6 papers
TUBO: A Tailored ML Framework for Reliable Network Traffic Forecasting
Zhihang Yuan, Leyang Xue, Waleed Ahsan +1
Traffic forecasting based network operation optimization and management offers enormous promise but also presents significant challenges from traffic forecasting perspective. While…
Towards Decentralized and Sustainable Foundation Model Training with the Edge
Leyang Xue, Meghana Madhyastha, Randal Burns +2
Foundation models are at the forefront of AI research, appealing for their ability to learn from vast datasets and cater to diverse tasks. Yet, their significant computational dema…
HybridServe: Efficient Serving of Large AI Models with Confidence-Based Cascade Routing
Leyang Xue, Yao Fu, Luo Mai +1
Giant Deep Neural Networks (DNNs), have become indispensable for accurate and robust support of large-scale cloud based AI services. However, serving giant DNNs is prohibitively ex…
MoE-Infinity: Efficient MoE Inference on Personal Machines with Sparsity-Aware Expert Cache
Leyang Xue, Yao Fu, Zhan Lu +2
This paper presents MoE-Infinity, an efficient MoE inference system designed for personal machines with limited GPU memory capacity. The key idea for MoE-Infinity is that on person…
PH-Dropout: Practical Epistemic Uncertainty Quantification for View Synthesis
Chuanhao Sun, Thanos Triantafyllou, Anthos Makris +4
View synthesis using Neural Radiance Fields (NeRF) and Gaussian Splatting (GS) has demonstrated impressive fidelity in rendering real-world scenarios. However, practical methods fo…
Learning High-Frequency Functions Made Easy with Sinusoidal Positional Encoding
Chuanhao Sun, Zhihang Yuan, Kai Xu +4
Fourier features based positional encoding (PE) is commonly used in machine learning tasks that involve learning high-frequency features from low-dimensional inputs, such as 3D vie…