CS7180: Special Topics in AI: Text Modeling for the Humanities and Social Sciences

Fall 2017

Class meeting: Tuesdays, 11:45-1:25, and Thursdays, 2:50-4:30, Ryder 124

Instructor: David Smith, Assistant Professor in Computer and Information Science (Office Hours: Thursdays, 12-2, or by appointment; WVH 356)

Course Description

Researchers and archivists have been digitizing the source materials for human history and culture for over half a century, but two further factors sped the emergence in the last decade of the digital humanities and computational social sciences. First, industrial scale scanning projects have increased the available evidence beyond the ability of individual scholars to manage them; second, born-digital traces of our social, cultural, economic, and political lives have become practically archivable and searchable on a massive scale. Much of this data is text—“unstructured” as the database people might say—providing opportunities for advances in natural language processing.

In this seminar, we will read and discuss papers about building models of text to answer questions in the humanities and social sciences. Students will take turns presenting and leading discussion of papers along with the relevant background material. All students will write short reviews of the papers we read and complete a course project and accompanying report.

Prerequisites

There are no official prerequisites; however, it is expected that students have some background either in NLP, in machine learning, or in working with text computationally in the humanities or social sciences.

Syllabus

Each week, we will read roughly two papers on a common theme. The papers could be tied together by methodology—e.g., text categorization or convolutional neural networks—or by subject matter—e.g., criminology or plot analysis.

1. September 7: Introduction: Human language and culture meet Big Government, Big Business, Big Science, and Big Data. I ended up talking about several books from, modally, 1983:
   • Walter J. Ong. Orality and Literacy: The Technologizing of the Word, Routledge, 1982.
   • Elizabeth Eisenstein. The Printing Press as an Agent of Change: Communications and Cultural Transformations in Early-Modern Europe, Cambridge University Press, 1979.
   • Ernest Gellner. Nations and Nationalism, Cornell University Press, 1983.
   • James C. Scott. Seeing Like a State: How Certain Schemes to Improve the Human Condition Have Failed, Yale University Press, 1998.
   • Benedict Anderson. Imagined Communities: Reflections on the Origins and Spread of Nationalism, Verso, 1983.
   • Georges Lefebvre. John Albert White, trans. The Great Fear of 1789: Rural Panic in Revolutionary France, Princeton University Press, 1983.
2. September 12: Models of Text in the Social Sciences and Humanities
   • Justin Grimmer and Brandon M. Stewart. Text as data: The promise and pitfalls of automatic content analysis methods for political texts. Political Analysis, 21:267–297, 2013. (McCabe)
   • Lara Putnam. The transnational and the text-searchable: Digitized sources and the shadows they cast. American Historical Review, 121(2):377–402, 2016. (Muther)
   • Ted Underwood. Theorizing research practices we forgot to theorize twenty years ago. Representations, 127:64–72, 2014. (Murali)
3. September 14: Bags of words and other text representations
   • Matthew J. Denny and Arthur Spirling. Text Preprocessing For Unsupervised Learning: Why It Matters, When It Misleads, And What To Do About It. Working paper, 2017. (MacLaughlin)
   • Jacob Eisenstein. What to do about bad language on the internet. In Proceedings of the North American Association for Computational Linguistics, 2013. (Sreekumar)
   • See also: Alexandra Schofield and David Mimno. Comparing apples to Apple: The effects of stemmers on topic models. Transactions of the Association for Computational Linguistics, 4:287–300, 2016.
4. September 19: Word vectors and distributed representations
   • Ryan James Heuser. Word Vectors in the Eighteenth Century. In Digital Humanities, 2017. See also a bloggier presentation, with more visualizations (Sreekumar).
   • Shyam Upadhyay, Manaal Faruqui, Chris Dyer, and Dan Roth. Cross-lingual models of word embeddings: An empirical comparison. In Processings of the Association for Computational Linguistics, 2016 (Gallagher).
   • See also: Sebastian Ruder. A survey of cross-lingual embedding models, arXiv, 2017.
5. September 21: Text categorization
   • Ruchita Sarawgi, Kailash Gajulapalli, and Yejin Choi. Gender Attribution: Tracing Stylometric Evidence Beyond Topic and Genre. In Proceedings of the Conference on Computational Natural Language Learning, 2011. (Nye)
   • Daniel Hopkins and Gary King. A method of automated nonparametric content analysis for social science. American Journal of Political Science, 54(1):229–247, 2010. (Gundogdu)
6. September 26: Language models and topic models
   • Lisa M. Rhody. Topic modeling and figurative language. Journal of Digital Humanities, 2(1), 2012. ()
   • Benjamin M. Schmidt. Words alone: Dismantling topic models in the humanities. Journal of Digital Humanities, 2(1), 2012. (Shugars)
   • Alexandra Schofield, Laure Thompson, and David Mimno. Quantifying the effects of text duplication on semantic models. In Empirical Methods in Natural Language Processing, 2017. (MacLaughlin)
7. September 28: Dynamic models and temporal change
   • William L. Hamilton, Jure Leskovec, and Dan Jurafsky. Diachronic word embeddings reveal statistical laws of semantic change. In ACL, 2016. (Nye)
   • Cristian Danescu-Niculescu-Mizil, Robert West, Dan Jurafsky, Jure Leskovec, and Christopher Potts. No country for old members: User lifecycle and linguistic change in online communities. In WWW, 2013. (Robertson)
8. October 3: Entity and relation extraction
   • Robert West and Jürgen Pfeffer. Armed conflicts in online news: A multilingual study. In ICWSM, 2017. (Sherkati)
   • Katherine Keith, Abram Handler, Michael Pinkham, Cara Magliozzi, Joshua McDuffie, and Brendan O'Connor. Identifying civilians killed by police with distantly supervised entity-event extraction. In EMNLP, 2017. (Gallagher)
9. October 5: Discuss project ideas
10. October 10: Plot and character
    • David Bamman, Ted Underwood, and Noah A. Smith. A Bayesian Mixed Effects Model of Literary Character. In Proceedings of the Association for Computational Linguistics, 2014. (McCabe)
    • Mohit Iyyer, Anupam Guha, Snigdha Chaturvedi, Jordan Boyd-Graber, and Hal Daumé III. Feuding Families and Former Friends: Unsupervised Learning for Dynamic Fictional Relationships. In Proceedings of the North American Association for Computational Linguistics, 2016. (Dong)
11. October 12: Language and power relations
    • Cristian Danescu-Niculescu-Mizil, Lillian Lee, Bo Pang and Jon Kleinberg. Echoes of power: Language effects and power differences in social interaction. In WWW, 2012. (Robertson)
    • Vinodh Krishnan and Jacob Eisenstein. “You're Mr. Lebowski, I'm The Dude”: Inducing address term formality in signed social networks. In NAACL, 2015. (Gundogdu)
12. October 17: Geographical and social variation
    • Su Lin Blodgett and Brendan O'Connor. Racial disparity in natural language processing: A case study of social media African-American English. In Fairness, Accountability, and Transparency in Machine Learning (FAT/ML) workshop at KDD, 2017. (Murali)
    • Dong Nguyen and Jacob Eisenstein. A Kernel Independence Test for Geographical Language Variation. Computational Linguistics, 2017. (Hou)
13. October 19: Dialogue and argumentation
    • Stephanie Lukin, Pranav Anand, Marilyn Walker, and Steve Whittaker. Argument strength is in the eye of the beholder: Audience effects in persuasion. In EACL, 2017. (Pandya)
    • Chenhao Tan, Vlad Niculae, Cristian Danescu-Niculescu-Mizil, and Lillian Lee. Winning arguments: Interaction dynamics and persuasion strategies in good-faith online discussions. In WWW, 2016. (Shugars)
14. October 24: Text Reuse
    • Cristian Danescu-Niculescu-Mizil, Justin Cheng, Jon Kleinberg, and Lillian Lee. You had me at hello: How phrasing affects memorability. In ACL, 2012. (Sherkati)
    • Maria Moritz, Andreas Wiederhold, Barbara Pavlek, Yuri Bizzoni, and Marco Büchler. Non-literal text reuse in historical texts: An approach to identify reuse transformations and its application to Bible reuse. In EMNLP, 2016. (Muther)
15. October 26: Information cascades
    • David Liben-Nowell and Jon Kleinberg. Tracing information flow on a global scale using Internet chain-letter data. PNAS, 105(12):4633–4638, 2008. (MacLaughlin)
    • Rahul Goel, Sandeep Soni, Naman Goyal, John Paparrizos, Hanna Wallach, Fernando Diaz, and Jacob Eisenstein. The Social Dynamics of Language Change in Online Networks. In The International Conference on Social Informatics (SocInfo), 2016. (Nye)
16. October 31: Community structure and communication
    • Dongxiao He, Zhiyong Feng, Di Jin, Xiaobao Wang, and Weixiong Zhang. Joint identification of network communities and semantics via integrative modeling of network topologies and node contents. In AAAI, 2017. (Gallagher)
    • Weiwei Yang, Jordan Boyd-Graber, and Philip Resnik. A discriminative topic model using document network structure. In ACL, 2016. (Sherkati)
17. November 2: Exploratory Data Analysis
    • Gary King, Patrick Lam, and Margaret E. Roberts. Computer-assisted keyword and document set discovery from unstructured text. American Journal of Political Science, 2017. (Gundogdu)
    • Jacob Eisenstein. Unsupervised learning for lexicon-based classification. In AAAI, 2017. (Murali)
18. November 7: Document analysis and recognition
    • Maria Ryskina, Hannah Alpert-Abrams, Dan Garrette, Taylor Berg-Kirkpatrick. Automatic compositor attribution in the First Folio of Shakespeare. In ACL, 2017. (Sreekumar)
    • Verónica Romero, Alicia Fornés, Nicolás Serrano, Joan Andreu Sánchez, Alejandro H. Toselli, Volkmar Frinken, Enrique Vidal, and Josep Lladós. The ESPOSALLES database: An ancient marriage license corpus for off-line handwriting recognition. Pattern Recognition 46(6):1658–1669, 2013. (Gundogdu)
19. November 9: Preliminary project presentations
20. November 14: Laws and legislatures
    • Kellen R. Funk and Lincoln A. Mullen. The spine of American law: Digital text analysis and U.S. legal practice. American Historical Review, 2018. (McCabe)
    • Alexander T. J. Barron, Jenny Huang, Rebecca L. Spang, and Simon DeDeo. Individuals, Institutions, and Innovation in the Debates of the French Revolution. arXiv, 2017.
21. November 16: Crime and Harm
    • Rob Voigt, Nicholas P. Camp, Vinodkumar Prabhakaran, William L. Hamilton, Rebecca C. Hetey, Camilla M. Griffiths, David Jurgens, Dan Jurafsky, and Jennifer L. Eberhardt. Language from police body camera footage shows racial disparities in officer respect. Proceedings of the National Academy of Sciences, 114(25):6521–6526, 2017. (Gallagher)
    • Munmun De Choudhury, Emre Kiciman, Mark Dredze, Glen Coppersmith, and Mrinal Kumar. Discovering shifts to suicidal ideation from mental health content in social media. In CHI, 2016. (Robertson)
22. November 21: Causal Inference
    • Naoki Egami, Christian J. Fong, Justin Grimmer, Margaret E. Roberts, and Brandon M. Stewart. How to make causal inferences using texts. In Political Methods, 2017. (McCabe)
    • Michael Paul. Feature Selection as Causal Inference: Experiments with Text Classification. In CoNLL, 2017. (Pandya)
23. November 23: Thanksgiving: No Class
24. November 28: Final project presentations
    • MacLaughlin
    • Gallagher
    • Murali and Sreekumar
    • Robertson
25. November 30: Final project presentations
    • McCabe and Sherkati
    • Muther
    • Gundogdu
    • Nye
    • Pandya
26. December 5: Some other topic
27. December 7: Interpretation
    • Lisa Samuels and Jerome McGann. Deformance and interpretation. New Literary History, 30(1): 25–56, 1999.
    • Lucas Sterckx, Jason Naradowsky, Bill Byrne, Thomas Demeester, and Chris Develder. Break it down for me: A study in automated lyric annotation. In EMNLP, 2017.