Computational Social Dynamics Lab

Archived Lab News & Research Highlights

• 2023/12 Paper accepted to appear in IEEE Big Data 2023: "HungerGist: An Interpretable Predictive Model for Food Insecurity." This paper presents HungerGist, a multi-task deep learning model that can both predict and interpret food insecurity from association with text data.

• 2023/12 Congrats to Ahmad, for passing his Comprehensive Exams!

• 2023/08 Congrats to Yu-Ru Lin on her new position as the Research & Academic Director, Institute for Cyber Law, Policy and Security (Pitt Cyber). Learn more about what motivates her research, broader impact on society, and ways to combat online misinformation here.

• 2023/07 Paper accepted to appear in IEEE VIS 2023: "VISPUR: Visual Aids for Identifying and Interpreting Spurious Associations in Data-Driven Decisions." This paper proposes VISPUR, a novel visual analytic system that helps interpret spurious associations, especially Simpson’s paradox, through the de-paradox workflow to combat confounding bias and examine heterogeneous subgroup patterns. Demo.

• 2023/07 Paper accepted to appear in SBP-BRiMS 2023: "The Dynamics of Political Narratives During the Russian Invasion of Ukraine." This work sheds light on how actors coordinate to advance particular viewpoints or act against one another in the context of global conflict.

• 2023/04 Congrats to Yongsu, for passing his Proposal Defense!

• 2023/03 Congrats to Nefriana, for passing her Preliminary Exams!

• 2023/02 Congrats to Muheng, for passing his Comprehensive Exams!

• 2021/03 Our paper received the Best Paper Award for 2020 in ACM Transactions on Interactive Intelligent Systems. Congrats to Yongsu! The award committee recognized our "Policy Flow: Interpreting Policy Diffusion in Context" as "... this paper demonstrates the advantages of interactive visual analytics for understanding the policy adoption and exemplifies the two key characteristics of an interactive intelligent system: intelligence and interactivity." Read more about the award and the paper.

• 2020/10 Paper accepted to appear in COLING 2020: "Inflating Topic Relevance with Ideology: A Case Study of Political Ideology Bias in Social Topic Detection Models." This work highlights the susceptibility of large, complex models to propagating the biases from input, leading to a deterioration of retrieval accuracy, and the importance of controlling for these biases. As a way to mitigate the bias, we propose to learn a text representation that is invariant to political ideology while still judging topic relevance.

• 2020/10 Congrats to Ang, for her paper with Rosta and collaborators receiving Honorable Mention Award from CSCW 2020!

• 2020/08 Paper accepted to appear in Humanities and Social Sciences Communications: "The dynamics of Twitter users' gun narratives across major mass shooting events." This article reveals how cross-cutting political talks about guns and gun policies happened on Twitter in a time of mass shootings.

• 2020/08 Paper accepted to appear in Big Data Research: "Data-Driven Computational Social Science: A Survey." This article presents a survey on data-driven computational social science, focusing on the state-of-the-art research on human dynamics from three aspects: individuals, relationships, and collectives.

• 2020/07 Paper accepted to appear in IEEE Journal of Biomedical and Health Informatics (JBHI): "StoCast: Stochastic Disease Forecasting with Progression Uncertainty." Congrats to Xian for publishing a new journal article! This work presents a new deep generative learning framework to deal with the uncertainty scenarios often seen in medical and other time-evolving datasets. The code is available at our github page

• 2020/06 Paper accepted to appear in ACM Transactions on Interactive Intelligent Systems (TiiS): "PolicyFlow: Interpreting Policy Diffusion in Context." Congrats to Yongsu for publishing a new journal article! This new visualization combines text and network analyses to help understand the diffusion of laws and regulations across political boundaries in the US.

• 2020/03 Yongsu's research was highlighted in the DATA STORIES podcast and there is also an interview featuring our IEEE VIS paper. The story is called "Visualizing Fairness in Machine Learning." The story introduced how visualization can play a role in its detection and mitigation in the world of "machine bias."

• 2020/01 Paper accepted to appear in ICWSM 2020: "MimicProp: Learning to Incorporate Lexicon Knowledge into Distributed Word Representation for Social Media Analysis." Congrats to Muheng and the team! The code is available at our github page.

• 2020/01 Congrats to Ang, Kai and Muheng, for their papers accepted to appear in The Web Conference 2020!

• 2019/11 Congratulations to Xidao and Mert for successfully defending their PhD dissertations!

• 2019/07 Two papers accepted to appear at IEEE VIS 2019: "FairSight: Visual Analytics for Fairness in Decision Making" and "FacIt: Factorizing Tensors into Interpretable and Scrutinizable Patterns." Congrats to Yongsu, Xidao and the team!

• 2019/06 Paper accepted to appear in ECML/PKDD 2019: "CASTNet: Community-Attentive Spatio-Temporal Networks for Opioid Overdose Forecasting." The deep learning code used in this paper is available at our github page.

• 2019/05 Our new paper on opioid prediction was covered by VentureBeat: "AI can predict opioid overdoses from crime and socioeconomic data".

• 2019/02 Paper accepted to appear at EPJ Data Science Journal: "Activism via Attention: Interpretable Spatiotemporal Learning to Forecast Protest Activities." Congrats to Mert and the team! The deep learning code used in this paper is available at the github page.

• 2019/01 Paper accepted to appear in WWW 2019: "Iterative Discriminant Tensor Factorization for Behavior Comparison in Massive Open Online Courses." Congrats to Xidao and the team!

• 2018/12 Two Postdoc Positions in Computational Social Science are available.

• 2018/04 Paper accepted to appear in IJCAI 2018: "Deep into Hypersphere: Robust and Unsupervised Anomaly Discovery in Dynamic Networks." The deep learning code used in this paper is available at our github page.

• 2018/02 Paper accepted to appear in ACM TKDD: "Event Analytics via Discriminant Tensor Factorization."

• 2017/11 Paper published at EPJ Data Science Journal: "Measuring and monitoring collective attention during shocking events." Check out the SpringerOpen blog post that featured this article. The dataset associated with our findings has been released.

• 2017/11 Dr. Adam Perer to visit us and speak in Big Data Colloquium Series on November 17, 2017.

• 2017/11 Code release -- MTHL (Multi-view Time-Series Hypersphere Learning) for detecting anomalous patterns from dynamic and multi-attributed networks. Check out the github page and our paper!

• 2017/10 Dr. Dan Cosley to visit us and speak in Big Data Colloquium Series on October 27, 2017.

• 2017/10 Our studies (about fear & distress) were cited by an article in Las Vegas Sun to explain how the latest mass shooting in Las Vegas could trigger many's lasting emotional reactions regardless of where they live.

• 2017/10 The paper "Quantifying Content Polarization on Twitter" won the Best Student Paper Award in CIC 2017. Congrats to Muheng and Xidao!

• 2017/09 Dr. Leman Akoglu to visit us and speak in Big Data Colloquium Series on September 29, 2017.

• 2017/09 Congratulations to Xian for receiving travel awards from NSF/SIGWEB/SIGIR to attend CIKM 2017!

• 2017/09 Paper to appear in CIC 2017: "Quantifying Content Polarization on Twitter." Congrats to Muheng and Xidao!

• 2017/08 Paper accepted to appear in CIKM 2017: "Anomaly Detection in Dynamic Networks Using Multi-View Time-Series Hypersphere Learning." Congrats to Xian and Xidao!

• 2017/07 Paper accepted to appear in VAST 2017: "Voila: Visual Anomaly Detection and Monitoring with Streaming Spatiotemporal Data." Watch the video preview of this work!

• 2017/05 Congratulations to Xidao Wen for successfully defending his dissertation proposal!

• 2017/05 Paper accepted to appear in WebSci 2017: "The Signals and Noise: Actionable Information in Improvised Social Media Channels During a Disaster" Congrats to Xingsheng and the team!

• 2017/03 Paper accepted to appear in Risk Analysis Journal: "Tracking and Analyzing Individual Distress Following Terrorist Attacks Using Social Media Streams."