Efficient evolution of asymmetric recurrent neural networks using a PDGP-inspired two-dimensional representation

Pujol, João Carlos Figueira; Poli, Riccardo

doi:10.1007/bfb0055933

Cited by 21 publications

(13 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Poli [1997] presented a very general scheme wherein nodes are connected in a graph which is placed over a two-dimensional grid. Later, recurrent artificial neural networks were designed such that the nodes were synchronously parallel (e.g., [Pujol & Poli, 1998]) and variants exist in which some nodes can update more frequently than others (see [Poli, 1999] for an overview). Miller (e.g., [1999]) has presented a graph-based representation scheme originally designed to consider the hardware implementation of the evolved program wherein a twodimensional grid of sequentially updating, connected logic blocks is produced.…”

Section: Evolving Graph-based Programsmentioning

confidence: 99%

On dynamical genetic programming: simple Boolean networks in learning classifier systems

Bull

2009

International Journal of Parallel, Emergent and Distributed Sys

View full text Add to dashboard Cite

Abstract. Many representations have been presented to enable the effective evolution of computer programs. Turing was perhaps the first to present a general scheme by which to achieve this end.Significantly, Turing proposed a form of discrete dynamical system and yet dynamical representations remain almost unexplored within conventional genetic programming. This paper presents results from an initial investigation into using simple dynamical genetic programming representations within a Learning Classifier System. It is shown possible to evolve ensembles of dynamical Boolean function networks to solve versions of the well-known multiplexer problem. Both synchronous and asynchronous systems are considered.

show abstract

Section: Evolving Graph-based Programsmentioning

confidence: 99%

On dynamical genetic programming: simple Boolean networks in learning classifier systems

Bull

2009

International Journal of Parallel, Emergent and Distributed Sys

View full text Add to dashboard Cite

show abstract

“…Many schemes were considered in preparation of these experiments, prioritising flexibility, scalability, difficulty and efficiency. These included Connectionist Encoding [17], Node Based Encoding [18], Graph Based Encoding [19], Layer Based Encoding [20], Marker Based Encoding [21], Matrix Re-writing [22], [23], Cellular Encoding [24], Weight-based encoding [25], [26] and Architecture encoding [27].…”

Section: Encoding and Decoding Schemesmentioning

confidence: 99%

The effect of noise on the performance of cultural evolution in multi-agent systems

Curran¹,

O’Riordan²

Proceedings of the 2004 Congress on Evolutionary Computation (IEEE Cat. No.04TH8753)

View full text Add to dashboard Cite

Abstract-This paper examines the effect of the addition noise to the cultural learning process of a population of agents. Experiments are undertaken using an artificial life simulator capable of simulating population learning (through genetic algorithms) and lifetime learning (through the use of neural networks). To simulate cultural learning, (the exchange of information through non-genetic means) a group of highly fit agents is selected at each generation to function as teachers which are assigned a number of pupils to instruct. Cultural exchanges occur through a hidden layer of an agent's neural network known as the verbal layer. Through the use of back-propagation, a pupil agent imitates the teacher's behaviour and overall population fitness is increased.We show that the addition of noise to cultural exchanges can improve on the performance of cultural learning.

show abstract

“…Significantly, recursive connections are permitted and each node is executed with synchronous parallelism for some number of cycles before an output node's value is taken. Poli (e.g., [39]) presented a similar scheme wherein the graph is placed over a two-dimensional grid and executes its nodes synchronously in parallel. Connections are directed upwards and are only permitted between nodes situated on adjacent rows; however by including identity (i.e., pass-through) functions, connections between non-adjacent layers are possible and thus any parallel distributed program may be represented.…”

Section: Graph Representationsmentioning

confidence: 99%

Fuzzy dynamical genetic programming in XCSF

Preen

Bull

2011

Proceedings of the 13th Annual Conference Companion on Genetic and Evolutionary Computation

View full text Add to dashboard Cite

Abstract-A number of representation schemes have been presented for use within Learning Classifier Systems, ranging from binary encodings to Neural Networks, and more recently Dynamical Genetic Programming (DGP). This paper presents results from an investigation into using a fuzzy DGP representation within the XCSF Learning Classifier System. In particular, asynchronous Fuzzy Logic Networks are used to represent the traditional condition-action production system rules. It is shown possible to use self-adaptive, open-ended evolution to design an ensemble of such fuzzy dynamical systems within XCSF to solve two well-known test problems.

show abstract

Efficient evolution of asymmetric recurrent neural networks using a PDGP-inspired two-dimensional representation

Cited by 21 publications

References 22 publications

On dynamical genetic programming: simple Boolean networks in learning classifier systems

On dynamical genetic programming: simple Boolean networks in learning classifier systems

The effect of noise on the performance of cultural evolution in multi-agent systems

Fuzzy dynamical genetic programming in XCSF

Contact Info

Product

Resources

About