St. Thomas Engineer Magazine

Faculty Profile

Behind the AUTOMATED ATTENDANT

If you have ever called a bank, insurance company or an airline, you know this scenario well. A digital operator will ask you to press numbers or say your preferences. Annoying or not, there is a lot of science behind that interaction. Graduate Programs in Software Assistant

relates to signal processing and programming languages. Kazemzadeh got an early start in this field, designing an acoustic model for immersive, speech-interactive virtual reality (VR) world language training which helped to teach both aid workers and soldiers about languages in Iraq, Indonesia and Afghanistan during the Gulf War. He designed a VR video game in which players talk to communicate with video characters to bring aid to villagers after a natural disaster or to set up a base. The speech recognition decoder evaluates the answer and the player advances based on their correct verbal answers. Quite recently, speech recognition has become more practical due to deep learning, a hot topic in machine learning. Deep learning uses neural networks, which boils down to people doing less by hand, because the computer learns how to do it. As speech recognition continues to improve based on deep learning, so do the fields of computer vision and image recognition, allowing for some cross pollination of knowledge. Researchers can use the same technology to further advances in both fields, which has helped accelerate the speech recognition field dramatically in the last 5-10 years.

Professor Abe Kazemzadeh, PhD, is an expert

at language, with degrees in linguistics,

computational linguistics and computer science, and he has

industry experience. He has worked on notable projects such as an automated pronunciation modeling for child literacy assessment, a VXML telephone survey in English and Spanish for autism outreach, and several components for processing the Dow Jones newswire, including named-entity recognition, full- text search and commodities event extraction. His passion for linguistics was fueled by a high school class trip to Costa Rica where he noticed that dialects caused some misunderstandings. He pursued a degree in linguistics at University of Southern California and an electrical engineering professor made the connection for him on how linguistics

St. Thomas Engineer 2020 Page 27

Powered by