tex/gostyle.bib

   1 % This file was created with JabRef 2.6b2.
   2 % Encoding: UTF-8
   3
   4 @MISC{GoR,
   5   author = {Ales Cieply and others},
   6   title = {EGF ratings system --- Sys\-tem des\-crip\-tion},
   7   owner = {pasky},
   8   timestamp = {2011.03.10},
   9   url = {http://www.europeangodatabase.eu/EGD /EGF\_rating\_system.php}
  10 }
  11
  12 @ARTICLE{CoverHart1967,
  13   author = {Thomas M. Cover and Peter E. Hart},
  14   title = {Nearest neighbor pattern classification},
  15   journal = {IEEE Transactions on Information Theory},
  16   year = {1967},
  17   volume = {13},
  18   pages = {21-27},
  19   number = {1},
  20   owner = {hellboy},
  21   timestamp = {2010.03.10}
  22 }
  23
  24 @INPROCEEDINGS{GellySilver2008,
  25   author = {Gelly, Sylvain and Silver, David},
  26   title = {Achieving master level play in 9x9 computer go},
  27   booktitle = {AAAI'08: Proceedings of the 23rd national conference on Artificial
  28         intelligence},
  29   year = {2008},
  30   pages = {1537--1540},
  31   publisher = {AAAI Press},
  32   abstract = {The UCT algorithm uses Monte-Carlo simulation to estimate the value
  33         of states in a search tree from the current state. However, the first
  34         time a state is encountered, UCT has no knowledge, and is unable
  35         to generalise from previous experience. We describe two extensions
  36         that address these weaknesses. Our first algorithm, heuristic UCT,
  37         incorporates prior knowledge in the form of a value function. The
  38         value function can be learned offline, using a linear combination
  39         of a million binary features, with weights trained by temporal-difference
  40         learning. Our second algorithm, UCT-RAVE, forms a rapid online generalisation
  41         based on the value of moves. We applied our algorithms to the domain
  42         of 9 × 9 Computer Go, using the program MoGo. Using both heuristic
  43         UCT and RAVE, MoGo became the first program to achieve human master
  44         level in competitive play.},
  45   isbn = {978-1-57735-368-3},
  46   location = {Chicago, Illinois}
  47 }
  48
  49 @BOOK{Jolliffe1986,
  50   title = {Principal Component Analysis},
  51   publisher = {Springer, New York},
  52   year = {1986},
  53   author = {I.T. Jolliffe},
  54   owner = {hellboy}
  55 }
  56
  57 @ARTICLE{Pearson,
  58   author = {R. L. Plackett},
  59   title = {Karl Pearson and the Chi-Squared Test},
  60   journal = {International Statistical Review},
  61   year = {1983},
  62   volume = {51},
  63   pages = {59--72},
  64   number = {1},
  65   issn = {03067734},
  66   owner = {pasky}
  67 }
  68
  69 @INPROCEEDINGS{Riedmiller1993,
  70   author = {Martin Riedmiller and Heinrich Braun},
  71   title = {A Direct Adaptive Method for Faster Backpropagation Learning: The
  72         RPROP Algorithm},
  73   booktitle = {IEEE INTERNATIONAL CONFERENCE ON NEURAL NETWORKS},
  74   year = {1993},
  75   pages = {586--591},
  76   owner = {hellboy},
  77   timestamp = {2010.03.07}
  78 }
  79
  80 @ARTICLE{MDP,
  81   author = {Zito Tiziano and Wilbert Niko and Wiskott Laurenz and Berkes Pietro},
  82   title = {Modular toolkit for Data Processing (MDP): a Python data processing
  83         framework},
  84   journal = {Frontiers in Neuroinformatics},
  85   year = {2008},
  86   volume = {2},
  87   abstract = {Modular toolkit for Data Processing (MDP) is a data processing framework
  88         written in Python. From the user&#39;s perspective, MDP is a collection
  89         of supervised and unsupervised learning algorithms and other data
  90         processing units that can be combined into data processing sequences
  91         and more complex feed-forward network architectures. Computations
  92         are performed efficiently in terms of speed and memory requirements.
  93         From the scientific developer&#39;s perspective, MDP is a modular
  94         framework, which can easily be expanded. The implementation of new
  95         algorithms is easy and intuitive. The new implemented units are then
  96         automatically integrated with the rest of the library. MDP has been
  97         written in the context of theoretical research in neuroscience, but
  98         it has been designed to be helpful in any context where trainable
  99         data processing algorithms are used. Its simplicity on the user&#39;s
 100         side, the variety of readily available algorithms, and the reusability
 101         of the implemented units make it also a useful educational tool.},
 102   doi = {10.3389/neuro.11/008.2008},
 103   issn = {ISSN 1662-5196},
 104   owner = {hellboy},
 105   timestamp = {2010.03.05}
 106 }
 107
 108 @comment{jabref-meta: selector_publisher:}
 109
 110 @comment{jabref-meta: selector_author:}
 111
 112 @comment{jabref-meta: selector_journal:}
 113
 114 @comment{jabref-meta: selector_keywords:}
 115