Detecting misinformation
Using LLMs, machine learning, and network science to spot emerging misinformation trends as they spread across platforms.
CIRTA Impact+ Postdoctoral Fellow · McMaster University
Studying how opinions and information (mostly misinformation) spread through the networks that shape public discourse.
I'm Sriniwas Pandey, a computational social scientist at McMaster's Digital Society Lab. I use large language models, machine learning, and network science to detect misinformation trends as they spread across online platforms, and to understand how these dynamics affect the health of democratic discourse.
McMaster University
Research
I study online behavior, opinion dynamics, and social systems shaped by algorithms: how platform design, recommendation, and network structure shape what people see, what gets attention, and what becomes controversial. My dissertation at Binghamton University (advised by Prof. Hiroki Sayama) introduced formal measures of opinion eccentricity and showed why off-center viewpoints attract disproportionate attention online.
Research themes
LLM-mediated argumentation
Controlled experiments on how LLM agents shift framing, sentiment, and eccentricity when they face sustained argumentative pushback.
Diversity-aware recommendation
Recommendation interventions that surface cross-cutting content through social neighbors first, to widen what people are exposed to.
Credibility via collective intelligence
Incentive-compatible, prediction-market-inspired aggregation for judging how credible a claim is.
Eccentricity & controversiality
Measures of collective and self-eccentricity, tested on data from GAB and Parler, linking off-center opinions to controversy and attention.
Recommendation & opinion diffusion
How recommendation strategies, homophily, and novelty bias shape how opinions spread, connectivity, innovation, and polarization.
Methods
- Agent-based modeling
- Network analysis
- NLP & LLMs
- Social media data
- Python · R
- HPC · Docker
- Reproducible workflows
Publications
In the news
Talks & Presentations
Teaching
Binghamton University & Cornell University
I've designed and taught undergraduate and graduate courses for large cohorts, bringing computational social science, reproducible workflows, and project-based mentorship into the classroom.
Roles
Experience & Education
Positions
Education
Service
Honors & Awards
Get in touch
I'm always happy to talk about misinformation, opinion dynamics, or possible collaborations.
Affiliation
Digital Society LabMcMaster University
Hamilton, Ontario, Canada