We study Humanitarian and Social Informatics via the interdisciplinary science of design, analysis, and evaluation of information systems to augment the information processing capabilities of organizational actors that are concerned with the welfare and governance of society. Our research aim is to create new knowledge in the semantic, structured information representation and processing of unstructured data and designing human-centered computing solutions to the fundamental challenge of information overload in realtime decision support for organizations.
Our lab researches basic human behavior modeling problems to address the information overload challenge for organizations (check projects) by mining online social and Web as well as offline data sources using NLP and machine learning methods guided by domain knowledge bases and socio-psychological theories. We strive to conduct foundational research and evaluate that with real-world applications in the crisis management cycle from mitigation to response and recovery of natural (e.g., hurricanes), societal (e.g., migration, gender stereotyping) and human (e.g., terrorism, cyber warfare) crises.
Datasets for academic and research communities are available here: https://ist.gmu.edu/~hpurohit/informatics-lab/crisis-data.html
Cooperation between Humans and AI systems to Enhance City Services
(Intent Mining and Realtime Social Computing)
Social media, Web and Internet of Things have revolutionized information generation, transmission, and consumption, especially unstructured data, which has presented an information overload challenge for answering a variety of questions for organizational sensemaking. This area of projects investigate the basic questions of organizational decision support and cooperation using novel behavioral data analytics methods, such as how to quantify actionable information attributes in the data streams to help articulated organizational work and rank serviceable public posts, how to extract information for such articulation tasks, whom to trust and engage in the online community for information, etc.
Collaborators: Fairfax County First Responders, Virginia; Montgomery County CERT
Project Supported by: NSF award #1657379 (CRII: CHS: Mining Intentions on Social Media to Enhance Situational Awareness of Crisis Response Organizations), NSF award #1815459 (III: Summarizing Heterogeneous Crowdsourced & Web Streams Using Uncertain Concept Graphs), NSF award # 2029719 (RAPID: Human-AI Teaming for Big Data Analytics to Enhance Response to the COVID-19 Pandemic), NSF award # 2043522 (SCC-CIVIC-PG Track B: Assessing the Feasibility of Systematizing Human-AI Teaming to Improve Community Resilience), and WNRI & Research Council of Norway award #261685 (INTPART: Transnational Partnership for Excellent Research and Education in Big Data and Emergency Management)
User attitudes & Engagement
(Semantic Computing & Machine Learning)
Do you know that gender-based violence is a public health crisis and what are its impact on the policies and healthcare system? Do you wonder why people migrate from certain regions and what are the implications of migration? The objective of this project is to reduce information overload for responding (non-)governmental organizations in gleaning insights about public awareness and attitudes towards societal crises prevalent in the modern age (e.g., views on violence against women and narratives about refugee migration), in order to inform the organizations in their evaluation of awareness campaigns, intervention support, and policy design practices.
Project Partially Supported By: GMU Startup Funds and NSF award #1707837 (EAGER: Social Media Participation as Indicator of Actors, Awareness, Attitudes, and Activities Related to STEM Education).
Comprehension & Cognitive Load
Can an organization mitigate the crisis of losing intellectual property in response to human-created crisis of cyber attacks? This project investigates how we can create information overload for hackers to prevent them from stealing information by better understanding of information comprehension and cognitive load. Such an approach can improve cyber defense tactices for human crisis management as well as create opportunities to design tools and techniques for managing information overload in any tech-assisted environment.
Project Supported By: Office of Naval Research award #N00014-16-1-2896 (Believable Fake Documents)
- Prof. Hemant Purohit
- Rahul Pandey (PhD student, full time)
- Yasas Senarath (PhD student, full time)
- Jitin Krishnan (co-advising PhD student, full time)
- Cooper Jessup (PhD student, part time)
- Dr. Prakruthi Karuna (PhD; currently: Research Scientist, Perspecta)
- Dr. Habib Karbasian (PhD, co-advised)
- Gaurav Bahl (MS; currently: Senior Analyst, Citi)
- Ganesh Nalluru (MS; currently: Senior Data Scientist, DevelapMe)
- Yogen Chaudhari (MS; currently: Senior Data Scientist, Comscore)
- Sharan Banola (MS; currently: Senior Data Analyst, Capital One)
- Voravan Charnsawat (MS; currently: SLS Engineer, Qualitest)
- Mohammad Rana (BS; currently: Senior Software Engineer, Capital One)
- Hridoy Dutta (Visiting PhD Scholar, IIIT Delhi)