From Contagion to Stability: Insights into Network Dynamics, Resilience and Stability

The resilience of interconnected systems, such as financial networks, supply chains, software systems, and social networks, is a critical concern in today's highly connected world. While interconnectedness enables efficient economic transactions, rapid social learning, and adaptability to shocks, it also increases vulnerability to systemic risks, where localized disruptions can propagate and cause widespread failures. This thesis addresses this fundamental challenge by examining how to reason about and reinforce the resilience of complex networks through theoretical and applied tools, including probability, statistics, algorithms, and network science, whereas we rely on centralized and decentralized decision-making to design interventions that mitigate cascading failures and bolster network stability. First, the thesis focuses on optimizing resource allocation in networks undergoing contagion, developing novel resilience metrics for supply chains, and creating efficient algorithms to prevent cascading failures. Secondly, this thesis studies models of contagion and gives a formalized definition of resilience. Then, this thesis explores decentralized privacy-aware decision-making to reconcile privacy with efficient social learning in risk-prone environments to ensure resilience. Finally, the thesis studies models of network formation and also suggest modern ways to view complex interconnected systems through the lens of LLMs.

Name thesis:degree_name

Ph. D., Computer Science

Level thesis:degree_level

Doctor of Philosophy

Discipline thesis:degree_discipline

Computer Science

Grantor

Cornell University

Year dc:date.issued

2025

Author dc:creator

• Papachristou, Marios

Committee members dc:contributor.committeemember

• Pierson, Emma
• Banerjee, Siddhartha

• contagion
• interventions
• LLMs
• networks
• privacy
• resilience

Language dc:language.iso

en

DOI

https://doi.org/10.7298/rm2m-az62

Dc Identifier Other

ProQuest Submission ID: 15071
ProQuest Publication ID: 32120004

OAI identifier oai:identifier

oai:ecommons.cornell.edu:1813/120828