• Skip down to publications.
• My ESSLLI 2013 course notes on grammar-based approaches to opinion mining

I am a Associate Senior Lecturer (equivalent to tenure-track assistant professor) in the Department of Philosophy, Linguistics, and Theory of Science at the University of Gothenburg in Sweden, affiliated with the Center for Linguistic Theory and Studies in Probability (CLASP).

I was previously a postdoctoral research associate at the University of the Saarland affiliated with the Multimodal Computing and Interaction Cluster of Excellence (MMCI) and the Computational Linguistics (COLI) research units. In 2011, I obtained a Ph.D. from the University of Maryland Department of Computer Science, and I was a member of the Laboratory for Computational Linguistics and Information Processing (CLIP). My experience runs the gamut from information retrieval to syntactic theory to sentiment analysis, and I have teaching experience as a discussion/lab section TA. My most recent ongoing project is about developing a robust semantics for a cognitively plausible parsing formalism; I still work with closely with Vera Demberg at Saarland.

My completed dissertation was in opinion mining and sentiment analysis in a social science context using robust distributional and syntax-oriented machine learning techniques. Since then, I moved on to computational psycholinguistics, particularly the role of semantics in adult sentence processing.

I am either a language-aware computer scientist or a computation-aware linguist depending on your perspective.

Advisor: Prof. Amy Weinberg

Achieved candidacy March 2010.

Defended and submitted August 2011.

Thesis title: "A distributional and syntactic approach to fine-grained opinion mining."

Thesis: Developing a Minimalist Parser for Free Word Order Languages
Advisor: Prof. Stan Szpakowicz

Honours Project: Parsing Techniques for Free Word Order Languages as Exemplified by Latin. Supervised by Prof. Stan Szpakowicz.

• Non-information-density approaches to online processing (Winter 2015/2016) -- NEW! starting October 28!
• Spoken dialog systems for elevator control (Summer 2015, with Ingmar Steiner and Arif Khan)
• Is there any time for scope (Winter 2014/2015)
• Understanding the theory of syntax (Summer 2014)
• Adaptive language generation (Winter 2013/2014, with Vera Demberg)
• The syntax/semantics interface in an incremental context (Winter 2012/2013, with Vera Demberg)

Asad Sayeed and Alessandra Zarcone. Explicit-world knowledge and distributional semantic representations. European Summer School of Logic, Language, and Information (ESSLLI 2017). One week course to be presented July 2017, Toulouse, France.

Ted Gibson, Tal Linzen, Asad Sayeed, Marten van Schijndel, and William Schuler. Cognitive Modeling and Computational Linguistics (CMCL 2017). Workshop held with EACL 2017, Valencia, Spain.

Asad Sayeed and Vera Demberg. Linguistic complexity and cognitive workload: measurement and management. European Summer School of Logic, Language, and Information (ESSLLI 2014). One week course presented August 2014, Tübingen, Germany.

Asad Sayeed. Grammar-based approaches to opinion mining. European Summer School of Logic, Language, and Information (ESSLLI 2013). One week course presented August 2013, Düsseldorf, Germany.

Asad Sayeed. Towards an annotation framework for incremental scope specification update. To appear 2017. Logic and machine learning in natural language processing (LaML 2017). Gothenburg, Sweden. (6 pages, oral presentation)

Asad Sayeed, Matthias Lindemann, and Vera Demberg. Scope ambiguity resolution, event structure bias, and German main clause scrambling. 2017. CUNY Sentence Processing Conference 2017. Boston, USA. (2 pages, abstract, poster presentation)

Ashutosh Modi, Ivan Titov, Manfred Pinkal, Vera Demberg, and Asad Sayeed. Modeling semantic expectation: using script knowledge for reference prediction. 2017. Transactions of the Association for Computational Linguistics (TACL). (12 pages)

Ottokar Tilk, Vera Demberg, Asad Sayeed, Dietrich Klakow, and Stefan Thater. Modelling semantic surprisal at the level of events.. 2016. Empirical Methods in Natural Language Processing (EMNLP 2016). Austin, Texas. (10 pages)

Asad Sayeed, Clayton Greenberg, and Vera Demberg. Thematic fit evaluation: an aspect of selectional preferences. 2016. Thematic fit evaluation: an aspect of selectional preferences. 2016. First workshop on evaluating vector-space representations for NLP (RepEval 2016) at ACL 2016. Berlin, Germany. (7 pages)

Asad Sayeed, Xudong Hong, and Vera Demberg. Roleo: visualising thematic fit spaces on the web. 2016. Assocation for Computational Linguistics Annual Meeting (ACL 2016), demonstration session. Berlin, Germany. (6 pages)

Florian Pusse, Asad Sayeed, and Vera Demberg. LingoTurk: managing crowdsourced tasks for psycholinguistics. 2016. North American Assocation for Computational Linguistics Annual Meeting (NAACL 2016), demonstration session. San Diego, USA. (5 pages)

Vera Demberg and Asad Sayeed. The frequency of rapid pupil dilations as a measure of linguistic processing difficulty. 2016. PLoS ONE 11(1). (28 pages)

Asad Sayeed, Vera Demberg, and Pavel Shkadzko. An exploration of semantic features in an unsupervised thematic fit evaluation framework. 2015. Italian Journal of Computational Linguistics, 1(1). (16 pages)

Asad Sayeed, Stefan Fischer, and Vera Demberg. Model adaptation -- to what extent to speakers adapt word duration to a domain?. Architectures and mechanisms in language processing (AMLaP), 2015. (1 page, abstract, poster presentation)

Asad Sayeed. Representing the effort in resolving ambiguous scope. Sinn und Bedeutung 20 (2015). Tübingen, Germany. (2 pages, abstract, poster presentation)

Asad Sayeed, Stefan Fischer, and Vera Demberg. Vector-space calculation of semantic surprisal for predicting word pronunciation duration. Association for Computational Linguistics annual meeting (ACL-IJCNLP 2015). Beijing, China. (9 pages, poster presentation, acceptance rate 25%)

Clayton Greenberg, Asad Sayeed, and Vera Demberg. Verb polysemy and frequency effects in thematic fit modeling. Workshop on Cognitive Modeling and Computational Linguistics (CMCL). North American Association for Computational Linguistics annual meeting (NAACL-HLT 2015). Denver, USA. (10 pages, oral presentation)

Clayton Greenberg, Asad Sayeed, and Vera Demberg. Improving unsupervised thematic fit evaluation via role-filler prototype clustering. North American Association for Computational Linguistics annual meeting (NAACL-HLT 2015). Denver, USA. (9 pages, oral presentation, 26% acceptance rate)

Asad Sayeed and Vera Demberg. Combining unsupervised syntactic and semantic models of thematic fit. First Italian conference on computational linguistics (CLIC-it 2014). Pisa, Italy. (5 pages, oral and poster presentation, Distinguished Young Paper award)

Vera Demberg and Asad Sayeed. Pupillometry - the Index of Cognitive Activity as a measure of linguistic processing difficulty. CUNY sentence processing conference 2014. Columbus, USA. (abstract and poster)

Vera Demberg, Asad Sayeed, Angela Mahr, and Christian Müller. Measuring linguistically-induced cognitive load during driving using the ConTRe task. Fifth international conference on automotive user interfaces and interactive vehicular applications (AutoUI 2013). Eindhoven, The Netherlands. (8 pages)

Asad Sayeed and Vera Demberg. The semantic augmentation of a psycholinguistically-motivated syntactic formalism. Cognitive modeling and computational linguistics 2013 (workshop at ACL 2013). Sofia, Bulgaria. (9 pages)

Vera Demberg, Evangelia Kiagia, and Asad Sayeed. The Index of Cognitive Activity as a Measure of Linguistic Processing. CogSci 2013. Berlin, Germany. (6 pages)

Nikolaos Engonopoulos, Asad Sayeed, and Vera Demberg. Language and cognitive load in a dual task environment. CogSci 2013. Berlin, Germany. (6 pages)

Asad Sayeed. An opinion about opinions about opinions: subjectivity and the aggregate reader. North American Association for Computational Linguistics (NAACL 2013). Atlanta, USA. (5 pages, oral presentation, 37% acceptance rate)

Asad Sayeed and Vera Demberg. Covert without overt: QR for movementless parsing frameworks. Generative Linguistics in the Old World 36 (2013), Lund Sweden. (2 pages, poster)

Asad Sayeed and Vera Demberg. Incremental neo-Davidsonian semantic construction for TAG. 2012. The 11th International Workshop on Tree-Adjoining Grammars and Related Formalisms (TAG+11). Paris, France. (9 pages)

Vera Demberg, Asad Sayeed, Philip Gorinski, and Nikos Engonopoulos. Syntactic surprisal has an effect on spoken word duration. Empirical Methods in Natural Language Processing (EMNLP 2012). Jeju, South Korea. (12 pages, 25% acceptance rate, appeared as poster at AMLaP 2012.)

Asad Sayeed, Jordan Boyd-Graber, Bryan Rusk, and Amy Weinberg. Grammatical structures for word-level sentiment detection. North American Association for Computational Linguistics (NAACL 2012). Montreal, Canada. (10 pages, 31% acceptance rate)

Asad Sayeed, Bryan Rusk, Martin Petrov, Hieu Nguyen, Timothy Meyer, and Amy Weinberg. Crowdsourcing syntactic relatedness judgements for opinion mining in the study of information technology adoption. ACL 2011 workshop on Language Technology for Cultural Heritage, Social Sciences, and Humanities (LaTeCH 2011), Portland, USA. (9 pages)

Asad Sayeed, Hieu Nguyen, Timothy Meyer, and Amy Weinberg. ``Expresses-an-opinion-about'': using corpus statistics in an information extraction approach to opinion mining. International Conference on Computational Linguistics (COLING 2010), Beijing, China. (poster volume pp.1095-1103, 41% acceptance rate)

Asad Sayeed, Timothy Meyer, Hieu Nguyen, Olivia Buzek, and Amy Weinberg. Crowdsourcing the evaluation of a domain-adapted named-entity recognition system. Annual meeting of the North American chapter of the Association for Computational Linguistics (NAACL-HLT 2010), Los Angeles, USA. (pp.345-348, 35% acceptance rate)

Chia-Jung Tsui, Ping Wang, Kenneth Fleischmann, Asad Sayeed, and Amy Weinberg. Building an IT taxonomy with co-occurrence analysis, hierarchical clustering, and multidimensional scaling. iConference 2010, Urbana-Champaign, USA. (pp.247-256)

Chia-Jung Tsui, Ping Wang, Kenneth Fleischmann, Douglas Oard, and Asad Sayeed. Understanding IT innovations by computational analysis of discourse. International Conference on Information Systems (ICIS 2009), Phoenix, USA. (9 pages, 27% acceptance rate)

Asad Sayeed, Soumitra Sarkar, Yu Deng, Rafah Hosn, Ruchi Mahindru, and Nithya Rajamani. Characteristics of document similarity measures for compliance analysis. ACM Conference on Information and Knowledge Management (CIKM 2009), Hong Kong. (pp. 1207-1216, 13% acceptance rate)

Asad Sayeed, Tamer Elsayed, Nikesh Garera, David Alexander, Tan Xu, Douglas Oard, David Yarowsky, and Christine Piatko. Arabic cross-document coreference detection. Annual meeting of the Association for Computational Linguistics (ACL 2009), Singapore. (pp.357-360, 26% acceptance rate)

James Mayfield, David Alexander, Bonnie Dorr, Jason Eisner, Tamer Elsayed, Tim Finin, Clay Fink, Marjorie Freedman, Nikesh Garera, Paul McNamee, Saif Mohammad, Douglas Oard, Christine Piatko, Asad Sayeed, Zareen Syed, Ralph Weischedel, Tan Xu, and David Yarowsky. Cross-document coreference resolution: a key technology for learning by reading. AAAI 2009 Spring Symposium on Learning by Reading and Learning to Read, Stanford, California. (pp.65-70)

Tan Xu, Douglas W. Oard, Tamer Elsayed, and Asad Sayeed. Knowledge representation from information extraction. Joint Conference on Digital Libraries (JCDL 2008), Pittsburgh, Pennsylvania.

Asad Sayeed. Minimalist parsing of subjects displaced from embedded clauses in a free word order language. Annual meeting of the Association for Computational Linguistics (ACL 2005), Ann Arbor, USA. (student session pp.97-102, 37% acceptance rate)

Asad Sayeed and Stan Szpakowicz. Estimating suboptimal grammaticality from a small Latin corpus. Recent Advances in Natural Language Processing (RANLP 2005), Borovets, Bulgaria. (7 pages, 14% acceptance rate)

Asad Sayeed and Stan Szpakowicz. Developing a minimalist parser for free word order languages with discontinuous constituency. In José Luis Vicedo, Patricio Martínez-Barco, Rafael Muñoz, and Maximiliano Saiz, editors, Advances in Natural Language Processing, 4th international conference (EsTAL). Lecture Notes in Computer Science 3230. Springer-Verlag, 2004. (pp. 115-126)

Asad Sayeed and Amy Weinberg. Complementizer-stranding by phase.. (59 pages)

Chia-Jung Tsui, Ping Wang, Kenneth Fleischmann, Douglas Oard, and Asad Sayeed. Exploring the relationships among ICTs: a scalable computational approach using KL-divergence and Hierarchical Clustering. HICSS-43 (2010). (10 pages)

Tenure-track faculty position in the Department of Philosophy, Linguistics, and Theory of Science, affiliated with CLASP, the Center for Linguistic Theory and Studies in Probability. Teaching computational linguistics, particularly in its machine-learning aspect. Strong emphasis on cognitive modeling and semantics.

• Current research focus is on incremental ambiguity resolution of higher-order semantic operators in sentence processing.

Dr. Demberg's group looks at incremental sentence processing from both a computational and psycholinguistic perspective, focusing on a psycholinguistically-plausible computational formalism, PL-TAG, as a means of modeling the human parser for both scientific and engineering purposes.

• Then focused initially on the construction of a semantic formalism suitable for PL-TAG incremental parsing.

The University of Maryland's Computational Linguistics and Information Processing Laboratory participates along with the College of Library and Information Studies in the PopIT Project. This project develops techniques to use text mining and sentiment analysis over the IP trade press to answer sociological questions about the propagation of technologies and attitudes towards them.

• Participated in the construction of a corpus of IT trade press articles suited towards topic classification, sentiment analysis, and other techniques.
• Constructed engine to measure relationships between IT-related discourses over time via KL-divergence.
• Work on sentiment analysis and opinion mining over magazine articles: exploring techniques to match opinion holders to their opinions and the targets of their opinions.
• Work on domain-adaptation of named-entity system.
• Supervision of two undergraduate programming assistants.
• Design and supervision of multiple annotation tasks.

I worked with IBM's T. J. Watson Research Center department for Services Research, most specifically the Engagement Information Leverage project. This project focused on advanced techniques to mine IBM's internal business documents to improve IBM's profitability.

• Win strategy mining: using patterns to extract text that reflects successful business strategies. I worked on ranking criteria.
• Statement-of-work compliance analysis: IBM contract negotiators produce statements-of-work based on templates. This project involved developing a novel technique to using text information retrieval tools such as Latent Semantic Analysis to determine the extent to which templates where being followed.
• Paper published in proceedings of CIKM 2009: Characteristics of document similarity measures for compliance analysis.

The University of Maryland's Computational Linguistics and Information Processing Laboratory is a partner with Johns Hopkins University and other institutions in the Human Language Technology Center of Excellence. I am involved in their project on multi-document named-entity linking.

• Developed edit distance-based entity-linking baseline experimental system.
• Worked on design of UIMA-based document preprocessing and experiment organization architecture.
• Conducted research into the use of text categorization for boosting scores of cross-document entity links.

IBM's T. J. Watson Research Center has a large research team involved in work on leading-edge Statistical Machine Translation techniques.

• Worked on English-to-Spanish real-time translation system.
• Research problem: correcting inflectional morphology (e.g. tense, gender) errors in automatic Spanish translation.
• Built system with Viterbi decoder and several different language and translation models, e.g., English source part-of-speech tags, Spanish morphological lexica.
• Presented experimental results at IBM and UMD.

The Espial Group develops API suites for developing consumer applications for embedded systems in Java.

• Responsible for Micro Edition API suite, including design, programming, and testing.
• Improved browser and email components as well as personal information manager widgets.
• Redesigned user interface.

OTI, now part of IBM Ottawa Labs, was an IBM-owned firm that creates object-oriented software development environments, including VisualAge for Java and Eclipse. I worked on the team that was just initiating development of the Eclipse environment.

• Product testing/release phase: tested product with team, wrote scenario tests/examples, suggested UI changes, corrected and updated documentation. (Product itself is confidential.)
• Designed and wrote integrated, extensible team regression testing/reporting system: almost completely independent on own large project.

• Ported UI testing framework between environments.
• Updated configuration management software to automate routine tasks
• Participated in testing phases

2016-2017 - Program co-chair for Cognitive Modeling and Computational Linguistics (CMCL) workshop at European Association for Computational Linguistics 2017 annual meeting, Valencia, Spain.

2016-2017 - Program co-chair for the computational linguistics poster session of the German Linguistics Society annual meeting.

2015 - Admissions committee member for Erasmus Mundus European Masters Program in Language and Communications Technology (LCT).

2015 - Admissions committee member for Saarland University Masters Program in Language Science and Technology (LST).

2012-present - program committee member for several conferences/workshops; reviewer for several journals.

2012-present - organizer of Saarland University "FEAST" lecture series in language science research.

2010-2011 - appointed to UMD Senate Faculty Affairs committee, served on sub-committee on working conditions of non-tenure-track instructors/lecturers.

2009-2010 - elected member of UMD Computer Science Department Council.

2009-2009 - elected member of UMD Computer Science Department education committee.

March 2008 - organized hospitality for SIGIR 2008 program committee meeting at UMD iSchool.

Summer 2007 - ran graduate discussion group on formal grammars and mathematical linguistics

Technology

• Python, C++, Java, Smalltalk, x86 Assembly language, Perl, Pascal, SQL, HTML, XML, XSLT, Prolog, Scheme, NetLogo, Ruby, OCaml, Javascript, Scala, and others.
• VisualAge for Java, VisualAge for Smalltalk, VisualAge Micro Edition, Borland C++, Visual C++, Emacs, gcc, ObjecTime, MIT Scheme, SWI Prolog.
• UNIX (Linux, X/Windows, etc.), OS/2, Windows XP/NT/98/95.

Other

• Writes very readable code, comments, skilled at writing documentation, very good with English.
• Languages: English (perfectly fluent), French (reasonable grammar and vocabulary). Some German experience (several years of classes). Some Urdu (maternal language, cultural experience as well). Experience with ancient Greek (evening classes) and Latin (self-study).
• Extensive study in linguistics, particularly syntactic theory, as well as computational linguistics and artificial intelligence. Machine learning, statistical and symbolic NLP, various semantic and syntactic formalisms.