A Review Of Methods For Encoding Neural Network Topologies In Evolutionary Computation

Fekiac, Jozef; Zelinka, Ivan; Burguillo, Juan C.

doi:10.7148/2011-0410-0416

Cited by 26 publications

(8 citation statements)

References 9 publications

(6 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There is no "cookbook" how to set up the number of layers and nodes in layers. There are optimization techniques, which help with the structure, connections and node number, like evolutionary computation (Fekiac, 2011) or pseudo neural networks allow to produce a relation between input and output without setting of above mentioned parameters. PNN do not have the structure similar to ANN.…”

Section: Sít! S Dop#edn!m "í#Ením Chybymentioning

confidence: 99%

Prediction Of Raw Material Batches For The Production Of Clinker By Means Of Artificial Neural Networks -Analysis Of Behaviour

Oplatková¹,

Šenkeřík

2015

ECMS 2015 Proceedings Edited By: Valeri M. Mladenov, Petia Georgieva, Grisha Spasov, Galidiya Petrova

View full text Add to dashboard Cite

This research deals with the analysis of the behaviour of artificial neural nets for prediction of raw material batches for the production of clinker. During the production several oxides that are present in raw materials in quarries have to be extracted for homogenization of the mixture suitable for clinker production. There is some delay between the measurement of the mixture and the material which is send from quarry. It is necessary to "send" precise chemical composition to ensure a good quality of clinker and resulting product -cement. Artificial neural networks (ANN) are suitable for such kind of timeindependent prediction. The results show that not all oxides are necessary to use for the prediction of one oxide. The ANN were designed into several nets with one input similarly as pseudo neural networks are able to work. The results will be used for the purpose of further research of pseudo neural nets which currently serve only as classifiers.

show abstract

Section: Sít! S Dop#edn!m "í#Ením Chybymentioning

confidence: 99%

Prediction Of Raw Material Batches For The Production Of Clinker By Means Of Artificial Neural Networks -Analysis Of Behaviour

Oplatková¹,

Šenkeřík

2015

ECMS 2015 Proceedings Edited By: Valeri M. Mladenov, Petia Georgieva, Grisha Spasov, Galidiya Petrova

View full text Add to dashboard Cite

show abstract

“…Prosperity Measure: The other main goal of this framework is enabling automatic design of agent architectures by means of Evolutionary Algorithms (EAs). Here, similar methods to neuro-evolution Fekiac et al (2011) can be used. While omitting multi-objective Fig.…”

Section: Neural Modulementioning

confidence: 99%

Reusable Reinforcement Learning for Modular Self Motivated Agents

Vítků

Nahodil

2014

ECMS 2014 Proceedings Edited By: Flaminio Squazzoni, Fabio Baronio, Claudia Archetti, Marco Castellani

View full text Add to dashboard Cite

Presented topic is from the research fields called Artificial Life and Artificial Intelligence (AI). In this paper, there is presented novel approach to designing agent architectures with its requirements. The approach is inspired by inherited modularity of biological brains and agent architectures are represented here as set of given reusable modules connected into a particular topology. This paper presents design of two particular modules for future use in more complex architectures. The modules are used for implementing model-free motivation-driven Reinforcement Learning (RL). First, the novel framework for these architectures is described together with a used simulator. Then, the design of two new reusable domain-independent components of agent architectures is described. Finally, experimental validation of these new components and their future use is mentioned.

show abstract

“…In case of design networks with no constraints on topology, the space of possible solutions grows extremely fast. Nowadays, there are many solutions how to pose constraints on the topology, so the problem becomes feasible for EAs (Fekiac et al 2011). As a good example can be mentioned Cartesian Genetic Programming (CGP) (Fišer et al 2010) used for automatical design of logic circuits.…”

Section: Evolutionary Design Of Hybrid Networkmentioning

confidence: 99%

Autonomous Design Of Modular Intelligent Systems

Nahodil¹,

Vítků²

2013

ECMS 2013 Proceedings Edited By: Webjorn Rekdalsbakken, Robin T. Bye, Houxiang Zhang

View full text Add to dashboard Cite

We propose our original system capable of autonomous design of general-purpose complex modular hybrid systems. The resulting hybrid systems will be able to employ various techniques of learning, decisionmaking, prediction etc. Presented topic is from Artificial Life domain, but contributes also to fields such as Artificial Intelligence, Biology, Computational Neuroscience, Ethology, Cybernetics and potentially into many other aspects of research. The autonomous design is implemented as an optimization of system topology with respect to given problem. The principle of design is based on modified neuro-evolution and can be compared to modular neural networks. One of the main requirements is standardization of communication between very different subsystems. Here, each subsystem -module implements arbitrary algorithm and is treated as a Multiple-Input Multiple-Output subsystem. First, the design of simulator used is described, then the basic principle of hybrid networks is explained with it benefits and drawbacks. Finally, simple example is mentioned.

show abstract

A Review Of Methods For Encoding Neural Network Topologies In Evolutionary Computation

Cited by 26 publications

References 9 publications

Prediction Of Raw Material Batches For The Production Of Clinker By Means Of Artificial Neural Networks -Analysis Of Behaviour

Prediction Of Raw Material Batches For The Production Of Clinker By Means Of Artificial Neural Networks -Analysis Of Behaviour

Reusable Reinforcement Learning for Modular Self Motivated Agents

Autonomous Design Of Modular Intelligent Systems

Contact Info

Product

Resources

About