Guolin Ke
Creator of LightGBM; SVP of AI at DP Technology and former senior researcher at Microsoft Research Asia
- SVP of AI, DP Technology1
Guolin Ke is SVP of AI at DP Technology, working on AI for Science. Ke created LightGBM, an open-source gradient boosting decision tree framework, in 2016 during an internship at Microsoft Research Asia, and is first author of the NIPS 2017 paper that introduced its Gradient-based One-Side Sampling and Exclusive Feature Bundling techniques. Ke later became a Senior Researcher in the Machine Learning Group at Microsoft Research Asia, and in 2021 led the development of Graphormer. Ke also led the development of Uni-Fold, a reimplementation of AlphaFold with open-source training code, and Ke's homepage lists team releases including the Uni-Mol 3D molecular pretraining framework and Uni-3DAR, an autoregressive model for a range of 3D tasks.123
Work
LightGBM (external site: github.com)Open-source code
Created LightGBM in 2016 during an internship at Microsoft Research Asia, and first author of "LightGBM: A Highly Efficient Gradient Boosting Decision Tree" (NIPS 2017); Ke's GitHub profile describes Ke as the creator of LightGBM.1324
Graphormer
Led the development of Graphormer in 2021, according to Ke's homepage.1
Uni-Mol (external site: github.com)Open-source code
Ke's homepage lists the 2022 release of Uni-Mol, a 3D molecular and pocket pretraining framework, and Ke as an author of the Uni-Mol paper (ICLR 2023); the MIT-licensed repository is pinned on Ke's GitHub profile.125
Earlier roles
- Senior Researcher, Machine Learning Group, Microsoft Research Asia1
This profile covers public professional work only, from the sources below. It is not an endorsement or a ranking. To correct or remove it, use the correction form.