Emory University, a Top-Ranked US Research University in Georgia, Leverages EdgeCloud for AI…

Emory University, a Top-Ranked US Research University in Georgia, Leverages EdgeCloud for AI Research in Disease Diagnosis, Transmission and Epidemic Modelling

We’re pleased to announce that Emory University, a top-tanked research university in Georgia, through its Melody Lab led by Assistant Professor Wei Jin, has joined Theta’s academic partner network by adopting Theta EdgeCloud Hybrid.

The Melody Lab — short for Machine Learning for Data Analytics — is dedicated to advancing machine learning methods capable of handling diverse data structures. Their research interests cover graph neural networks and graph machine learning, time-series analysis (including forecasting and classification), AI for epidemiology, science, and healthcare, as well as machine learning robustness, interpretability, and large language models (LLMs). By combining methodological innovation with real-world applications, the Melody Lab has gained international recognition for its contributions to graph learning, time-series forecasting, and interpretable AI systems.

Prof. Jin is widely recognized for his research in graph neural networks, time-series forecasting, and AI for healthcare, and has received honors including the INNS Doctoral Dissertation Runner-up Award, KAUST Rising Stars in AI, AAAI New Faculty Highlights, and the Snap Research Fellowship. His work has been published extensively at top AI venues such as ICML, NeurIPS, KDD, ICLR, and NAACL, establishing him as one of the leading young scholars advancing robust and interpretable machine learning.

Recent Publications from the Melody Lab

The Melody Lab’s pioneering work has been published at top-tier international conferences in 2025 and 2024, including:

• MICCAI 2025: Achieving fairness in skin disease diagnosis through group-specific experts (early accept, top 9%).
• ICML 2025: Epidemiology-Aware Neural ODE with Continuous Disease Transmission Graph.
• KDD 2025: Graph ODEs and Beyond: a comprehensive survey integrating differential equations with graph neural networks.
• NAACL 2025 (Proceedings): STRUX: An LLM for Decision-Making with Structured Explanations.
• ICLR 2025: Precedence-Constrained Winter Value for effective graph data valuation.
• DATE 2024: Enabling on-device learning via experience replay with efficient dataset condensation.
• NeurIPS 2024: Text-space Graph Foundation Models: comprehensive benchmarks and new insights.
• SIGKDD 2024: A Review of Graph Neural Networks in Epidemic Modeling.

These publications highlight the lab’s leadership in graph machine learning, large language models, biomedical AI, and time-series forecasting, reinforcing why flexible, decentralized compute infrastructure like Theta EdgeCloud is critical for accelerating breakthroughs.

“Theta EdgeCloud Hybrid provides our lab with the ability to conduct large and small-scale experiments across heterogeneous GPU environments, ranging from NVIDIA H100s to community-driven GPUs and mobile Android GPUs,” said Prof. Wei Jin. “This unique flexibility is not found in any other cloud provider or GPU platform. This allows us to accelerate research on graph neural networks, time series forecasting, and AI for healthcare specifically in disease diagnosis, transmission and epidemic modeling. EdgeCloud enables us to reduce infrastructure bottlenecks and iterate faster.”

Accelerating Healthcare Research at the Melody lab with Theta EdgeCloud Hybrid

EdgeCloud Hybrid is the world’s first decentralized hybrid GPU platform, combining thousands of community-run NVIDIA GPUs (30/40/50 series), high-end enterprise A100s and H100s, and AWS’s next-generation Inferentia and Trainium AI chips.

The Melody Lab chose EdgeCloud because it delivers a unique balance of scale, flexibility, and affordability, allowing the team to seamlessly move between intensive large-scale experiments and lighter prototyping. This capability not only accelerates iteration but also removes infrastructure bottlenecks. For a lab focused on graph learning, time-series forecasting, and AI for healthcare, access to diverse, cost-effective compute power means they can train large language models, advance graph neural networks, and deploy real-world AI applications without being constrained by conventional cloud limitations.

“We are excited to welcome Dr. Wei Jin and the Melody Lab at Emory University to Theta’s academic partner network with nearly 30 top tier global AI labs,” said Mitch Liu, CEO and Co-Founder of Theta Labs. “Emory’s groundbreaking research in AI for healthcare, disease and epidemic modeling perfectly demonstrates the kind of innovation EdgeCloud was built to support. We’re the only hybrid platform that enables Emory to do large and small scale AI experimentation at scale and cost-effectively. This is only possible with our AWS Inferentia and Trainium chips on one end, and community run NVIDIA 30–40–50 series GPUs on the other end.”

About Emory University

Founded in 1836 and located in Atlanta, Georgia, Emory University is a top-ranked private research institution renowned for excellence across health sciences, liberal arts, business, and law. With more than 15,000 students and nine academic divisions, Emory is a global leader in interdisciplinary research and innovation, consistently ranked among the best U.S. universities for research funding and impact.

Emory University, a Top-Ranked US Research University in Georgia, Leverages EdgeCloud for AI… was originally published in Theta Network on Medium, where people are continuing the conversation by highlighting and responding to this story.