SELENE: Software Engineering Laboratory for Next-Gen Ecosystems

SELENE (seh-LEE-nee) integrates classical principles of Software Engineering (SE) with state-of-the-art Artificial Intelligence (AI) to develop and explore automated methodologies that make software more reliable (with fewer bugs) and easier to maintain. Rather than merely applying AI as-is, we emphasize ensuring that the resulting methodologies are trustworthy and practical in real-world settings.

Our Principles

  1. Commitment to Better Software: In the rapidly evolving software ecosystem, we continuously redefine what it means to build better software, explore this vision from multiple perspectives, and engage with all challenges necessary to realize it.

  2. Practical Impact: Our goal is to develop flexible, practical, and effective methodologies that can be applied in real-world environments. To this end, we validate our approaches on real systems, sometimes using open-source projects and, when possible, through industry collaborations.

  3. Trust & Explainability: We focus not only on applying AI to software engineering but also on automatically verifying the trustworthiness of its outcomes, ensuring that AI-driven tools remain transparent, reliable, and dependable.

Research Topics & Selected Publications

🐛 Debugging Automation & Debugging Hints

Related Publication 🤖 👫 🏭
Finding the Needle in the Crash Stack: Industrial-Scale Crash Root Cause Localization with AutoCrashFL O   O
COSMosFL: Ensemble of Small Language Models for Fault Localisation O O  
A Quantitative and Qualitative Evaluation of LLM-based Explainable Fault Localization O O  
Fonte: Finding Bug Inducing Commits from Failures   O O
Automatically Identifying Shared Root Causes of Test Breakages in SAP HANA     O
FDG: A Precise Measurement of Fault Diagnosability Gain of Test Cases   O  
Reducing the Search Space of Bug Inducing Commits using Failure Coverage   O  
Assisting Bug Report Assignment Using Automated Fault Localisation: An Industrial Case Study     O

🌦️ Proactive Accuracy Forecasting for AI-Driven SE

Related Publication 🤖 👫 🏭
Lachesis: Predicting LLM Inference Accuracy using Structural Properties of Reasoning Paths O O  
Ensemble & confidence measurement in AutoFL, COSMosFL, AutoCrashFL O O  
Confidence quantification via metamorphic query in METAMON O O  

🧪 Test Automation & Optimization

Related Publication 🤖 👫 🏭
Evaluating Machine Learning-Based Test Case Prioritization in the Real World: An Experiment with SAP HANA O   O
FDG: A Precise Measurement of Fault Diagnosability Gain of Test Cases   O  
Just-in-Time Flaky Test Detection via Abstracted Failure Symptom Matching     O

🧹 Software Maintenance

Related Publication 🤖 👫 🏭
METAMON: Finding Inconsistencies between Program Documentation and Behavior using Metamorphic LLM Queries O O  
Iterative Refactoring of Real-World Open-Source Programs with Large Language Models O O  
  • 🤖 AI-for-SE
  • 👫 Evaluated on open-source projects
  • 🏭 Industry collaboration / industrial case study