M
Mu Li
Researcher at Amazon.com
Publications - 75
Citations - 12622
Mu Li is an academic researcher from Amazon.com. The author has contributed to research in topics: Deep learning & Computer science. The author has an hindex of 27, co-authored 72 publications receiving 9489 citations. Previous affiliations of Mu Li include Google & Baidu.
Papers
More filters
Posted Content
Empirical Evaluation of Rectified Activations in Convolutional Network.
TL;DR: The experiments suggest that incorporating a non-zero slope for negative part in rectified activation units could consistently improve the results, and are negative on the common belief that sparsity is the key of good performance in ReLU.
Posted Content
MXNet: A Flexible and Efficient Machine Learning Library for Heterogeneous Distributed Systems
Tianqi Chen,Mu Li,Yutian Li,Min Lin,Naiyan Wang,Minjie Wang,Tianjun Xiao,Bing Xu,Chiyuan Zhang,Zheng Zhang +9 more
TL;DR: The API design and the system implementation of MXNet are described, and it is explained how embedding of both symbolic expression and tensor operation is handled in a unified fashion.
Proceedings ArticleDOI
Scaling Distributed Machine Learning with the Parameter Server
TL;DR: View on new challenges identified are shared, and some of the application scenarios such as micro-blog data analysis and data processing in building next generation search engines are covered.
Proceedings ArticleDOI
Scaling distributed machine learning with the parameter server
Mu Li,David G. Andersen,Jun Woo Park,Alexander J. Smola,Amr Ahmed,Vanja Josifovski,James Long,Eugene J. Shekita,Bor-Yiing Su +8 more
TL;DR: In this paper, the authors propose a parameter server framework for distributed machine learning problems, where both data and workloads are distributed over worker nodes, while the server nodes maintain globally shared parameters, represented as dense or sparse vectors and matrices.
Proceedings ArticleDOI
Bag of Tricks for Image Classification with Convolutional Neural Networks
TL;DR: This article examined a collection of such refinements and empirically evaluated their impact on the final model accuracy through ablation study, and showed that by combining these refinements together, they are able to improve various CNN models significantly.