Matthew Roddy


View My GitHub Profile


[Google Scholar] [Github] [CV] [LinkedIn] [Twitter]

I’m a research scientist with a focus on machine learning for speech technologies. I received my PhD from the Electrical Engineering Department at Trinity College Dublin under the supervision of Dr. Naomi Harte. I have published first-author papers at a number of top-tier conferences such as ACL, Interspeech, ICMI, and Eusipco. I have also acted as a reviewer for conferences such as ICASSP, Interspeech, and ICMI. My PhD topic was on the design of conversational turn-taking models for spoken dialogue systems (SDSs). More specifically, my PhD focused on extracting information from multiple modalities to train deep learning models that predict conversational turn-taking behaviors. The research incorporates elements from pure machine-learning, natural language processing, audio-visual signal processing, and social sciences. During my PhD I was part of the Sigmedia Lab and my work was funded by the ADAPT Centre.

PhD Thesis