Paper Available

[Toshiaki Aida]

Dear colleagues,

The following paper is now available at

   http://xxx.lanl.gov/abs/cond-mat/9911474 ,

in which we discussed the control of bin size for optimal on-line
learning of probability distributions.


TITLE:
Field Theoretical Analysis of On-line Learning of Probability
Distributions,
Toshiaki Aida, Physical Review Letters 83 (1999) 3554-3557.

ABSTRACT:
On-line learning of probability distributions is analyzed from the
field theoretical point of view. We can obtain an optimal on-line
learning algorithm, since renormalization group enables us to control
the number of degrees of freedom of a system according to the number
of examples. We do not learn parameters of a model, but probability
distributions themselves. Therefore, the algorithm requires no a priori
knowledge of a model.


Best regards,

Toshiaki Aida
Department of Physics, Tokyo Institute of Technology,
Department of Aeronautics, Tokyo Metropolitan College
of Aeronautical Engineering,
E-mail: aida at stat.phys.titech.ac.jp