
Building: TCS Hall 310 and Zoom
Speaker's Name: YUCHEN SHEN
Speaker Website: se-phd.s3d.cmu.edu…
Speaker's Professional Title: Ph.D. Student, Ph.D. Program in Software Engineering, Software and Societal Systems Seminar, Carnegie Mellon University
Talk Title: NLP-Assisted Preference Elicitation in Requirements Engineering
Event Poster Title: Poster
Event Poster URL: www.cs.cmu.edu…
For More Information: cherold@cs.cmu.edu | aroudebu@andrew.cmu.edu
Affiliations: Software and Societal Systems Department (S3D)
Organization(s): School of Computer Science