Towards Practical AI for Networked Systems

   Artificial intelligence has made significant advances in fields like vision and language, but its application in real-world networked systems remains limited due to challenges like unreliable data transmission, dynamic environments, and strict system constraints. This dissertation introduces a suite of learning-enabled systems designed to address these challenges, including a signal propagation model for LoRa-based orchard networks, a model predictive control framework for groundwater recharge, a causality-aware time series model, and a scalable system for real-time visual captioning. Together, these contributions demonstrate how co-designing AI with domain and systems knowledge enables robust and practical intelligence in complex networked environments.
   
   Yuning Chen is a Ph.D. candidate in the EECS department of UC Merced, under the guidance of Prof. Wan Du. He works on machine learning and its applications in networked systems, with experience in generative AI, time series causal learning and system optimization. Yuning’s work has been published in conferences such as MLSys and KDD, and he has served as a reviewer and area chair for leading machine learning venues.