Fecha: jueves 8 de mayo
Hora: 16:15
Speaker: Mariano Vazquez (Universidad de Chile)
Title: Cyber risk prevention under risk averse spectral criteria
Abstract: This talk presents a model for the prevention and mitigation of cyber risks in networked systems, using a risk-averse approach based on spectral criteria.
The network is modeled as an initial graph, where attacks infect edges and trigger a Susceptible-Infected-Susceptible (SIS) process with controlled infection and recovery rates. Over time, edges are removed when infected and reinserted when recovered, generating a continuous-time Markov chain over the evolving subgraphs.
We study an insured agent who minimizes network risk through self-protection and self-insurance efforts, affecting infection probabilities and loss sizes. Risk is measured using a coherent, law-invariant measure $\rho$, applied to a spectral function $\varphi$ based on the Laplacian matrix. This function captures network resilience and increases with edge additions.
The optimization problem adjusts protection, insurance efforts, and coverage levels $\alpha$, balancing costs and premiums. Theoretical results show conditions for monotonicity in first-order stochastic dominance (FSD), and key properties of the risk measures used.