June 27, 2014
Baltimore, MD, USA
8:50 - 9:50 Invited talk by Derek Ruths "The Promises and Pitfalls of Demographic Inference in Social Media"
9:50 - 10:45 Oral Session 1:
10:45 - 10:55 Coffee break
10:55 - 11:30 Oral Session 2:
11:30 - 12:30 Panel (Hanna Wallach, Jacob Eisenstein, Cristian Danescu-Niculescu-Mizil and Jimmy Lin)
12:30 - 1:30 Lunch Break
1:30 - 2:25 Oral Session 3:
2:25 - 2:35 Coffee break
2:35 - 3:35 Invited talk by Mark Dredze "Opportunities from Social Media Data for Public Health"
3:35 - 3:45 Coffee Break
3:45 - 4:30 Oral Session 4:
4:30 - 5:30 Poster Session
5:30 - 5:40 Closing Remarks
The list of accepted papers (oral talks and posters) is available here.
The Workshop on Latent Attribute Prediction in Social Media and the Workshop on NLP and Social Dynamics have merged!
NLP and Social Dynamics Aspect: Language is a set of publicly agreed conventions that serves the purpose of inter-personal communication. Speakers (or writers) try to convey a message, instill an idea or make an impression on the listeners. Listeners (or readers), in turn, are affected by the message and may respond to it. Language, in that sense, is an important vehicle that shapes (and is shaped by) social dynamics.
Traditional NLP research, however, focuses on "documents" (either of full length or on the sentence level), rather than on the communication process as reflected by language use. Common examples of traditional NLP research are parsing, document classification, machine translation, and sentiment analysis at the sentence and document level without considering the social dynamics of the people who are writing and reading those texts.
We propose to move beyond analyzing the informational aspect of documents and discuss ways in which NLP can contribute to gaining insights about the interplay between language use and various levels of social dynamics.
Personal Analytics in Social Media Aspect: There are many important social science questions and commercial applications that are impacted by the large amounts of diverse personalized data emerging from social media. These data can reveal user interests, preferences and opinions, as well as trends and activity patterns for companies and their products.The automatic prediction of latent attributes from discourse in social media includes topics such as:
We invite original and unpublished research papers on all topics related to text-driven latent attribute prediction in social media, including but not limited to the topics listed below:
NLP and Social Dynamics: