Controlled Sensing in Social Networks

Tuesday, May 15, 2018 - 11:00am - 11:50am
Lind 305
Vikram Krishnamurthy (Cornell University)
This talk describes stochastic control problems involving social sensing models. Humans can be viewed as social sensors that input information to a social network. The interaction of social sensors present unique challenges: sensors interact with and influence other social sensors resulting in herding behaviour. Social sensors are risk averse decision makers. This talk describes how herding can be mitigated by providing incentives to individual sensors. Also experimental data on human interaction will be discussed to motivate the models and algorithms presented. The talk draws from ideas in behavioral economics, structural results in partially observed Markov decision processes and social networks.