- November, 2018: Just in time for the election, Jane showed that Russian trolls are still active on Twitter and trying to interact with major news reporters. Timely stuff!
- November, 2018: Jane presents her work on Wikipedia conflict resolution at CSCW and David and Zijian meet up in Brussels to talk about access support in online communities. The lab hopes that Jane will bring back Montreal-style bagels too.
- October, 2018: David is off to NWAV47 to talk about Computational Sociolinguistics! He came back with a mountain of Montreal-style Bagels and a new appreciation for the Northern Cities Vowel Shift.
- September, 2018: New PhD students Jane Im and Minje Choi arrive at UMSI! Welcome Jane and Minje!
- August, 2018: New EMNLP paper on supportive/unsupportive language accepted with undergraduate first author Zijian Wang! In online conversations, users who indicate they are women really do receive more unsupportive replies--yet they also receive more supportive replies. Lots of interesting follow-up questions on gendered interactions online #FoodForThought
- August, 2018: Visiting students Akshita Jha, Qi Sun, and Nan Gu depart physically but remain with us in spirit and co-authorship. Wonderful having you here with us this summer!
The Bringing together Language and Behavior for Large-scale Analytical Breakthroughs Laboratory (Blablablab)
Why do we say the things we do and what does our language say about us in return? The Blablablab develops new scientific methods to answer socially-impactful questions about human behavior from what we observe people say and do. Our work broadly intersects with computational social science, natural language processing, computational sociolinguistics, data mining, and social computing. Along the way, we release new software tools and datasets for the community to help reveal the social dimensions of language, and, with a bit of luck, make our online world a better place.
The Blablablab is officially a part of the University of Michigan School of Information (UMSI) and is a member of the Information Analysis and Retrieval (IAR) group within UMSI. We widely collaborate on interdisciplinary projects with researchers and students in the Departments of Computer Science and Linguistics, the School of Public Health, the Center for the Study of Complex Systems, and the Michigan Institute for Data Science. But mostly, we work with amazing people doing socially-impactful work.
Some Blablablabers: (top) Thomas Horak, Jane Im, Minje Choi, Lingding Chen, Yaoyang Lin, Carol Zheng, Shangming Zhao; (bottom) Justin Chen, David Jurgens, Nan Gu, Qi Sun
Professors
PhD Students
Masters Students
Stella Choi
(School of Information)
Undergraduates
Sanjana Kolisetty
(Computer Science)
Sam Lee
(Computer Science)
Trevor Li
(Computer Science)
Yaoyang Lin
(Computer Science)
Adi Mannari
(Computer Science)
Carol Zheng
(Computer Science)
Alumni
Undergradutes
Zijian Wang
(BS Computer Science, 2018)
First position after: MS Student in Symbolic Systems at Stanford
Xinyi Wu
(BS Computer Science, 2018)
First position after: MA Student in Computational Linguistics at the University of Washington
Visiting Students
Nan Gu
(BS Electrical Engineering @ Tsinghua University)
Akshita Jha
(MS Computer Science @ IIIT)
Qi Sun
(BS Computer Science @ Peking University)
Shangming Zhao
(BS Software Engineering @ Tsinghua University)
Language and Identity
Identities are complex and in any given situation, we may choose different language--or even pronunciation--to signal different aspects of who we are. A core project in the Blablablab is focused on the interplay between linguistic style and the construction of identity. Here, work in the emerging field of Computational Sociolinguistics to help bring together theories, observational analyses, and large scale models to advance the fields of Linguistics and NLP.
Promoting Healthier Interactions
Online environments can be very supportive or very toxic places. We believe that the absence of anti-social behavior doesn't necessary make for a positive online environment. Here in the Blablablab, we develop new methods to study what leads to healthy communities and how to nudge people to behave in more pro-social ways through both system design, social analysis, and NLP tools.
Information Engagement
Our world viewpoints are often shaped by the information we're exposed to. The Blablablab has multiple projects examining how people engage with different kinds of news sources and how this exposure influences our world views. We also examine this within the broader context, examining both multilingual and international exchanges of ideas.
People and Place
The internet brings people together but geography still plays a major role in our lives. In a series of interrelated projects, the Blablabab examines the impact geography plays on our social lives, how we can the characteristics and culture of a place, and how to support broader survey methods through location inference. We also like to make cool maps.
2018
 |
It's going to be okay: Measuring Access to Support in Online Communities. Zijian Wang and David Jurgens. Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), 2018 pdf · supplementary · website and data · code (all forthcoming) |
 |
RtGender: A Corpus of Responses to Gender for Studying Gender Bias. Rob Voigt, David Jurgens, Vinodkumar Prabhakaran, Dan Jurafsky, and Yulia Tsvetkov. Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC), 2018 pdf · data |
 |
Measuring the Evolution of a Scientific Field through Citation Frames. David Jurgens, Srijan Kumar, Raine Hoover, Dan McFarland, Dan Jurafsky. Transactions of the Association for Computational Linguistics (TACL). 2018. pdf · website and data · code |
We showed up to UMSI!
Software
- GenderPerformr: Infers gender performance for any kind of online username, name, handle, etc.
- Social Support: Infers how supportive, neutral or unsupportive is an online reply
- Citation Function: Infers why an author cited another paper
Data
- Citation Function Data: labeled training data of citations by their rhetorical function.
- Gender-Labeled Online Conversations: 100M online dyadic conversations from Reddit, Wikipedia, and StackExchange, labeled by gender salience. To respect the privacy and dignity of individuals, this dataset is available for non-commercial research purposes online; please email the lab PI to obtain access.
Graduate students
Prospective PhD graduate students interested in joining Blablablab should apply to one of our affiliated programs. We typically accept most students through UMSI though we occasionally will accept students who apply to the Computer Science division of EECS.
Current U-M graduate students of any program are welcome to email about potential research collaborations. Blablablab is highly interdisciplinary and we especially enjoy working with social scientists.
Undergraduates
Current U-M undergraduates who want to do research during the school year should contact the lab and describe a bit about their background (e.g., have you done research before? what classes have you taken? why do you want to research?) and what kind of project they want to be on. Scanning our current list of publications will give you a sense of what topics we research and can be a good jumping off point for a new project. We often have a few spots for the year for all levels of experience but please know we expect at least 10 hours per week of research. Research takes time and it will be difficult for you (or anyone) to make progress with just a few hours per week.
Undergraduates interested for summer research opportunities should look for an announcement sometime in the early to mid-winter period with a description of the projects. We have hosted lots of wonderful undergraduate students throughout the summer. Women and underrepresented minorities are encouraged to contact us early, as we often can apply for special summer funding opportunities (e.g., through the B.A. Rudolph Foundation or the NSF) to provide financial support for your summer stay.
Visiting Students
We often have self-funded visiting students for the summer and (rarely) during the academic year. In these cases, usually the student has some prior research experience, similar research interests, and (due to luck) there is a current project going on in the Blablablab where they would be a good fit. If you are a self-funded student who wants to come join us for the summer, you should send an email in February or March before the summer that
- Describes your research experience and state what parts are relevant to the work going on in the Blablablab.
- Clearly states why you want to work in the Blablablab.
- Discusses what you want to get out of a summer research experience and what you want to learn — this helps us make sure that the trip is a success for you as well.
We don't consider self-funded visiting students as "free labor" and strongly want to make sure that your stay is productive and a success for your career goals.
Postdocs
We have no openings at this time for postdocs. :(