Natural Language Processing for Software Engineering

Description

This research area explores the application of state-of-the-art natural language processing methods to support, enhance and extend traditional software engineering and requirements engineering tasks. We focus on multiple areas within the software development lifecycle, including requirements specification, code generation and feedback analysis, among others. More recently, we are exploring how large language models can be leveraged to improve such tasks and generate new use cases.

The research aims at contributing both from a scientific and technical perspective, building open-source tools to be exploited by both researchers and practitioners.

Team

Quim Motger

UPC (Principal Investigator)

Xavier Franch

UPC

Jordi Marco

UPC

Marc Oriol

UPC

Max Tiessler

UPC/TU Wien - MSc

Projects Overview

Title	Project type	Goal
HIVEMIND (2025-2027)	Horizon Europe	Promote responsible software engineering practices that accelerate all stages of the software development lifecycle, leveraging novel AI and data technologies to design and develop an adaptive LLM-based multi-agent framework.
NLP4RE (2021-2024)	Non-funded	Analyse the application of NLP techniques to enhance their effectiveness, efficiency and adoption in the context of requirements specification and validation tasks.
AI4Software (2023-2025)	REDES (AEI)	Fostering collaboration between national research groups to enhance the application of AI methods within the software development lifecycle
OpenReq (2017-2020)	Horizon 2020	Build an intelligent recommendation and decision system for community-driven requirements engineering.

Outcomes/Main Contributions

Research:

Requirements Traceability
- Requirements Dependency Extraction by Integrating Active Learning with Ontology-Based Retrieval [doi]
- Improved Management of Issue Dependencies in Issue Trackers of Large Collaborative Projects [doi]
- NLP-based Relation Extraction Methods in Requirements Engineering [pre-print]
Feedback Analysis
- T-FREX: A Transformer-based Feature Extraction Method from Mobile App Reviews [doi]
- Leveraging Large Language Models for Mobile App Review Feature Extraction [pre-print]
Competition and Market Analysis
- Unveiling Competition Dynamics in Mobile App Markets Through User Reviews [doi]
Conversational Agents (Chatbots)
- Software-Based Dialogue Systems: Survey, Taxonomy, and Challenges [doi]
- Adaptive Task-Oriented Chatbots Using Feature-Based Knowledge Bases [doi]

Technology Transfer:

Requirements Traceability
- OpenReq-DD: A Requirements Dependency Detection Tool [paper, GitHub]
- ORSI: A Similarity Detection Service for Requirements [GitHub]
Requirements Analysis and Knowledge Base Design
- App Scanner Service: a tool to support data collection from heterogeneous mobile application repositories [GitHub]
- MApp-KG: Mobile App Knowledge Graph for Document-Based Feature Knowledge Generation [doi, replication package]
Feedback Analysis
- TransFeatEx: a NLP pipeline for feature extraction [paper, GitHub]
- T-FREX: A Transformer-based Feature Extraction Method from Mobile App Reviews [paper, GitHub, HuggingFace models]
- RE-Miner: Mining Mobile User Reviews with Feature Extraction and Emotion Classification [paper, GitHub (back-end), GitHub (front-end)]

Team Leader and Contact Person

Quim Motger
Contact Quim Motger