Current Research
My research focuses on the development of multilingual NLP techniques and systems. While language technology research has traditionally focused on English and a few other common world languages, the global penetration of the internet requires a new, multilingual, approach. By taking such an approach, we open the door to several exciting avenues of research:
Without the benefit of hand-annotated training data, NLP models suffer in terms of accuracy and robustness. However, by leveraging a large set of languages, we can treat each language as a data-point in the set of all possible human languages. By gathering enough such data-points, we can build unsupervised learning methods which focus on hypotheses which are universally plausible.
A multilingual computational approach can also help us understand why the space of possible languages looks the way it does. Linguists have long debated the existence and nature of "language universals" and the related question of whether humans possess an innate language instinct. Computational modeling has the potential to shed much light on these controversies.
Finally, multilingual modeling can help us develop practical technology to aid in human language acquisition. Millions of people struggle everyday with a foreign language. My aim is to develop practical NLP technology that leverages our understanding of human cognitive processes to aid in language learning.
Awards
Media
PhD Students
Biography
Benjamin Snyder is an Assistant Professor at the University of Wisconsin-Madison in the Department of Computer Sciences. His research areas include natural language processing, computational linguistics, and cognitive science. Ben received his Ph.D. in EECS from MIT in 2010. He received his B.A in philosophy from the University of Pennsylvania in 2003. During the fall of 2010, Ben worked as a post-doc in the CIS department at Penn as a member of Ben Taskar's research group.
Publications
Training a Korean SRL System with Rich Morphological Features
Young-Bum Kim, Heemoon Chae, Benjamin Snyder and Yu-Seop Kim. ACL 2014 (short paper)
Unsupervised Consonant-Vowel Prediction over Hundreds of Languages
Young-Bum Kim and Benjamin Snyder. ACL 2013
Modeling Child Divergences from Adult Grammar
Sam Sahakian and Benjamin Snyder. Transactions of the ACL 2013
Optimal Data Set Selection: An Application to Grapheme-to-Phoneme Conversion
Young-Bum Kim and Benjamin Snyder. NAACL 2013
Universal Grapheme to Phoneme Prediction over Latin Alphabets
Young-Bum Kim and Benjamin Snyder. EMNLP 2012
⇒ Data
Automatically Learning Measures of Child Language Development
Sam Sahakian and Benjamin Snyder. ACL 2012 (short paper)
Universal Morphological Analysis using Structured Nearest Neighbor Prediction
Young-Bum Kim, Joao Graca, and Benjamin Snyder. EMNLP 2011
Unsupervised Multilingual Learning. PhD Thesis.
Electrical Engineering and Computer Science, MIT. September 2010.
⇒ Winner of the ACM dissertation award honorable mention.
⇒ Winner of the George M. Sprowls Award for best doctoral thesis.
A Statistical Model for Lost Language Decipherment
Benjamin Snyder, Regina Barzilay, and Kevin Knight. ACL 2010
⇒Write-up in National Geographic News
⇒Write-up in MIT news
⇒Interview on Swedish National Radio
Climbing the Tower of Babel: Unsupervised Multilingual Learning
Benjamin Snyder and Regina Barzilay. ICML 2010
Multilingual Part-of-Speech Tagging: Two Unsupervised Approaches
Tahira Naseem, Benjamin Snyder, Jacob Eisenstein, and Regina Barzilay. JAIR 36 (2009)
Unsupervised Multilingual Grammar Induction
Benjamin Snyder, Tahira Naseem, and Regina Barzilay. ACL 2009
⇒ Slides: (PDF) (Keynote)
Adding More Languages Improves Unsupervised Multilingual Part-of-Speech Tagging:
A Bayesian Non-Parametric Approach
Benjamin Snyder, Tahira Naseem, Jacob Eisenstein, and Regina Barzilay. NAACL 2009
⇒ Slides: (PDF) (Keynote)
Unsupervised Multilingual Learning for POS Tagging
Benjamin Snyder, Tahira Naseem, Jacob Eisenstein, and Regina Barzilay. EMNLP 2008
⇒ Code & Data Slides: (PDF) (Keynote)
Unsupervised Multilingual Learning for Morphological Segmentation
Benjamin Snyder and Regina Barzilay. ACL 2008
⇒ Slides: (PDF) (Keynote)
Cross-lingual Propagation for Morphological Analysis
Benjamin Snyder and Regina Barzilay. AAAI 2008
Multiple Aspect Ranking for Opinion Analysis
Benjamin Snyder. Master's Thesis, MIT 2007
Incremental Text Structuring with Online Hierarchical Ranking
Erdong Chen, Benjamin Snyder, and Regina Barzilay. EMNLP 2007
Multiple Aspect Ranking using the Good Grief Algorithm
Benjamin Snyder and Regina Barzilay. NAACL 2007
⇒ (Data and Code) (Slides)
See my Master's Thesis for improved results, generalizations, and theoretical analysis.
Restaurant Browser system
Database-Text Alignment via Structured Multilabel Classification
Benjamin Snyder and Regina Barzilay. IJCAI 2007
⇒ (Data)