Computer Science Department
School of Computer Science
Carnegie Mellon University
Email: <my first name>k at cs dot cmu dot edu
Office: GHC 6221
Ph.D. candidate in Computer Science, CMU.
Tom Mitchell is my advisor.
S.B. in Computer Science, MIT (2008)
I've been a teaching assistant for:
11-713 - Advanced NLP Seminar (CMU, Fall 2012)
15-213 - Introduction to Computer Systems (CMU, Spring 2012)
10-701 - Machine Learning (CMU, Fall 2010)
6.857 - Computer and Network Security (MIT, Spring 2009)
6.006 - Introduction to Algorithms (MIT, Fall 2008)
I'm interested in natural language semantics, specifically in producing computer programs that understand natural language. The two questions I'm most interested in are:
I work on the Never-Ending Language Learner (NELL), a system which learns to read using a corpus of webpages. My research includes both work on the underlying knowledge representation and interpreting text given a knowledge representation. I recently gave a talk at the CMU Machine Learning lunch about my recent semantic parsing work.
Vector Space Semantic Parsing: A Framework for Compositional Vector Space Models. In Proceedings of the 2013 Workshop on Continuous Vector Space Models and their Compositionality, 2013. PDF
Toward Interactive Grounded Language Acquisition. In Proceedings of Robotics: Science and Systems, 2013. PDF
Weakly Supervised Training of Semantic Parsers. In Proceedings of the 2012 Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning (EMNLP-CoNLL), 2012. PDF
Learning to Parse and Ground Natural Language Commands to Robots. In Proceedings of the AAAI 2012 Workshop on Grounding Language for Physical Systems, 2012. PDF
Which Noun Phrases Denote Which Concepts? In Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics (ACL), 2011. PDF
CrossBridge: Finding Analogies using Dimensionality Reduction. In Proceedings of the AAAI Fall Symposium on Common Sense Knowledge, November 2010. PDF
Finding Analogies in Semantic Networks using the Singular Value Decomposition. Master's thesis, Massachusetts Institute of Technology, 2009. PDF CrossBridge, the analogy algorithm from the paper, is now included in Divisi.
An Interface for Targeted Collection of Common Sense Knowledge using a Mixture Model. In Proceedings of Intelligent User Interfaces, 2009. PDF
The MD6 Hash Function -- A Proposal to NIST for SHA-3. Submitted to NIST on October 27, 2008. PDF The latest version of the MD6 report, along with other related material and code is available from the MD6 website.
Binary Search Tees. Clothing for computer scientists. Funny T-shirts that only a grad student could love.