Demo Data from EDBT 2026 / reviewers in the wild / expert
Qingliang Ou
dblp:416/7541
· DBLP profile ↗
1ranked-venue papers
0as first author
1since 2021 · last 2025
0009-0006-6762-8056ORCID · reported
Domains — the database's venue-derived domains; a paper can count in several
Software engineering, systems software and programming languages · 1 · 1 since 2021
Expertise — from the expertise taxonomy: the topics of the expert's papers under the CCF categories. A weight counts papers with recency: 1 for a paper about the topic, 0.3 when the topic is its context, halved every five years.
| Artificial intelligence
1 paper |
Efficient and distributed learning · 50% Deep learning architectures and training · 50% | |
| Computer architecture, parallel and distributed computing, and storage systems
1 paper |
Hardware accelerators and domain-specific architectures · 77% GPUs and heterogeneous computing · 23% |
Topics — the 4 heaviest of 4, each with the papers that count most for it
| Topic | Weight | Papers | Last | Evidence papers |
|---|---|---|---|---|
Machine learning › Deep learning architectures and training › mixture of experts
mixture-of-experts inference |
0.9 | 1 | 2025 | KTransformers: Unleashing the Full Potential of CPU/GPU Hybrid Inference for MoE Models · SOSP 2025 |
Machine learning › Efficient and distributed learning
model inference |
0.9 | 1 | 2025 | KTransformers: Unleashing the Full Potential of CPU/GPU Hybrid Inference for MoE Models · SOSP 2025 |
Hardware accelerators and domain-specific architectures
machine learning accelerator |
0.9 | 1 | 2025 | KTransformers: Unleashing the Full Potential of CPU/GPU Hybrid Inference for MoE Models · SOSP 2025 |
GPUs and heterogeneous computing
CPU-GPU heterogeneous computing |
0.3 | 1 | 2025 | KTransformers: Unleashing the Full Potential of CPU/GPU Hybrid Inference for MoE Models · SOSP 2025 |
Methods — techniques the papers use, named apart from their topics
CPU/GPU hybrid inference · 1.7
| Year | Publication | Venue | Position |
|---|---|---|---|
| 2025 | KTransformers: Unleashing the Full Potential of CPU/GPU Hybrid Inference for MoE ModelsabstractDue to the sparse nature of Mixture-of-Experts (MoE) models, they are particularly suitable for hybrid CPU/GPU inference, especially in low-concurrency scenarios. This hybrid approach leverages both the large, cost-effective memory capacity of CPU/DRAM and the high bandwidth of GPU/VRAM. However, existing hybrid solutions remain bottlenecked by CPU computation limits and CPU-GPU synchronization overheads, severely restricting their ability to efficiently run state-of-the-art large MoE models, such as the 671B DeepSeek-V3/R1. Hongtao Chen, Weiyu Xie, Boxin Zhang, Jingqi Tang, Shaoyuan Chen, Ziwei Yuan, Chengyu Qiu, Yuening Zhu, Qingliang Ou, Jiaqi Liao, Xianglin Chen, Zhiyuan Ai, Yongwei Wu 0001 |
SOSP | 12 |