À propos
I (Mao Yang, 杨懋) received my Ph.D degree in computer science from Beijing University, China, in 2006. Before that, I received my B.S., M.S. in computer science from Harbin Institute of Technology, Harbin, in 2000 and 2002, respectively. Since 2006, I have been with Microsoft Research Asia, Beijing, as researcher manager for Systems and Networking Research Group (Asia) (opens in new tab)
My research interests are in distributed systems, information retrieval systems, machine learning systems, and multimedia systems, especially for design, implement and deploy practical systems.
I am also an architect, and I worked on the following projects at Microsoft BING (opens in new tab) team:
- The design and implementation of Cougar, a new ranking system for supporting the-state-of-art semantic ranking models. The system starts to serve all web queries since from 2013.
- The design and implementation of Tiger, a new generation flash memory based index serving platform, and the system starts to serve all web queries since from 2012.
- The design and implementation of replication and fail over protocol of Kirin, a new web store and processing system, and the system starts to process many billions of web data since from 2010.
- Proposed a Web scale Q&A system that build into Web search engine. The system starts to provide directly answers in Bing since from 2016.
Some other research projects I’ve worked on include:
- TLA Made Live: a formal method to build distributed systems.
- The design and implementation of a large scale distributed storage system prototype PacificA. The protocol is also used by several open source projects, such as rDSN (opens in new tab), Kafka (opens in new tab).
- Reconfiguration protocol for a paxos based replication state machine library.
My current research focus is on the AI infrastructure and Tools, and algorithms for Web search. We released several projects:
- OpenPAI (opens in new tab) : An open source platform that provides complete AI model training and resource management capabilities, it is easy to extend and supports on-premise, cloud and hybrid environments in various scale.
- NNI (neural network intelligence) (opens in new tab) : An open source AutoML toolkit for neural architecture search and hyper-parameter tuning.
- MMdnn (opens in new tab): A comprehensive, cross-framework solution to convert, visualize and diagnose deep neural network models.
- Visual Studio Tools for AI (opens in new tab) / Visual Studio Code Tools for AI (opens in new tab) and GitHub repo (opens in new tab) : An extension for Visual Studio and Visual Studio Code to build, test, and deploy Deep Learning / AI solutions.