Guanyu Wang

About Me

Software Engineer @ Google

       I'm a Software Engineer at Google now. At the end of 2013, I graduated from School of Computer Science, Carnegie Mellon University with a M.S. degree specialized in Very Large Information Systems. My work involves with applying machine learning approaches to knowledge mining and natural language understanding.

Education

Carnegie Mellon University
Master of Science in Very Large Information System (August 2012 - December 2013)

GPA: 3.92.
Mainly focus on machine learning on big data, distributed system.

Tsinghua University
Master of Science in Computer Science (August 2010 - July 2012)

GPA: 92.
Mainly focus on algorithm design and analysis for distributed system.

University of Science and Technology Beijing
Bachelor of Engineering in Computer Science and Technology (August 2006 - July 2010)

GPA: 91; Major GPA: 93.
Medal of President award, USTB, 2010.
National Scholarship, three times in a row, 2007-2009.


Project
(Selected)

Slot Filling with Scalable Machine Learning
May 2013 - Dec 2013

  1. Applied the distant(weak) supervision learining methods to train the relation classifiers on terabyte size data set, in order to answer the questions like ”who did create Google” or ”what is Obama’s job title”, etc.
  2. Built a efficient system which can retrieve a collection of entities from the target large text corpus to fill in values for predefined slots (relation attributes) for a given entity in a reference Knowledge Base (Knowledge Base Population).
  3. Proposed an algorithm called Focus-Entity Slot Filling algorithm for Slot Filling task, which inputs a single ``focus entity'' ef and a set of relations R, then finds all other entities that the focus entity is related to via some relation in R from a large given unstructured natural language corpus.

"Who To Follow" Recommendation System
Dec 2012 - May 2013

  1. Explored the "who to follow" recommendation problem in Twitter like social networks.
  2. Incorporated the social network structure with user and item features to provide recommendation by using the personalized random walk.
  3. Implemented distributed feature extraction and recommendation system with MapReduce framework and GraphChi (process very large graph computations on a single machine).
  4. Tested system on KDDCUP12 Tencent Dataset, got rank 50th among about 700 teams.

Programming Skill

  • C/C++
  • Java
  • Python
  • C#
  • Javascript

Interests

  • Machine Learning
  • Distributed System
  • Algorithm
  • CG Painting
  • Playing Basketball
  • Watching Anime