Autonomous Evolution of Topographic Regularities in Artificial Neural Networks

By Jason Gauci, Kenneth O. Stanley. 2010

164 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Cellular and Molecular Neuroscience, Cognitive Neuroscience, Optimization & Training, Neural Architecture Search & AutoML

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Cellular and Molecular Neuroscience

In Cognitive Neuroscience

In Optimization & Training

In Neural Architecture Search & AutoML

Citations in the corpus

Recurrent World Models Facilitate Policy Evolution (2018, methodology): “Before the popularity of Deep RL methods [54], evolution-based algorithms have been shown to be effective at solving RL tasks [18, 21, 22, 88, 92].”
World Models (2018, methodology): “Before the popularity of Deep RL methods (Mnih et al., 2013), evolution-based algorithms have been shown to be effective at solving RL tasks (Stanley & Miikkulainen, 2002; Gomez et al., 2008; Gomez & Schmidhuber, 2005; Gauci & Stanley, 2010; Sehnke et al., 2010; Miikkulainen, 2013).”
Generative NeuroEvolution for Deep Learning (2013, methodology): “Hypercube-based NEAT (HyperNEAT; [6, 11]) is a GDS extension of NEAT that enables effective evolution of high-dimensional ANNs.”
An efficient automated parameter tuning framework for spiking neural networks (2014, methodology): “…networks with as many as 8 million connections and networks evolved using NEAT have been used in food-gathering tasks (Stanley et al., 2009), in a checkers-playing ANN that exhibits topographic mappings (Gauci and Stanley, 2010), and in evolving robot gaits in hardware (Yosinski et al., 2011).”
Ieee Transactions on Evolutionary Computation on the Performance of Indirect Encoding across the Continuum of Regularity (unknown): “In the HyperNEAT algorithm , CPPNs encode ANNs instead of pictures, and evolution modifies the population of CPPNs [14], [19].”
Neuromodulation Improves the Evolution of Forward Models (2016, background): “Because we are interested in learning not only how to produce effective forward models for robotic control, but also to learn about the evolution of forward models in animals, the forward models in this paper are produced by evolving neural networks, a field called neuroevolution [5, 6, 10, 12, 27].”
Evolutionary Computation for Reinforcement Learning (2012, methodology): “Figure taken with permission from (Gauci and Stanley, 2010).”
Single-unit pattern generators for quadruped locomotion (2013, methodology): “HyperNEAT has been successful in a variety of challenging domains that require the discovery of regularities [11, 12, 34, 37].”
Reinforcement Learning for Improving Agent Design (2018, background): “Subsequent works further investigated morphology evolution [4, 8, 9, 11, 37, 44, 64, 65, 70], modular robotics [39, 45, 48, 75], and evolving soft robots [17, 20], using indirect encoding [5, 6, 7, 23, 61].”
Neuroevolutionary reinforcement learning for generalized control of simulated helicopters (2011, background): “Recent advances in indirect encodings such as HyperNEAT [22, 23, 75] could further improve performance.”
Indirectly Encoding Neural Plasticity as a Pattern of Local Rules (2010, methodology): “For a complete overview of HyperNEAT, see Gauci and Stanley [6] and Stanley et al.”
A unified approach to evolving plasticity and neural geometry (2012, background): “To this end, as such algorithms are asked to evolve increasingly large and complex structures, interest has increased in recent years in indirect neural network encodings, wherein the description of the solution is compressed such that information can be reused [2] [4] [14] [16] [27] [29].”
Guided self-organization in indirectly encoded and evolving topographic maps (2014, background): “proven in a number of challenging domains that require discovering regularities [3, 8, 9, 29].”