Deep Agent: Studying the Dynamics of Information Spread and Evolution in Social Networks

Abstract

This paper explains the design of a social network analysis framework, developed under DARPA's SocialSim program, with novel architecture that models human emotional, cognitive and social factors. Our framework is both theory and data-driven, and utilizes domain expertise. Our simulation effort helps understanding how information flows and evolves in social media platforms. We focused on modeling three information domains: cryptocurrencies, cyber threats, and software vulnerabilities for the three interrelated social environments: GitHub, Reddit, and Twitter. We participated in the SocialSim DARPA Challenge in December 2018, in which our models were subjected to an extensive performance evaluation for accuracy, generalizability, explainability, and experimental power. This paper reports the main concepts and models, utilized in our social media modeling effort in developing a multi-resolution simulation at the user, community, population, and content levels.

A model simulation of political segmentation through an estimation of the entropy

“…Even if the voting opinions of citizens are many the behaviors that arise from those opinions become a challenge to estimate not only on an individual basis but as well for its placement within the state space of a simulation. Efforts such as those described in [21,22] have combined big data, machine learning and agent based simulations in an effort to predict online social activity. The challenges of the arising complexity of societal systems is discussed in [23,24] which describes how aggregate behaviors inherent in simple earthquake and avalanche models can be found in many complex systems, such as financial systems, which display the power law distributions for their oscillations.…”

Exploring Usability of Reddit in Data Science and Knowledge Processing

“…Hence, this category was checked specifically. Even though many works focus on information spreading in online communities [13,41,11,12], there is hardly any focus purely on information/knowledge retrieval. There are precisely two papers (1% of the considered work) related to knowledge processing (specifically, knowledge graphs [6,43]).…”