PhD Thesis Defense: Yu-Zhen (Janice) Chen, Toward Resource-Efficient Decision-Making for Networked Systems
Content
Speaker:
Abstract:
Networked systems are prevalent in critical domains such as environmental monitoring, healthcare, recommendation systems, and wireless communications, where resource constraints -- including limitations on energy, samples, and communication -- pose significant challenges to effective decision-making. This dissertation develops resource-efficient decision-making algorithms to optimize the performance of networked systems under such constraints.
The first part of this dissertation focuses on multi-agent systems and develops communication-efficient policies for cooperative bandit learning, with applications including multi-server recommendations and distributed clinical trials. The second part addresses sequential distributed parameter estimation in wireless sensor networks, where sensing, data collection, and transmission must be balanced under energy constraints to maximize estimation accuracy. The third part studies sequential change detection, with the goal of achieving sample-efficient detection in complex environments characterized by uncertainty in change distributions and limited observability of internal network states.
Overall, this dissertation advances resource-efficient estimation, learning, and decision-making by providing theoretical foundations and developing practical algorithms for scalable and reliable networked systems.
Advisor:
Don Towsley