Quantitative Methods for Dynamics on Networks

The past decade witnessed a growing interest in building analytical tools for the study of out-of-equilibrium cascading dynamics in networked disordered systems, whereby information, viruses, and failures propagate through their edges via the interactions between individual constituents. These dynamic processes are ubiquitous, and include: epidemic spreading; propagation of information and innovations in social media; avalanche dynamics of magnetic and glassy systems; point processes; and communication protocols, such as gossip algorithms and peer-to-peer file sharing on computer networks.

This workshop seeks perspectives on quantitative methods for solving the cascading dynamics on heterogeneous networks, as well as applications in inference, optimization, control, and learning tasks for spreading processes. Popular problems, such as learning of the network structure and parameters; inference of the epidemic origin; targeted resource allocation for maximizing information dissemination or mitigating the epidemic spread, emerged among important applications of interest to LANL and DOE.

We invite experts in methods that include but are not limited to: dynamic message-passing algorithms; dynamic cavity method; edge-based compartmental modeling; heterogeneous mean-field methods; probability generating functional techniques; statistical learning; and dynamical replica analysis, to discuss the recent advances in approximate solutions of inference problems on networks.

Organizers:

Andrey Lokhov (Los Alamos National Laboratory)

Mateusz Wilinski (Tampere University)