SocialSim investigated computational approaches to modeling and forecasting online social behavior. My work centered on a multi-platform simulation framework and the Multi-Action Cascade Model, connecting behavioral assumptions about influence with scalable implementation.
DARPA’s Social Simulation for Evaluating Online Messaging Campaigns (SocialSim) program funded research into computational frameworks for modeling and forecasting how online social behavior — and the spread of information through it — unfolds across platforms and over time.
As part of the UCF Complex Adaptive Systems Lab team, Chathura developed a programming framework in NetLogo for simulating information spread and evolution across multiple social media platforms, including Twitter, Reddit, GitHub, YouTube, and Telegram.
Working in a team of two, Chathura co-originated the Multi-Action Cascade Model (MACM) — a framework in which each type of user action (e.g., posting, replying, resharing) exerts a distinct influence on the receiver, contributing the behavioral premise that influence varies by relationship and content domain.
MACM’s core idea is that not all actions carry equal behavioral weight: the influence one user’s action has on another depends on the type of action and the domain of the content involved, rather than treating every interaction identically.
MACM was implemented jointly in NetLogo and, in the same two-person team, ported to a CUDA environment using Python — reducing execution time from roughly 15 days to roughly 3 hours. From 2021 onward, Chathura became sole developer of MACM, independently extending it with Bayesian inference of influence parameters and Markovian context handling, so that influence estimates update as evidence accumulates and an agent’s response depends on the recent history of its interactions rather than each message in isolation. Additional implementations followed in C++/Repast.
Chathura was sole developer of the multi-platform simulation framework, co-originated MACM, and was its sole developer for the post-2021 extensions, which included the CUDA performance work and the Bayesian/Markovian modeling extensions described above. This work was presented as a major component of the deliverables submitted to the DARPA SocialSim program across its multi-year funding cycle.
MACM’s premise — that influence is not uniform across interaction types — carries directly into Chathura’s subsequent work on transfer-entropy-based influence measurement for DARPA MIPS.
My contribution
Chathura was sole developer of the multi-platform SocialSim simulation framework, co-originated the Multi-Action Cascade Model in a team of two, and became its sole developer from 2021 onward, extending it with Bayesian inference of influence parameters and Markovian context handling.
Collaborators
Publications
2019
Gunaratne, C., Senevirathna, C., Jayalath, C., Baral, N., Rand, W., Garibay, I.
5th International Conference on Computational Social Science
Introduces the Multi-Action Cascade Model (MACM), in which each action type exerts a distinct influence on the receiver — the founding publication behind the DARPA SocialSim simulation work.
Software
SocialSim / MACM Framework
Simulation framework
Lead developer
A DARPA-funded multi-platform social-media simulation environment. Chathura was sole developer of the simulation framework covering Twitter, Reddit, GitHub, YouTube, and Telegram, and co-originated the Multi-Action Cascade Model (MACM) in a team of two, later becoming its sole developer from 2021 onward.
Related