陆昱成现任上海纽约大学计算机科学助理教授,并当人 HeavyBall 研究团队负责人。他的研究聚焦于高效机器学习系统的软硬件协同设计,涵盖训练与推理引擎、GPU 内核优化以及大规模人工智能应用。在加入上海纽约大学前,他曾任Together AI研究工程师,并曾在谷歌、微软与亚马逊等科技企业完成多项研究实习。陆教授于2023年在康奈尔大学获得计算机科学博士学位, 师从Chris De Sa教授。
代表性论著
- Jue Wang*, Yucheng Lu*, Binhang Yuan, Beidi Chen, Percy Liang, Chris De Sa, Chris Ré, Ce Zhang. CocktailSGD: Fine-tuning Foundation Models over 500Mbps Networks. In the Fortieth International Conference on Machine Learning (ICML) 2023.
- Yucheng Lu, Conglong Li, Minjia Zhang, Chris De Sa, Yuxiong He. Maximizing Communication Efficiency for Large-scale Training via 0/1 Adam. In the Eleventh International Conference on Learning Representations (ICLR) 2023.
- Yucheng Lu, Wentao Guo, Chris De Sa. GraB: Finding Provably Better Data Permutations than Random Reshuffling. In the Thirty-sixth Conference on Neural Information Processing Systems (NeurIPS) 2022.
- Yucheng Lu, Chris De Sa. Optimal Complexity in Decentralized Training. In the Thirty-eighth International Conference on Machine Learning (ICML) 2021.
- Yucheng Lu, Chris De Sa. Moniqua: Modulo Quantized Communication in Decentralized SGD. In the Thirty-seventh International Conference on Machine Learning (ICML) 2020.
教育背景
- 康奈尔大学 计算机科学博士
- 上海交通大学 电子工程学士
研究兴趣
机器学习系统