Options
Nicolau, Miguel
Preferred name
Nicolau, Miguel
Official Name
Nicolau, Miguel
Research Output
Now showing 1 - 10 of 12
- PublicationA GAuGE Approach to Learning DFA from Noisy Samples(2004-06-30)
; ; This paper describes the adaptation of the GAuGE system to classify binary sequences generated by random DFA. Experiments were conducted, which, although not highly successful, illustrate the potential of applying GAuGE like systems to this problem domain.104 - PublicationCrossover, Population Dynamics and Convergence in the GAuGE SystemThis paper presents a study of the effectiveness of a recently presented crossover operator for the GAuGE system. This crossover, unlike the traditional crossover employed previously, preserves the association of positions and values which exists in GAuGE genotype strings, and as such is more adequate for problems where the meaning of an allele is dependent on its placement in the phenotype string. Results obtained show that the new operator improves the performance of the GAuGE system on simple binary problems, both when position-sensitive data is manipulated and not.
360Scopus© Citations 1 - PublicationHow Functional Dependency Adapts to Salience Hierarchy in the GAuGE SystemGAuGE is a position independent genetic algorithm that suffers from neither under nor over-specification, and uses a genotype to phenotype mapping process. By specifying both the position and the value of each gene, it has the potential to group important data together in the genotype string, to prevent it from being broken up and disrupted during the evolution process. To test this ability, GAuGE was applied to a set of problems with exponentially scaled salience. The results obtained demonstrate that GAuGE is indeed moving the more salient genes to the start of the genotype strings, creating robust individuals that are built in a progressive fashion from the left to the right side of the genotype.
405Scopus© Citations 6 - PublicationIntroducing Grammar Based Extensions for Grammatical EvolutionThis paper presents a series of extensions to standard Grammatical Evolution. These grammar-based extensions facilitate the exchange of knowledge between genotype and phenotype strings, thus establishing a better correlation between the search and solution spaces, typically separated in Grammatical Evolution. The results obtained illustrate the practical advantages of these extensions, both in terms of convenience and potential increase in performance.
581 - PublicationEfficient Crossover in the GAuGE SystemThis paper presents a series of context-preserving crossover operators for the GAuGE system. These operators have been designed to respect the representation of genotype strings in GAuGE, thereby making sensible changes at the genotypic level. Results on a set of problems suggest that some of these operators can improve the maintenance and propagation of building blocks in GAuGE, as well as its scalability, and could be of use to other systems using structural evolving genomes.
341Scopus© Citations 3 - PublicationFunctional Dependency and Degeneracy: Detailed Analysis of the GAuGE SystemThis paper explores the mapping process of the GAuGE system, a recently introduced position-independent genetic algorithm, that encodes both the positions and the values of individuals at the genotypic level. A mathematical formalisation of its mapping process is presented, and is used to characterise the functional dependency feature of the system. An analysis of the effect of degeneracy in this functional dependency is then performed, and a mathematical theorem is given, showing that the introduction of degeneracy reduces the position specification bias of individuals. Experimental results are given, that backup these findings.
329Scopus© Citations 5 - PublicationDeep Evolution of Feature Representations for Handwritten Digit Recognition(IEEE, 2015-05-28)
; ; ; ; ; A training protocol for learning deep neural networks, called greedy layer-wise training, is applied to the evolution of a hierarchical, feed-forward Genetic Programming based system for feature construction and object recognition. Results on a popular handwritten digit recognition benchmark clearly demonstrate that two layers of feature transformations improves generalisation compared to a single layer. In addition, we show that the proposed system outperforms several standard Genetic Programming systems, which are based on hand-designed features, and use different program representations and fitness functions.313Scopus© Citations 11 - PublicationGenetic Algorithms Using Grammatical EvolutionThis paper describes the GAUGE system, Genetic Algorithms Using Grammatical Evolution. GAUGE is a position independent Genetic Algorithm that uses Grammatical Evolution with an attribute grammar to dictate what position a gene codes for. GAUGE suffers from neither under-specification nor over-specification, is guaranteed to produce syntactically correct individuals, and does not require any repair after the application of genetic operators. GAUGE is applied to the standard onemax problem, with results showing that its genotype to phenotype mapping and position independence nature do not affect its performance as a normal genetic algorithm. A new problem is also presented, a deceptive version of the Mastermind game, and we show that GAUGE possesses the position independence characteristics it claims, and outperforms several genetic algorithms, including the competent genetic algorithm messyGA.
552Scopus© Citations 31 - PublicationGenetic Operators and Sequencing in the GAuGE SystemThis paper investigates the effects of the mapping process employed by the GAuGE system on standard genetic operators. It is shown that the application of that mapping process transforms these operators into suitable sequencing searching tools. A practical application is analysed, and its results compared with a standard genetic algorithm, using the same operators. Results and analysis highlight the suitability of GAuGE and its operators, for this class of problems.
393 - PublicationSolving Sudoku with the GAuGE SystemThis paper presents an evolutionary approach to solving Sudoku puzzles. Sudoku is an interesting problem because it is a challenging logical puzzle that has previously only been solved by computers using various brute force methods, but it is also an abstract form of a timetabling problem, and is scalably difficult. A different take on the problem, motivated by the desire to be able to generalise it, is presented. The GAuGE system was applied to the problem, and the results obtained show that its mapping process is well suited for this class of problems.
609Scopus© Citations 12