Contact

Tony Mullen
410 Thompson Hall
Dept. of Computer Science
University of Puget Sound
1500 N. Warner St.
Tacoma, WA 98416

Phone: 253.879.3562
Email: tmullen@pugetsound.edu

Research

My academic research has mostly been in the fields of natural language processing and computer-assisted language learning. Less formally, I am interested in computer graphics, web development, and interaction design. I am particularly interested in projects that bring as many of these diverse areas together as possible.

Computer assisted language-learning: For some time I have been involved in developing online environments for tandem language exchange between learners of each others' native languages. My work in this field goes back to 1998, when I implemented the Electronic Tandem Resources (ETR) web application used in language courses at Trinity College, Dublin and Open University of Catalunya. ETR enabled language learners to engage in asynchronous language exchange in an environment that was similar to webmail but incorporated features to encourage language learning, support tandem best practices, and facilitate data collection and analysis by instructors and researchers. Later, I turned my attention to synchronous, VoIP-based exchanges, focusing on designing task-based environments to increase engagement between learners. Currently, I am interested in using web-based real-time communication (WebRTC), interactive web technologies, ideas from social networking, and innovative approaches to scheduling support to create engaging environments for language exchange.

Natural language processing: My earliest academic research was in the field of statistical parsing of natural language; I explored the use of maximum entropy-based models to select the best from a large number of parses generated by a wide-coverage, highly ambiguous natural language grammar. Later, I investigated maximum entropy-based approaches to named-entity recognition and rhetorical zone analysis of biomedical research texts. I also became interested in the field of sentiment analysis, where I have made my most influential academic contributions. This field is concerned with the identification and analysis of opinion and emotional content of natural language texts. I used methods from machine learning to recognize positive and negative reviews of products and music. My interest in this area led to research in political sentiment analysis, specifically the analysis of informal political discussion texts to automatically identify political leanings.

In the future, I hope to apply techniques from NLP and machine learning to analyzing data acquired from within the tandem language exchange environment mentioned previously. The creation of a data set of spoken second-language conversation would be of great interest to researchers in the field of automatic spoken language assessment.

Web/UX design: In addition to working on the design of the task-based tandem language exchange environment, I also am interested in other interactive educational applications. In particular, I am working on a web-based visualization tool for learners of computation theory that will enable users to create, edit, save, and share finite-state automata, pushdown automata, and Turing machines.