Jump to content

ALOPEX

From Wikipedia, the free encyclopedia

ALOPEX (an abbreviation of "algorithms of pattern extraction") is a correlation based machine learning algorithm first proposed by Tzanakou and Harth in 1974.

Principle

[edit]

In machine learning, the goal is to train a system to minimize a cost function or (referring to ALOPEX) a response function. Many training algorithms, such as backpropagation, have an inherent susceptibility to getting "stuck" in local minima or maxima of the response function. ALOPEX uses a cross-correlation of differences and a stochastic process to overcome this in an attempt to reach the absolute minimum (or maximum) of the response function.

Method

[edit]

ALOPEX, in its simplest form is defined by an updating equation:

Δ Wij(n)=γ Δ Wij(n−1)Δ R(n)+ri(n)

where:

  • n≥0 is the iteration or time-step.
  • Δ Wij(n) is the difference between the current and previous value of system variable  Wij at iteration n.
  • Δ R(n) is the difference between the current and previous value of the response function  R, at iteration n.
  • γ is the learning rate parameter (γ <0 minimizes R, and γ >0 maximizes R )
  • ri(n)∼ N(0,σ 2)

Discussion

[edit]

Essentially, ALOPEX changes each system variable Wij(n) based on a product of: the previous change in the variable ΔWij(n−1), the resulting change in the cost function ΔR(n), and the learning rate parameter γ. Further, to find the absolute minimum (or maximum), the stochastic process rij(n) (Gaussian or other) is added to stochastically "push" the algorithm out of any local minima.

References

[edit]
  • Harth, E.; Tzanakou, E. (December 1974). "ALOPEX: A stochastic method for determining visual receptive fields". Vision Research. 14 (12): 1475–1482. doi:10.1016/0042-6989(74)90024-8. PMID 4446379.