Seminar - "The spread of fake news: a mathematical model"

Wednesday, November 15, 2017 3:30 pm - 4:30 pm EST (GMT -05:00)

SPEAKER: Rahul Roy, Indian Statistical Institute, New Delhi, India

TOPIC:  The spread of fake news: a mathematical model           

LOCATION: E5 4106-4128

INVITED BY: Ravi Mazumdar

ABSTRACT:

Consider a model where individuals are placed at each point of a non-negative integer lattice. Individuals are of two kinds - gullible and astute. With probability p an individual is gullible and with probability 1 - p he/she is astute, independently of other individuals.

A rumour is started at the origin and is shared with all individuals in the random interval [0;R0]. If a gullible person at location i hears the rumour, he/she spreads it to the individuals in the interval [i; i + Ri]. Here R0;R1; : : : are i.i.d. random variables.

The rumour dies if only finitely many people hear the rumour and, it survives if with a positive probability everybody hears the rumour. We obtain conditions on p and R0 for the survival of the rumour. We also generalise this model to gullible/astute individuals located according to a Markov process on the line and also to individuals located in d dimensions for d > 2. We also study the case where instead of gullible individuals we have sceptical individuals who spread the rumour only if he/she receives the rumour from two different sources.

BIOGRAPHY:

Rahul Roy is with the Indian Statistical Institute, New Delhi, India. He obtained his PhD at Cornell University, USA.  After serving on the faculty in the Department of Mathematics, Duke University, USA he returned to India in 1988 to join the Indian Statistical Institute where he is a Professor.

His research is in Percolation Theory and applied probability models. He serves as an Associate Editor of the Journal of Applied Mathematics and Communications on Stochastic Analysis. He is the author of 2 monographs in the areas of percolation theory and applications.

                                                  All are welcome!