Dr. Vasile Rus' research interests lie at the intersection of machine learning, data mining, computational linguistics/natural language processing, and cognitive science with an emphasis on developing interactive intelligent systems based on strong theoretical findings in order to solve critical challenges that would change the world.

More broadly and as a primary research paradigm, Dr. Rus uses big data driven approaches to address challenging problems. In particular, he combines machine learning, statistical, and computational linguistics techniques to infer models that take advantage of the rich (large and diverse) sources of data available nowadays to address big challenges such as automatically answering questions, furthering the effectiveness of educational technologies such as intelligent tutoring systems on STEM-C topics, increasing the pace of discovery in biomedical domains, developing interactive systems that optimize the effectiveness of substance abuse treatments, or improving the quality of very large software projects. Dr. Rus has been developing advanced, optimized models from structured, semi-structured, and open-structure, i.e. free text, data collections using both supervised and unsupervised techniques.

Dr. Rus has been exploring Bayesian methods through the development of hierarchical models for many of the applications mentioned above, sequential data models based on Hidden Markov Models, Maximum Entropy Models and Conditional Random Fields, Markov Logic for discovering learners' proficiency during tutoring, deep neural network models for semantic processing in dialogues, as well as unsupervised, semi-supervised, co-training, and active learning methods. In a world flooded by streams of data and large collections of data there is an acute need for methods that scale and are unsupervised or semi-supervised (annotating large quantities of data is out of the question as the costs would be prohibitive). This will be a primary area of research for Dr. Rus for the next decade with a focus on fundamentals and applications to many of the challenges mentioned above. In particular, two major research goals Dr. Rus will be pursuing would be to improve learning ecosystem's effectiveness and cost-efficiency as well as learners and instructors' engagement and satisfaction through the design and development of advanced, data-driven educational technologies as well as the development of intelligent interactive systems for precision medicine in the healthcare domain, e.g. to optimize substance abuse treatments and healthcare services for the elderly.

During his graduate studies Dr. Vasile Rus mainly worked on knowledge representation and automated reasoning methods to boost the performance of automated Question Anwering systems from very large, unstructured (free text) collections of documents.

Dr. Rus was involved in the development of SMU's Question Answering system that participated in the Question Answering competition organized by NIST (National Institute for Standards and Technology) in 1999 as part of TREC-9 conference (Text REtrieval Conference) as well as in 2000 and 2001. His team won the competition in all three years. Vasile has also worked on Question Answering on Semantic Web. All the following publications are copyrighted by the author.
 

Selected Publications
 

Theses


Books


Selected Refereed Journal Papers and Book Chapters


Selected Refereed Conference Papers


Program Committees