CS student Elza Jung presented her project “Just-in-Time Defect Prediction Using Cost-Efficient Boosting Models” at the 39th Canadian Conference on Artificial Intelligence, under the supervision of Prof. Md Asif Khan. The project was recognized for its practical contribution to software engineering and machine learning, focusing on how defect-prone code changes can be detected earlier in the software development process. By demonstrating that a carefully tuned XGBoost model with engineered features and imbalance-aware sampling can improve defect detection while remaining computationally efficient, the work highlights a promising alternative to more costly deep-learning approaches. The study shows strong potential for helping software teams improve code quality, reduce review effort, and identify risky commits before they affect production systems. Congratulations, Elza! Her initial f

Computer Science student presented research at Canadian AI 2026

June 15, 2026
Print | PDF

Computer Science student Elza Jung presented her project “Just-in-Time Defect Prediction Using Cost-Efficient Boosting Models” at the 39th Canadian Conference on Artificial Intelligence, under the supervision of Prof. Md Asif Khan. The project was recognized for its practical contribution to software engineering and machine learning, focusing on how defect-prone code changes can be detected earlier in the software development process.

By demonstrating that a carefully tuned XGBoost model with engineered features and imbalance-aware sampling can improve defect detection while remaining computationally efficient, the work highlights a promising alternative to more costly deep-learning approaches. The study shows strong potential for helping software teams improve code quality, reduce review effort, and identify risky commits before they affect production systems. Congratulations, Elza!

Her initial findings were also showcased as a poster at the 2026 FOSSA Undergraduate Research Conference, held at Wilfrid Laurier University’s Waterloo campus.