Professor Alvarez does research in automated inductive learning. She uses grammatical inference methods based on mixing states that produce finite state automata models. She has demonstrated that symbolic inductive learning methods (not based on probability) can produce competitive results while enabling a more natural representation of knowledge. She has contributed new algorithms and applied them to natural language learning, prediction of poliprotein cleavage sites, and automated recommendation.
Currently Funded Research
Hidden Markov models in natural language recognition. Product configuration and assembly in a software production line (grant: Javeriana).