I am a computational linguist and cognitive scientist specializing in modeling meaning, particularly meaning in context. I spent my research career developing mathematical models of conversational meaning, and testing those models on empirical data. I applied this work in the area of natural language generation, aiming to create more human-like dialogue systems. Most recently I have turned my focus to combining linguistic theory and machine learning to improve natural language generation and understanding, and I am currently applying these techniques in the biopharmaceutical domain as an NLP developer at AbbVie.
Skills
- Computational modeling
- Data analysis, visualization and collection
- Computational semantics (word embeddings, models of discourse)
- Machine learning (classification, regression, deep learning)
- Natural language understanding and generation
- Corpus linguistics
- Game theory
- Information theory
- Experiment design
- Syntactic formalisms (TAG, CCG)
- Teaching university-level courses
- Public speaking
- Fluent in German
Experience
NLP software developer
AbbVie, July 2018-present
Develop and contribute to software projects aiding in the analysis and curation of scientific literature; apply linguistic analysis and NLP tools to improve the accessibility of knowledge to researchers aiming to discover and develop new drugs
Researcher in computational linguistics
Ohio State University, August 2016-July 2018
Led research on computational models of linguistic meaning and cognition; designed experiments to test models; analyzed and visualized behavioral data; collaborated on research publications; taught courses
Consultant on potential commercial project
Noema Inc., NY, April 2017-July 2018
Designed algorithms that applied game-theoretic models of pragmatics to disambiguate entities in a named entity recognition (NER) system
Data science consultant
Data Science Retreat, Berlin, May 2016-July 2016
Sub-contracted to design natural language processing modules for an IBM Big Data University course on data science with Scala and Spark
Researcher in computational linguistics
ZAS Berlin, PRAGSales project, July 2013-June 2016
Led research on developing novel methods of generating natural answers to user questions in a dialogue system; aided with the creation of a corpus and database ontology
Education
PhD in Linguistics, University of Pennsylvania, 2013
Emphasis on game theory, mathematical models of communication and computational models of language learning.
BA/MA in Linguistics, Ohio State University, 2008
Programming Skills
Python
(Keras, SpaCy, SciPy, pandas, gensim, scikit-learn)
R
Scala
Spark
(SparkML, MLib)
SQL
Bash
LaTeX
HTML
Academic Works: Highlights
Oxford Research Encylopedia entry on the Pragmatics of Focus.
2018. Benz, Anton and Jon Stevens. Game-theoretic approaches to pragmatics. Annual Review of Linguistics.
2017. Stevens, Jon, Lila Gleitman, John Trueswell & Charles Yang. The pursuit of word meanings.
Cognitive Science 41(S4): 638-676.
2016. Stevens, Jon, Anton Benz, Sebastian Reuße & Ralf Klabunde. Pragmatic question answering:
A game-theoretic approach. Data & Knowledge Engineering 106: 52-69.
2016. Stevens, Jon. Focus games. Linguistics and Philosophy 39: 395-441.
Conferences: Highlights
Stevens, Jon, Marie-Catherine de Marneffe, Shari Speer & Judith Tonhauser. Rational use of prosody predicts projection in manner adverb utterances. Paper presented at CogSci 2017, London, July 26-29.
Stevens, Jon, Anton Benz, Sebastian Reuße & Ralf Klabunde. A strategic reasoning model for
generating alternative answers. Poster presented at ACL 2015, Beijing, China, July 26-31 2015.