Computational Neuroscience Seminar - LCN
25.08.09 Tuesday, 12h15, AAB032
Daan Wierstra, IDSIA Lugano, Switzerland (Homepage)
Natural Evolution Strategies
Abstract:
We introduce a novel stochastic search method, Natural
Evolution Strategies,
which constitutes a principled alternative to conventional
evolutionary
methods. Like evolutionary algorithms, this method is used
to optimize an
unknown `fitness' function. But instead of using
evolutionary concepts such as
mutation and selection, our algorithm maintains,
batch-by-batch, a multinormal
search distribution whose characterizing weights are
updated using the
natural gradient in the direction of higher expected
fitness. Heavily inspired
by recent developments in the policy gradient framework,
this method is derived
from first principles. Experimental work shows
the viability of this approach on a standard benchmark set
of both unimodal en
multimodal functions, and shows superior results on
learning the weights of
a difficult reinforcement-learned recurrent neural network
for non-Markov double pole balancing. back |