![]() |
5th year PhD Student in the Department of Computer and Information Science and Institute for Research in Cognitive Science at the University of Pennsylvania Office: IRCS 424 Email: Prove you're not a robot to get it Details: CV, ConstantineLignos on Github |
Computational modeling provides a unique opportunity to test the predictions of theories of language and cognition. I seek to improve our understanding of language in the mind by developing cognitively-plausible models of language acquisition, processing, and change that are informed by both experimental and theoretical work. I work primarily with Mitch Marcus and Charles Yang. The main problems I've worked on are:
- Word segmentation and morphology learning: In the first years of life, children learn to segment the speech stream into structured words. I work on algorithmic-level, online models of acquisition that are computationally efficient, cognitively plausible, and predict the ways in which infants learn over time.
- Morphological processing: I use computational modeling to extend experimental work by evaluating theories of morphological processing using large scale data. My work focuses on the debate surrounding the role of frequency in facilitating processing and the nature of morphological decomposition.
- Natural language understanding for autonomous systems: In situations like search and rescue, we need robots to perform complicated tasks under limited supervision. As a part of the SUBTLE MURI project, I've developed SLURP, a system that allows users to specify simple tasks for a robot to perform using natural language by translating their instructions into a satisfiable logical plan.
I've also done research in more traditional natural language processing. I developed MORSEL, a cognitively-motivated state-of-the-art unsupervised morphological analyzer. From June-August 2012, I participated in the SCALE summer workshop at the Johns Hopkins Center of Excellence in Human Language Technology. With Mitch Marcus, I developed Codeswitchador, a system for identifying code-switching in social media data. This work enables the creation of large scale corpora of code-switching and identification of bilingual users.
Recent and Upcoming Presentations
From lexicon to grammar in infant word segmentation. Presentation at the 87th Annual Meeting of the Linguistic Society of America, January 5, 2013. [ slides ]
Toward web-scale analysis of codeswitching. (with Mitch Marcus) Poster at the 87th Annual Meeting of the Linguistic Society of America, January 5, 2013. [ poster ]
Modeling domain-narrowing phonological change. Presentation at West Coast Conference on Formal Linguistics 31, Feburary 8-10, 2013. [ slides ]
Symbolic constraints and statistical methods: Use together for best results. Presentation at Johns Hopkins Human Language Technology Center of Excellence, February 18, 2013. [ slides ]
Articles
Erwin Chan and Constantine Lignos. Investigating the relationship between linguistic representation and computation through an unsupervised model of human morphology learning. Research on Language and Computation, 8 (2), 209-238. Springer Netherlands, 2010. [ paper ]
Constantine Lignos, Erwin Chan, Mitchell Marcus, and Charles Yang. A rule-based acquisition model adapted for morphological analysis. In C. Peters, G. Di Nunzio, M. Kurimo, T. Mandl, D. Mostefa, A. Penas, and G. Roda, editors, Multilingual Information Access Evaluation I. Text Retrieval Experiments. Lecture Notes in Computer Science, 6241, 658-665. Springer Berlin / Heidelberg, 2010. [ preprint ]
Conference/Workshop Proceedings
Daniel J. Brooks, Constantine Lignos, Cameron Finucane, Mikhail S. Medvedev, Ian Perera, Vasumathi Raman, Hadas Kress-Gazit, Mitch Marcus, Holly A Yanco. Make it so: Continuous, flexible natural language interaction with an autonomous robot. Proceedings of the Grounding Language for Physical Systems Workshop at the Twenty-Sixth AAAI Conference on Artificial Intelligence, Toronto, Ontario, Canada, July 23, 2012. AAAI. [ paper ]
Constantine Lignos and Kyle Gorman. Revisiting frequency and storage in morphological processing. To appear, Proceedings of the 48th Annual Meeting of the Chicago Linguistic Society, April 19-21, 2012. [ paper ]
Constantine Lignos. Infant word segmentation: An incremental, integrated model. Proceedings of the West Coast Conference on Formal Linguistics 30, April 13-15, 2012. [ paper ]
Constantine Lignos. You can't get there from here: On interpreting learning experiments. To appear, proceedings of Penn Linguistics Colloquium 36, March 23-25, 2012. [ paper ]
Constantine Lignos. Modeling infant word segmentation. In Proceedings of the Fifteenth Conference on Computational Natural Language Learning, pages 29-38, Portland, Oregon, USA, June 2011. Association for Computational Linguistics. [ paper ]
Constantine Lignos. Learning from Unseen Data. In Mikko Kurimo, Sami Virpioja, and Ville T. Turunen, editors, Proceedings of the Morpho Challenge 2010 Workshop, pages 35-38, Helsinki, Finland, September 2-3 2010. Aalto University School of Science and Technology. Best paper award. [ paper ]
Constantine Lignos and Charles Yang. Recession segmentation: Simpler online word segmentation using limited resources. In Proceedings of the Fourteenth Conference on Computational Natural Language Learning, pages 88-97, Uppsala, Sweden, July 2010. Association for Computational Linguistics. [ paper ]
Constantine Lignos, Erwin Chan, Charles Yang, and Mitchell P. Marcus. Evidence for a morphological acquisition model from development data. In Katie Franich, Kate M. Iserman, and Lauren L. Keil, editors, Proceedings of the 34th Annual Boston University Conference on Language Development. Cascadilla Press, 2010. [ paper ]
Constantine Lignos, Erwin Chan, Mitchell P. Marcus, and Charles Yang. A Rule-Based Unsupervised Morphology Learning Framework. In Working Notes of the 10th Workshop of the Cross-Language Evaluation Forum, CLEF 2009, Corfu, Greece, September 30-October 2 2009. [ paper ]
Emily Wang, Constantine Lignos, Ashish Vatsal, and Brian Scassellati. Effects of head movement on perceptions of humanoid robot behavior. In HRI '06: Proceedings of the 1st ACM SIGCHI/SIGART Conference on Human-Robot Interaction, pages 180-185, New York, NY, USA, 2006. ACM. [ paper ]
Conference Posters/Presentations
Productivity in analogical change. Presentation at the Manchester and Salford New Researchers Forum in Linguistics, November 2-3, 2012. [ slides ]
Constantine Lignos and Charles Yang. Cuing infants in: From universal to language-specific cues in word segmentation. Presentation at ICIS 2012. June 6-9th, 2012. [ slides ]
Constantine Lignos and Kyle Gorman. The subtleties of frequency in morphological processing. Poster at CUNY Conference on Human Sentence Processing. March 14-16, 2012.
Laurel MacKenzie and Constantine Lignos. Size matters: The effect of subject length on contraction. Presentation at Georgetown University Round Table on Languages and Linguistics. March 8-11, 2012.
Constantine Lignos. The journey to word segmentation. Presentation at BUCLD 36, November 4-6 [ slides ]
Constantine Lignos and Charles Yang. Using prosody to bootstrap word segmentation in a more realistic learning environment. Presentation at BCCCD in the symposium "The Role Of Prosody In Guiding Language Learning In Pre-Lexical Infants" organized by Mohinish Shukla, January 14-16 2011. [ slides ]
Constantine Lignos and Jana Beck. The power of objects in morphology learning. Presentation at LSA Annual Meeting, January 6-9 2011. [ slides ]
