Dynamics of Interactive Social Sensors
Abstract
This project deals with the statistical signal processing and controlled sensing of interacting social sensors. Social sensors are agents that provide information about their environment (state of nature) to a social network after interaction with other agents. This research aims to develop a mathematical theory and practical methodologies for interactive sensing and decision making amongst social sensors. The objectives fall under three inter-related themes: (i) Bayesian social sensing in finite networks where herding, data incest, learning rates and controlled sensing are considered; (ii) interactive sensing in large scale networks (such as power law networks) where adaptive estimation of degree distribution dynamics, infection dynamics and achievable bounds are considered; and (iii) analyzing YouTube datasets to justify the models, algorithms and analysis developed. The research involves the interplay of Bayesian social learning, mean field dynamics, and partially observed Markov decision process (stochastic control). The problems considered transcend classical statistical signal processing (which deals with extracting signals from noisy measurements) to address the deeper issue of how multi agent decision systems and signal processing algorithms interact over a network
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Oct 11, 2018
- Source ID
- W911NF1710335
Entities
People
- Vikram Krishnamurthy
Organizations
- Army Contracting Command
- Cornell University
- Office of the Secretary of Defense