Eduard Hovy

I investigate various aspects of human language, developing both theories and applications for natural language processing and human language technology. Ideas my students and I develop include both theoretical and computational facets. I am specifically interested in the following areas:

Semantics for NLP
In this work, I explore structured distributional semantics — the interrelationships between traditional propositional semantics and modern statistics-based distributional semantics. To test this, we are developing sophisticated machine-reading, information-extraction and text-analysis technology. My prior work also includes efforts in text summarization and automated question-answering systems. 

Pragmatics and Social Media
Our research in this area aims to define and automate the discovery of the personality profiles, interaction patterns and interpersonal relationships of people active on social media. 

Ontologies and Large Knowledge Resources
My research in this topic focuses on developing large concept taxonomies/ontologies through a combination of merging existing ontologies, adding to the knowledge by extracting information from online text and enriching interdependency relations by extracting information from dictionaries. We develop techniques to mine large amounts of instance- and concept-level information from online text. 

Digital Government, Homeland Security, and Law Enforcement
Work in this topic aims to assist emergency response teams through event discovery and tracking in online social media, assisting law enforcement in the fight against human trafficking and developing technology to assist people regarding cybersecurity.