Profiles2021-2025
Xiaonan Nie
Lab of High Confidence Software Technologies (MOE)
9Papers
0Last author
17Unique co-authors
3First author
0Solo
First and last author come from byline position, which some venues assign alphabetically.
Recent papers9
- MEMO: Fine-grained Tensor Management For Ultra-long Context LLM TrainingAuthor 12 of 15SIGMOD 2025 · 4 citations
- NetMoE: Accelerating MoE Training through Dynamic Sample PlacementAuthor 6 of 7ICLR 2025
- LSH-MoE: Communication-efficient MoE Training via Locality-Sensitive HashingAuthor 1 of 9NeurIPS 2024 · 10 citations
- Galvatron: Efficient Transformer Training over Multiple GPUs Using Automatic ParallelismAuthor 5 of 7VLDB 2023 · 113 citations
- FlexMoE: Scaling Large-scale Sparse Pre-trained Model Training via Dynamic Device PlacementAuthor 1 of 6SIGMOD 2023 · 40 citations
- HET: Scaling out Huge Embedding Model Training via Cache-enabled Distributed FrameworkAuthor 4 of 7VLDB 2022 · 70 citations
- TSPLIT: Fine-grained GPU Memory Management for Efficient DNN Training via Tensor SplittingAuthor 1 of 4ICDE 2022 · 26 citations
- HET-GMP: A Graph-based System Approach to Scaling Large Embedding Model TrainingAuthor 5 of 7SIGMOD 2022 · 24 citations
- Heterogeneity-Aware Distributed Machine Learning Training via Partial ReduceAuthor 2 of 6SIGMOD 2021 · 64 citations
