Titre : Time Is Of The Essence: Spreading Information Among Interacting Groups

Conférencier : Imane Hafnaoui

Résumé :
When an intruder is sensed, hyenas are known to use unique whoops, specific to every individual, to identify the members of the clan and then use the whoops to coordinate the hunt against the intruder. This type of collective behavior is not only observed in the animal kingdom but also among humans and even down to the microscopic cells.

Generally, a collective behavior does not emerge from the state of the individual entities in a group, whether that be emotions of uncertainty, imagery or strain in the natural order. It is rather the result of the information shared between the individuals in a communication network. On these accounts, understanding the flow of information between individuals can help decipher their behavior.

In this presentation, we present a probabilistic model of the information propagation in networks in a way that is sufficiently simple yet effective for the engineering of distributed systems. We showcase its ability to explain certain events and observed behaviors. Considering that one of the incentives to studying collective behavior in nature is to gather the knowledge to engineer new systems, we take it a step further and provide upper bounds for the time for the information to propagate in a network, as well as identify individuals that are responsible for slow convergence.

Bio :
Imane Hafnaoui has been conducting her PhD since 2014 at Ecole Polytechnique de Montreal as a member of the MistLab. She obtained her master degree in 2013 from the Electrical and Electronics Engineering Institute in Algeria. Her research interests span the fields of swarm intelligence, bio-inspired optimization techniques and real-time embedded systems.

Bienvenue à tous!