A carregar...
Projeto de investigação
Sem título
Financiador
Autores
Publicações
General purpose optimization library (Gpol)
Publication . Bakurov, Illya; Buzzelli, Marco; Castelli, Mauro; Vanneschi, Leonardo; Schettini, Raimondo; NOVA Information Management School (NOVA IMS); Information Management Research Center (MagIC) - NOVA Information Management School; MDPI - Multidisciplinary Digital Publishing Institute
Several interesting libraries for optimization have been proposed. Some focus on individual optimization algorithms, or limited sets of them, and others focus on limited sets of problems. Frequently, the implementation of one of them does not precisely follow the formal definition, and they are difficult to personalize and compare. This makes it difficult to perform comparative studies and propose novel approaches. In this paper, we propose to solve these issues with the General Purpose Optimization Library (GPOL): a flexible and efficient multipurpose optimization library that covers a wide range of stochastic iterative search algorithms, through which flexible and modular implementation can allow for solving many different problem types from the fields of continuous and combinatorial optimization and supervised machine learning problem solving. Moreover, the library supports full-batch and mini-batch learning and allows carrying out computations on a CPU or GPU. The package is distributed under an MIT license. Source code, installation instructions, demos and tutorials are publicly available in our code hosting platform (the reference is provided in the Introduction).
A Study of Generalization and Fitness Landscapes for Neuroevolution
Publication . Rodrigues, Nuno M.; Silva, Sara; Vanneschi, Leonardo; NOVA Information Management School (NOVA IMS); Information Management Research Center (MagIC) - NOVA Information Management School; Institute of Electrical and Electronics Engineers (IEEE)
Fitness landscapes are a useful concept for studying the dynamics of meta-heuristics. In the last two decades, they have been successfully used for estimating the optimization capabilities of different flavors of evolutionary algorithms, including genetic algorithms and genetic programming. However, so far they have not been used for studying the performance of machine learning algorithms on unseen data, and they have not been applied to studying neuroevolution landscapes. This paper fills these gaps by applying fitness landscapes to neuroevolution, and using this concept to infer useful information about the learning and generalization ability of the machine learning method. For this task, we use a grammar-based approach to generate convolutional neural networks, and we study the dynamics of three different mutations used to evolve them. To characterize fitness landscapes, we study autocorrelation, entropic measure of ruggedness, and fitness clouds. Also, we propose the use of two additional evaluation measures: density clouds and overfitting measure. The results show that these measures are appropriate for estimating both the learning and the generalization ability of the considered neuroevolution configurations.
Vectorial GP for Alzheimer’s Disease Prediction Through Handwriting Analysis
Publication . Azzali, Irene; Cilia, Nicole Dalia; De Stefano, Claudio; Fontanella, Francesco; Giacobini, Mario; Vanneschi, Leonardo; NOVA Information Management School (NOVA IMS); Information Management Research Center (MagIC) - NOVA Information Management School
Alzheimer’s Disease (AD) is a neurodegenerative disease which causes a continuous cognitive decline. This decline has a strong impact on daily life of the people affected and on that of their relatives. Unfortunately, to date there is no cure for this disease. However, its early diagnosis helps to better manage the course of the disease with the treatments currently available. In recent years, AI researchers have become increasingly interested in developing tools for early diagnosis of AD based on handwriting analysis. In most cases, they use a feature engineering approach: domain knowledge by clinicians is used to define the set of features to extract from the raw data. In this paper, we present a novel approach based on vectorial genetic programming (VE_GP) to recognize the handwriting of AD patients. VE_GP is a recently defined method that enhances Genetic Programming (GP) and is able to directly manage time series in such a way to automatically extract informative features, without any need of human intervention. We applied VE_GP to handwriting data in the form of time series consisting of spatial coordinates and pressure. These time series represent pen movements collected from people while performing handwriting tasks. The presented experimental results indicate that the proposed approach is effective for this type of application. Furthermore, VE_GP is also able to generate rather small and simple models, that can be read and possibly interpreted. These models are reported and discussed in the Last part of the paper.
Using rapid chlorophyll fluorescence transients to classify vitis genotypes
Publication . Silva, Jorge Marques da; Figueiredo, Andreia; Cunha, Jorge; Eiras-Dias, José Eduardo; Silva, Sara; Vanneschi, Leonardo; Mariano, Pedro; NOVA Information Management School (NOVA IMS); Information Management Research Center (MagIC) - NOVA Information Management School; MDPI - Multidisciplinary Digital Publishing Institute
When a dark-adapted leaf is illuminated with saturating light, a fast polyphasic rise of fluorescence emission (Kautsky effect) is observed. The shape of the curve is dependent on the molecular organization of the photochemical apparatus, which in turn is a function of the interaction between genotype and environment. In this paper, we evaluate the potential of rapid fluorescence transients, aided by machine learning techniques, to classify plant genotypes. We present results of the application of several machine learning algorithms (k-nearest neighbors, decision trees, artificial neural networks, genetic programming) to rapid induction curves recorded in different species and cultivars of vine grown in the same environmental conditions. The phylogenetic relations between the selected Vitis species and Vitis vinifera cultivars were established with molecular markers. Both neural networks (71.8%) and genetic programming (75.3%) presented much higher global classification success rates than k-nearest neighbors (58.5%) or decision trees (51.6%), genetic programming performing slightly better than neural networks. However, compared with a random classifier (success rate = 14%), even the less successful algorithms were good at the task of classifying. The use of rapid fluorescence transients, handled by genetic programming, for rapid preliminary classification of Vitis genotypes is foreseen as feasible.
A GP approach for precision farming
Publication . Abbona, Francesca; Vanneschi, Leonardo; Bona, Marco; Giacobini, Mario; NOVA Information Management School (NOVA IMS); Information Management Research Center (MagIC) - NOVA Information Management School
Livestock is increasingly treated not just as food containers, but as animals that can be susceptible to stress and diseases, affecting, therefore, the production of offspring and the performance of the farm. The breeder needs a simple and useful tool to make the best decisions for his farm, as well as being able to objectively check whether the choices and investments made have improved or worsened its performance. The amount of data is huge but often dispersive: it is therefore essential to provide the farmer with a clear and comprehensible solution, that represents an additional investment. This research proposes a genetic programming approach to predict the yearly number of weaned calves per cow of a farm, namely the measure of its performance. To investigate the efficiency of genetic programming in such a problem, a dataset composed by observations on representative Piedmontese breedings was used. The results show that the algorithm is appropriate, and can perform an implicit feature selection, highlighting important variables and leading to simple and interpretable models.
Unidades organizacionais
Descrição
Palavras-chave
Contribuidores
Financiadores
Entidade financiadora
Fundação para a Ciência e a Tecnologia
Programa de financiamento
3599-PPCDT
Número da atribuição
PTDC/CCI-INF/29168/2017