Unsupervised Conformal Novelty Detection for Hierarchical Data
A conformal e-value framework for unsupervised novelty detection in hierarchically structured data, achieving FDR control at both group and unit levels.
Paper
Multi-Agent Large Language Model-Based Economic Dispatch Framework for Optimal Energy Management of Data Center Microgrids
A multi-agent LLM framework that turns natural-language prompts into optimal economic dispatch for data center microgrids, validated on a 400 MW MIT TX-GAIA load profile.