Erik van der Werf

About me
• I'm a Senior Research Scientist and System Architect in the research department of GN ReSound in Eindhoven, where I've been working since 2005. GN is an international market leader in digital hearing instruments, headsets, and audiological diagnostics equipment with headquarters in Copenhagen and R&D sites in the US, the Netherlands, Denmark and China.
  I have a Ph.D. degree in Artificial Intelligence from Maastricht University , and a M.Sc. degree in Applied Physics from the Delft University of Technology. For my M.Sc. research I worked in the Pattern Recognition Group on non-linear feature extraction. For my Ph.D. research I worked on computer Go (Baduk,Weiqi). The research resulted in the thesis 'AI techniques for the game of Go', several publications, and the two Go programs Magog and Migos". The 9x9 Go program Magog participated in several international events, and has won the bronze medal in the 9^th Computer Olympiad. The program Migos was the first program in the world to solve 5x5 Go. In 2008 Migos also found the world's first solutions for 3x7, 4x6, 4x7 and 5x6 (all published human solutions for these boards were wrong). Migos still holds the record for the largest Go boards solved by computer.
  In 2007 I wrote a new Go program called Steenvreter which won the 12^th Computer Olympiad and still performs well in more recent events.
  
  My main interests and expertise are in artificial intelligence, pattern recognition, machine learning, digital signal processing, neural networks, search techniques, games, data mining, perception, and of course computer Go.

Publications (click here)

Go Programs
• Steenvreter.
• Magog.
• Migos.

Go Research
• Migos has solved most rectangular Go boards up to 30 intersections. Here's an overview of first player scores.
• My thesis AI techniques for the game of Go is available online. Contact me if you're interested in a hardcopy.
• A brief overview of my Ph.D. research is available in this presentation (presented at MSRC).
• Magog won the bronze medal in the 9^th Computer Olympiad
• My program Migos solved 5x5 Go for all possible opening moves.
• I supervised Reindert-Jan Ekker for his M.Sc thesis on "Reinforcement Learning and Games"

The Game of Go
• Introduction to the game of Go by the BGA
  
• Play Go on The Kiseido Go Server
  
• Go software available for download from BGA software catalogue
  
• Research in Go
  Publications available on-line
  
• Other Go-links
  Sensei's Library
  Go.pagina.nl
  GoBase.org by Jan van der Steen
  Dutch Go Association
  British Go Association
  European Go Federation
  American Go Association
  Eindhovense Go Club
  

Academic links

• The Games and AI Group of Maastricht University
  
• The University of Alberta GAMES Group
  
• The MSRC Applied Games group
  
• Hiroyuki Iida's research Unit for Computers and Games
  
• Bruno Bouzy, Tristan Cazenave, Keh-Hsun Chen, Rémi Coulom, Bernard Helmstetter, Levente Kocsis, Lukasz Lew, Martin Müller, Jan Ramon, Nic Schraudolph, Mark Winands