25,161
edits
Changes
no edit summary
'''[[Main Page|Home]] * [[People]] * Alan J. Lockett'''
[[FILE:alan-lockett.jpg|border|right|thumb|link=http://nn.cs.utexas.edu/?alanlockett| Alan J. Lockett <ref>[http://nn.cs.utexas.edu/?alanlockett - NNRG People - Alan J. Lockett]</ref> ]]
'''Alan J. Lockett''',<br/>
an American computer scientist, Ph.D. from [https://en.wikipedia.org/wiki/University_of_Texas_at_Austin University of Texas at Austin] under [[Risto Miikkulainen]].
In his thesis covering on [[Learning|machine learning]] and [https://en.wikipedia.org/wiki/Global_optimization global optimization], he introduced the [https://en.wikipedia.org/wiki/Martingale_(probability_theory) martingale] based '''Evolutionary Annealing''',
which resembles an [[Genetic Programming#EvolutionaryAlgorithms|evolutionary algorithms]] to approximates samples from an increasingly sharp
[https://en.wikipedia.org/wiki/Boltzmann_distribution Boltzmann distribution], asymptotically focusing on the global optima.
'''Neuroannealing''' applies evolutionary annealing to evolve and learn [[Neural Networks|neural networks]]<ref>[[Alan J. Lockett]] ('''2012'''). ''General-Purpose Optimization Through Information Maximization''. Ph.D. thesis, [https://en.wikipedia.org/wiki/University_of_Texas_at_Austin University of Texas at Austin], advisor [[Risto Miikkulainen]]</ref>.His further research interests include [https://en.wikipedia.org/wiki/Humanoid_robot humanoid robotics], [[Deep Learning|deep learning]] and [[Opponent Model Search|opponent modelling]] in [[Games|games]].
=Selected Publications=
