T L H Watkin and J -P Nadal 1994 J. Phys. A: Math. Gen. 27 1899 doi:10.1088/0305-4470/27/6/016
T L H Watkin and J -P Nadal
Show affiliationsWe introduce an inferential approach to unsupervised learning which allows us to define an optimal learning strategy. Applying these ideas to a simple, previously studied model, we show that it is impossible to detect structure in data until a critical number of examples have been presented-an effect which will be observed in all problems with certain underlying symmetries. Thereafter, the advantage of optimal learning over previously studied learning algorithms depends critically upon the distribution of patterns; optimal learning may be exponentially faster. Models with more subtle correlations are harder to analyse, but in a simple limit of one such problem we calculate exactly the efficacy of an algorithm similar to some used in practice, and compare it to that of the optimal prescription.
05.10.-a Computational methods in statistical physics and nonlinear dynamics
02.50.-r Probability theory, stochastic processes, and statistics
07.05.Mh Neural networks, fuzzy logic, artificial intelligence
82C32 Neural nets (See also 68T05, 91E40, 92B20)
68Q32 Computational learning theory (See also 68T05)
62H30 Classification and discrimination; cluster analysis (See also 68T10)
Issue 6 (21 March 1994)
T L H Watkin and J -P Nadal 1994 J. Phys. A: Math. Gen. 27 1899
N Brunel et al 1992 J. Phys. A: Math. Gen. 25 5017
Y-H Chen 2008 J. Phys.: Conf. Ser. 96 012212
S S Sridharamurthy et al 2008 Smart Mater. Struct. 17 012001
Mikhail Yu. Kalmykov et al JHEP02(2007)040
Mikhail Yu. Kalmykov et al JHEP10(2007)048
Mikhail Yu. Kalmykov JHEP04(2006)056
Mikhail Y. Kalmykov et al JHEP11(2007)009
Yunxia Chen et al 2007 Nanotechnology 18 135708
I A Konyakhin et al 2006 J. Phys.: Conf. Ser. 48 998