Alvin Grissom II
Department of Computer Science
Haverford College
Computational Approaches to Sentence-final Verb Prediction and What We Can Learn from Counterintuitive Errors in Machine Learning
In this talk, I will present an overview of some of my work on incremental natural language processing. After providing a short general introduction to machine learning-based computational linguistics, I will discuss my previous work incremental language processing in the context of machine learning-based simultaneous interpretation from verb-final to verb-medial languages (e.g., Japanese to English) and the related problem of incremental verb prediction by computers its human verb prediction. I will then discuss previous work on “pathological” errors in neural machine learning systems and offer thoughts on what these counterintuitive errors suggest deep learning systems are, and are not, learning.