Zirou Qiu

Summary. We developed a principled model for resource allocation with externalities (e.g., mutual influences among the resources or the population) in complex systems. We proposed efficient strategies that guarantees optimal or near-optimal welfare outcomes for real-world re- source allocation scenarios

Keywords. Resource allocation, optimization, large systems planning, computational social science

Summary. We proposed efficient noise-tolerant algorithms for learning the behaviors of large-scale systems in realistic noisy environments, with a provable high learning accuracy. Our algorithms are highly scalable to large networks, and the amount of training data required is provable near-optimal.

Keywords. Machine learning on graphs; learing under noise; complex systems; sample complexity

We estabilished a theoretical groundwork for dimensionality reduction in Bayesian inference. Further, we proposed efficient algorithms for dimensionality reduction with provable guarantees.

Keywords. Bayesian Networks; Parent divorcing transformation; Network embedding; Theoretical foundations

We proposed the first efficient learning method for inferring the behaviors of large-scale multilayer social systems and multi-agent systems. Further, we lay a theoretical foundation of learning multilayer systems with an exact analysis of the model complexity and expressiveness

Keywords. Machine learning on graphs; multilayer graphs; complex systems; sample complexity

We investigate the core principles of learning both the behaviors and the topology of black-box social systems and multi-agent systems. Towards this end, we proposed several efficient learning algorithms with provable guarantees for realistic scenarios of learning large systems.

Keywords. Machine learning on graphs; complex systems; sample complexity;

We developed a formal model for real-world spatial resources allocation problems, such as public housing assignment and urban spatial planning. Further, we presented rigorous methodologies for resource allocation with provable close-to-optimal guarantees, while achieving social diversity and fairness.

Keywords. Resource allocation and planning; optimization; fairness and diversity

We presented a theoretical groundwork for stability and convergence in real-word strategic situations such as resources sharing and competition within a society.

Keywords. Game theory; Nash equilibrium, convergence time

We introduced a principled modeling framework that couples the fast dynamics of viral loads in the local scale and the slow dynamics of disease progression at the population level. Further, we derived policy guidelines to lessen the negative impact of epidemics.

Keywords. Group gathering; multiple time scales; large-scale modeling; pandemic science;

We presented a fundamental model for dueling dynamics of social and biological contagions both cooperating and completing in a population. Through our modeling, we uncovered a collection of previously unknown phenomena in contagion dynamics. We derived corresponding policy guidelines to suppress epidemics and prevent their resurgence.

Keywords. Dueling dynamics, multilayer networks, non-pharmaceutical interventions, social science

We established the principles of finding equilibrium points in contagion diffusion within a population, with the goal of minimizing the overall scale of infection in a society.

Keywords. Optimization; complex systems; fixed point; contagion diffusion

We presented the first collection of active learning based methods for inferring both the behaviors and the topology of black-box social systems and multi-agent systems. Further, we conducted an extensive experimental study on the empirical performs on the proposed algorithms on real-world networks over various learning scenarios.

Keywords. Machine learning on graphs; Active learning; query model; complex systems; query complexity

Summary. We propose novel algorithms for the network alignment problem that outperforms the state-of-the-art methods, with applications in computational biology and social network analysis.

Keywords. Optimization; network alignment; social network analysis