Department of Computer Science
United States Naval Academy
(Publishes as Nathanael Chambers)
New research on event schema induction at EMNLP-2013.
Event/time extraction from raw text at TempEval-3; ranked 2nd.
Undergrad research on sentiment; highly ranked at SemEval-2013
Radio interview on WNAV Dec 14
Mood of the Nation live feed posted.
|Spring 2014:||IC211, Object-Oriented Programming|
|Fall 2013:||SI485i, Natural Language Processing|
|SI110, Cyber Security I|
|Spring 2013:||IT452, Advanced Web & Internet Programming|
|Fall 2012:||SI485i, Natural Language Processing|
|IC210, Intro to Computer Science|
|Spring 2012:||IT452, Advanced Web & Internet Programming|
|Fall 2011:||IT350, Web and Internet Programming|
Graduate Studies: info for CS/IT majors and M.S. degrees.
I research natural language processing (NLP) and learning about the world from text with minimal supervision. I apply modern NLP and machine learning techniques to the Web to learn how events and people interact. I am also interested in information extraction to help people better understand large amounts of text. I also model time: when events start, how long, and in what order they occur.
I am always open to advising undergraduate research. Come talk to me if you're interested. Topics ranging from social media like Twitter to intelligent systems that read the news are ripe for study.
Ph.D., Computer Science, Stanford University, 2011
M.S., Computer Science, University of Rochester, 2003
B.S., Computer Science, University of Rochester, 2002
Selected Publications (view all 33 with downloads)
Event Schema Induction with a Probabilistic Entity-Driven Model
NavyTime: Event and Time Ordering from Raw Text
Deterministic coreference resolution based on entity-centric, precision-ranked rules
Labeling Documents with Timestamps: Learning from their Time Expressions
Learning the Central Events and Participants in Unlabeled Text
Learning for Microblogs with Distant Supervision: Political Forecasting with Twitter
Template-Based Information Extraction without the Templates
A Multi-Pass Sieve for Coreference Resolution
Unsupervised Learning of Narrative Schemas and their Participants
Jointly Combining Implicit Constraints Improves Temporal Ordering
Unsupervised Learning of Narrative Event Chains
Classifying Temporal Relations Between Events
PLOW: A Collaborative Task Learning Agent (outstanding paper award)
Using Semantics to Identify Web Objects
Chester: Towards a Personal Medical Advisor
Journal Biomedical Informatics
Narrative Cloze Evaluation (Gigaword Corpus NYT labeled)
Past and present web design projects I have been involved with:
A collaborative website for teachers (founder, no longer affiliated)