Share on Facebook Tweet on Twitter Share on LinkedIn Share by email
Weizhu Chen

Weizhu Chen

HI, I am in Cosmos Scope Team in Bellevue working on distributed machine learning on an internal map-reduce infrastructure. Before this, I worked in BING Data Mining team and Microsoft Research Asia since 2005 as a Researcher in the Machine Learning group. My research interests include large scale machine learning, data mining, information retrieval and I enjoy employing them to solve industrial challenges, mostly in Web search and online advertising. I finished my PhD in HKUST and my PhD supervisor is Prof.Qiang Yang with a thesis on click model. Prior to joining Microsoft, I worked in IBM research lab for about one year.  

My current research focus includes parallel machine learning and click model. My work in click model has been utilized by BING with achieving significant NDCG gains and acting as a sub-ranker since 2010 May. Meanwhile, my curiosity in data-driven research leads me to study parallel machine learning algorithms. We have published a toolkit, see SIGMA and SIGMA_AZURE. I also help the product team to parallelize the search ranking algorithm in BING and the CTR prediction algorithm in Ads Center R&R.

News:  Our team, including Nathan Liu, Botao Hu and I, win the champion of the Relevance Prediction Challenge organized by Yandax in WSDM 2012.  A report is shared in here.

Recent publications:

I have been fortunate to work closely with many talent interns. Most of them were undergraduate students in China.  They worked with me very closely and then went to universities for their PhD. The list for recent 2 years includes:

2011: Danqi Chen (Stanford), Yitao Gao (Facebook), Botao Hu (Stanford), Yuchen Zhang (UC Berkeley), Dong Wang(UCLA), Si Shen (HKUST)

2010: Zeyuan Allen Zhu (MIT), Chenguang Zhu (Stanford), , Zhanglong Ji (UCSD), Qifeng Chen (Stanford), Dakan Wang (Stanford)



PhD Thesis: An Associative Characterization of Click Models in Web Search Link