인공지능 및 SW공학 연구실

(Artificial Intelligence and Software Engineering Laboratory)

Professor

My research field is a study that combines artificial intelligence and software engineering. Specifically, research on applying artificial intelligence techniques to improve software quality (AI4SE) and software engineering techniques to improve the quality of artificial intelligence-based software systems (SE4AI) are conducted. For artificial intelligence-based software analysis, mining software repository, software defect prediction, software bug report prediction, software reliability engineering, web service QoS prediction, and software fault localization research are being conducted. In particular, in the field of software defect prediction, hyperparameter optimization using metaheuristic algorithms, cross-project defect prediction models to solve cold-start problems, and heterogeneous defect prediction models that can be built even if the feature dimensions are different between the training set and the test set. In order to solve the imbalance of the ratio of defective and non-defective instances (class imbalance), a cost-sensitive modeling technique and a boosting technique were newly devised. To this end, various machine learning and deep learning techniques were applied. These include multi-output convolutional neural networks, generative adversarial neural networks, Bayesian optimization, Explainable AI, AutoML, TabNet, self-supervised learning, deep domain adaptation, deep metric learning, image conversion techniques, etc. In the Web service QoS prediction study, a collaborative filtering method based on a matrix factorization model commonly used in recommendation systems was newly devised to analyze the quality of web services in a cloud environment.

Research Areas

Education