Sensemaking at Scale: Solving Intelligence Analysis Mysteries with Crowdsourcing
Making sense of large amounts of information is difficult for both human and machine intelligence. Computational methods offer increasingly powerful capabilities for data analysis, but deciphering rich information encoded in the raw data is still AI-hard and requires human intelligence. Meanwhile, understanding and adopting machine learning models is difficult for domain experts. Crowdsourcing provides an effective paradigm to augment human intelligence and bridge the gap between information overload and individual competence. However, novice crowds usually struggle to accomplish cognitively demanding and complex goals. In this talk, I will focus on sharing my work on developing novel computational techniques to achieve holistic sensemaking through distributed contributions from crowds. I will also present my research in collaboration with Microsoft Research (MSR), Cloudera, and Informatica on facilitating human-AI interaction in domain-specific sensemaking.
Tianyi Li is a Ph.D. candidate in Computer Science at Virginia Tech. Her research combines methods from crowdsourcing, visual analytics and human-computer interaction (HCI) to support complex sensemaking. She develops novel mechanisms and software to enable individuals, crowds, and artificial intelligence (AI) to collaborate on analytic tasks in different critical domains. Tianyi’s work has been published in top-tier computer science journals and conference proceedings, such as ACM Computer Supportive Cooperative Work (CSCW), ACM Intelligent User Interfaces (IUI), ACM Human Factors in Computing Systems (CHI), and IEEE Visualization (VIS), and has resulted in two patents. Tianyi has also served on organizing and program committees at leading venues such as CHI, IUI, and ACM Recommender Systems (RecSys). She received her Bachelor’s degree in Computer Science from the University of Hong Kong in 2015. During her Ph.D., Tianyi interned with research labs at Microsoft Research (MSR), Cloudera, and Informatica.