Skip to main content

Resultados relacionados

Resultados totales (Incluyendo duplicados): 25
Encontrada(s) 1 página(s)

RUC. Repositorio da Universidade da Coruña
Publication Contribución de congreso (ConferenceOutput)

A Machine Learning Solution for Distributed Environments and Edge Computing

  • Penas-Noce, Javier
  • Fontenla-Romero, Óscar
  • Guijarro-Berdiñas, Bertha
[Abstract] In a society in which information is a cornerstone the exploding of data is crucial. Thinking of the Internet of Things, we need systems able to learn from massive data and, at the same time, being inexpensive and of reduced size. Moreover, they should operate in a distributed manner making use of edge computing capabilities while preserving local data privacy. The aim of this work is to provide a solution offering all these features by implementing the algorithm LANN-DSVD over a cluster of Raspberry Pi devices. In this system, every node first learns locally a one-layer neural network. Later on, they share the weights of these local networks to combine them into a global net that is finally used at every node. Results demonstrate the benefits of the proposed system., This research was funded by the Spanish Secretaría de Estado de Universidades e I+D+i (Grant TIN2015-65069-C2-1-R), Xunta de Galicia (Grants ED431C2018/34, ED341D R2016/045) and FEDER funds., Xunta de Galicia; ED431C2018/34, Xunta de Galicia; ED341D R2016/045
RUC. Repositorio da Universidade da Coruña
Publication Artículo científico (JournalArticle)

Fast Distributed kNN Graph Construction Using Auto-tuned Locality-sensitive Hashing

  • Eiras-Franco, Carlos
  • Martínez Rego, David
  • Kanthan, Leslie
  • Piñeiro, César
  • Bahamonde, Antonio
  • Guijarro-Berdiñas, Bertha
  • Alonso-Betanzos, Amparo
© 2020 Authors|ACM. This is the author's version of the work. It is posted here for your personal use. Not for redistribution. The definitive Version of Record was published in ACM Trans. Intell. Syst. Technol. 11, 6, Article 71 (October 2020). https://doi.org/10.1145/3408889., [Abstract]: The k-nearest-neighbors (kNN) graph is a popular and powerful data structure that is used in various areas of Data Science, but the high computational cost of obtaining it hinders its use on large datasets. Approximate solutions have been described in the literature using diverse techniques, among which Locality-sensitive Hashing (LSH) is a promising alternative that still has unsolved problems. We present Variable Resolution Locality-sensitive Hashing, an algorithm that addresses these problems to obtain an approximate kNN graph at a significantly reduced computational cost. Its usability is greatly enhanced by its capacity to automatically find adequate hyperparameter values, a common hindrance to LSH-based methods. Moreover, we provide an implementation in the distributed computing framework Apache Spark that takes advantage of the structure of the algorithm to efficiently distribute the computational load across multiple machines, enabling practitioners to apply this solution to very large datasets. Experimental results show that our method offers significant improvements over the state-of-the-art in the field and shows very good scalability as more machines are added to the computation., This research has been supported in part by the Spanish Ministerio de Economía y Competitividad (Projects No. TIN 2015-65069-C2-1-R and No. 2-R), partially funded by FEDER funds of the EU and by the Xunta de Galicia (Project No. ED431C 2018/34 and Centro singular de investigación de Galicia, accreditation 2016–2019), Xunta de Galicia; ED431C 2018/34
RUC. Repositorio da Universidade da Coruña
Publication Artículo científico (JournalArticle)

Parallel feature selection for distributed-memory clusters

  • González-Domínguez, Jorge
  • Bolón-Canedo, Verónica
  • Freire, Borja
  • Touriño, Juan
Versión final aceptada de: https://doi.org/10.1016/j.ins.2019.01.050, This manuscript version is made available under the CC-BY-NC-ND 4.0 license https://creativecommons.org/ licenses/by-nc-nd/4.0/. This version of the article: González-Domínguez, J. et al. (2019) ‘Parallel feature selection for distributed-memory clusters’, has been accepted for publication in Information Sciences, 496, pp. 399–409. The Version of Record is available online at: https://doi.org/10.1016/j.ins.2019.01.050, [Abstract]: Feature selection is nowadays an extremely important data mining stage in the field of machine learning due to the appearance of problems of high dimensionality. In the literature there are numerous feature selection methods, mRMR (minimum-Redundancy-Maximum-Relevance) being one of the most widely used. However, although it achieves good results in selecting relevant features, it is impractical for datasets with thousands of features. A possible solution to this limitation is the use of the fast-mRMR method, a greedy optimization of the mRMR algorithm that improves both scalability and efficiency. In this work we present fast-mRMR-MPI, a novel hybrid parallel implementation that uses MPI and OpenMP to accelerate feature selection on distributed-memory clusters. Our performance evaluation on two different systems using five representative input datasets shows that fast-mRMR-MPI is significantly faster than fast-mRMR while providing the same results. As an example, our tool needs less than one minute to select 200 features of a dataset with more than four million features and 16,000 samples on a cluster with 32 nodes (768 cores in total), while the sequential fast-mRMR required more than eight hours. Moreover, fast-mRMR-MPI distributes data so that it is able to exploit the memory available on different nodes of a cluster and then complete analyses that fail on a single node due to memory constraints. Our tool is publicly available at https://github.com/borjaf696/Fast-mRMR., This research has been partially funded by projects TIN2016-75845-P and TIN-2015-65069-C2-1-R of the Ministry of Economy, Industry and Competitiveness of Spain, as well as by Xunta de Galicia projects ED431D R2016/045 and GRC2014/035, all of them partially funded by FEDER funds of the European Union. We gratefully thank CESGA for providing access to the Finis Terrae II supercomputer., Xunta de Galicia; ED431D R2016/045, Xunta de Galicia; GRC2014/035
RUC. Repositorio da Universidade da Coruña
Publication Artículo científico (JournalArticle)

Scalable Feature Selection Using ReliefF Aided by Locality-Sensitive Hashing

  • Eiras-Franco, Carlos
  • Guijarro-Berdiñas, Bertha
  • Alonso-Betanzos, Amparo
  • Bahamonde, Antonio
Financiado para publicación en acceso aberto: Universidade da Coruña/CISUG, [Abstract] Feature selection algorithms, such as ReliefF, are very important for processing high-dimensionality data sets. However, widespread use of popular and effective such algorithms is limited by their computational cost. We describe an adaptation of the ReliefF algorithm that simplifies the costliest of its step by approximating the nearest neighbor graph using locality-sensitive hashing (LSH). The resulting ReliefF-LSH algorithm can process data sets that are too large for the original ReliefF, a capability further enhanced by distributed implementation in Apache Spark. Furthermore, ReliefF-LSH obtains better results and is more generally applicable than currently available alternatives to the original ReliefF, as it can handle regression and multiclass data sets. The fact that it does not require any additional hyperparameters with respect to ReliefF also avoids costly tuning. A set of experiments demonstrates the validity of this new approach and confirms its good scalability., This study has been supported in part by the Spanish Ministerio de Economía y Competitividad (projects PID2019-109238GB-C2 and TIN 2015-65069-C2-1-R and 2-R), partially funded by FEDER funds of the EU and by the Xunta de Galicia (projects ED431C 2018/34 and Centro Singular de Investigación de Galicia, accreditation 2016-2019). The authors wish to thank the Fundación Pública Galega Centro Tecnolóxico de Supercomputación de Galicia (CESGA) for the use of their computing resources. Funding for open access charge: Universidade da Coruña/CISUG, Xunta de Galicia; ED431C 2018/34
RUC. Repositorio da Universidade da Coruña
Publication Artículo científico (JournalArticle)

Wavefront Marching Methods: A Unified Algorithm to Solve Eikonal and Static Hamilton-Jacobi Equations

  • Cancela, Brais
  • Alonso-Betanzos, Amparo
© 2020 IEEE. This version of the article has been accepted for publication, after peer review. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. The Version of Record is available online at: https://doi.org/10.1109/TPAMI.2020.2993500, [Abstract]: This paper presents a unified propagation method for dealing with both the classic Eikonal equation, where the motion direction does not affect the propagation, and the more general static Hamilton-Jacobi equations, where it does. While classic Fast Marching Method (FMM) techniques achieve the solution to the Eikonal equation with a O(M log M) (or O(M) assuming some modifications), solving the more general static Hamilton-Jacobi equation requires a higher complexity. The proposed framework maintains the O(M log M) complexity for both problems, while achieving higher accuracy than available state-of-the-art. The key idea behind the proposed method is the creation of ‘mini wave-fronts’, where the solution is interpolated to minimize the discretization error. Experimental results show how our algorithm can outperform the state-of-the-art both in precision and computational cost., The authors would like to thank to the financial support of the Spanish Ministerio de Economıa y Competitividad (research project TIN2015-65069-C2-1-R), the Xunta de Galicia (research projects ED431C 2018/34 and Centro Singular de Investigacion de Galicia, accreditation 2016-2019) and by the European Union (European Regional Development Fund). Brais Cancela acknowledges the support of the Xunta de Galicia under its postdoctoral program., Xunta de Galicia; ED431C 2018/34
RUC. Repositorio da Universidade da Coruña
Publication Artículo científico (JournalArticle)

A scalable saliency-based feature selection method with instance-level information

  • Cancela, Brais
  • Bolón-Canedo, Verónica
  • Alonso-Betanzos, Amparo
  • Gama, João
© 2019. This manuscript version is made available under the CC-BY-NC-ND 4.0 license https://creativecommons.org/licenses/by-nc-nd/4.0/. This version of the article [Cancela, Brais, et al. «A Scalable Saliency-Based Feature Selection Method with Instance-Level Information». Knowledge-Based Systems, vol. 192, marzo de 2020, p. 105326] has been accepted for publication in "Knowledge-Based Systems". The Version of Record is available online at https://doi.org/10.1016/j.knosys.2019.105326., [Abstract]: Classic feature selection techniques remove irrelevant or redundant features to achieve a subset of relevant features in compact models that are easier to interpret and so improve knowledge extraction. Most such techniques operate on the whole dataset, but are unable to provide the user with useful information when only instance-level information is required; in other words, classic feature selection algorithms do not identify the most relevant information in a sample. We have developed a novel feature selection method, called saliency-based feature selection (SFS), based on deep-learning saliency techniques. Our algorithm works under any architecture that is trained by using gradient descent techniques (Neural Networks, SVMs, …), and can be used for classification or regression problems. Experimental results show our algorithm is robust, as it allows to transfer the feature ranking result between different architectures, achieving remarkable results. The versatility of our algorithm has been also demonstrated, as it can work either in big data environments as well as with small datasets., This research was financially supported in part by the Spanish Ministerio de Economía y Competitividad (research project TIN2015-65069-C2-1-R), by the Xunta de Galicia (research projects ED431C 2018/34 and Centro Singular de Investigación de Galicia, accreditation 2016–2019) and by the European Union (European Regional Development Fund) . We also gratefully acknowledge the support of NVIDIA Corporation with the donation of the Titan Xp GPU used for this research. Brais Cancela acknowledges the support of the Xunta de Galicia under its postdoctoral program., Xunta de Galicia; ED431C 2018/34
RUC. Repositorio da Universidade da Coruña
Publication Artículo científico (JournalArticle)

Multithreaded and Spark parallelization of feature selection filters

  • Eiras-Franco, Carlos
  • Bolón-Canedo, Verónica
  • Ramos Garea, Sabela
  • González-Domínguez, Jorge
  • Alonso-Betanzos, Amparo
  • Touriño, Juan
©2016 Elsevier B.V. All rights reserved. This manuscript version is made available under the CC-BY-NC-ND 4.0 license https://creativecommons.org/licenses/bync-nd/4.0/. This version of the article has been accepted for publication in Journal of Computational Science. The Version of Record is available online at https://doi.org/10.1016/j.jocs.2016.07.002, Versión final aceptada de: C. Eiras-Franco, V. Bolón-Canedo, S. Ramos, J. González-Domínguez, A. Alonso-Betanzos, and J. Touriño, "Multithreaded and Spark parallelization of feature selection filters", Journal of Computational Science, Vol. 17, Part 3, Nov. 2016, Pp. 609-619, [Abstract]: Vast amounts of data are generated every day, constituting a volume that is challenging to analyze. Techniques such as feature selection are advisable when tackling large datasets. Among the tools that provide this functionality, Weka is one of the most popular ones, although the implementations it provides struggle when processing large datasets, requiring excessive times to be practical. Parallel processing can help alleviate this problem, effectively allowing users to work with Big Data. The computational power of multicore machines can be harnessed by using multithreading and distributed programming, effectively helping to tackle larger problems. Both these techniques can dramatically speed up the feature selection process allowing users to work with larger datasets. The reimplementation of four popular feature selection algorithms included in Weka is the focus of this work. Multithreaded implementations previously not included in Weka as well as parallel Spark implementations were developed for each algorithm. Experimental results obtained from tests on real-world datasets show that the new versions offer significant reductions in processing times., This work has been financed in part by Xunta de Galicia under Research Network R2014/041 and project GRC2014/035, and by Spanish Ministerio de Economía y Competitividad under projects TIN2012-37954 and TIN-2015-65069-C2-1-R, partially funded by FEDER funds of the European Union. V. Bolón-Canedo acknowledges support of the Xunta de Galicia under postdoctoral Grant code ED481B 2014/164-0. Additionally, the collaboration of Jorge Veiga on setting up and using the MREv tool for Spark execution was essential for this work., Xunta de Galicia; R2014/041, Xunta de Galicia; GRC2014/035, Xunta de Galicia; ED481B 2014/164-0
RUC. Repositorio da Universidade da Coruña
Publication Artículo científico (JournalArticle)

Ensembles for feature selection: A review and future trends

  • Bolón-Canedo, Verónica
  • Alonso-Betanzos, Amparo
© 2019. This manuscript version is made available under the CC-BY-NC-ND 4.0 license https://creativecommons.org/licenses/by-nc-nd/4.0/. This version of the article: Bolón-Canedo, V. and Alonso-Betanzos, A. (2019) ‘Ensembles for Feature Selection: A Review and Future Trends’ has been accepted for publication in: Information Fusion, 52, pp. 1–12. The Version of Record is available online at https://doi.org/10.1016/j.inffus.2018.11.008., [Abstract]: Ensemble learning is a prolific field in Machine Learning since it is based on the assumption that combining the output of multiple models is better than using a single model, and it usually provides good results. Normally, it has been commonly employed for classification, but it can be used to improve other disciplines such as feature selection. Feature selection consists of selecting the relevant features for a problem and discard those irrelevant or redundant, with the main goal of improving classification accuracy. In this work, we provide the reader with the basic concepts necessary to build an ensemble for feature selection, as well as reviewing the up-to-date advances and commenting on the future trends that are still to be faced., This research has been financially supported in part by the Spanish Ministerio de Economa y Competitividad (research project TIN 2015-65069-C2-1-R), by the Xunta de Galicia (research projects GRC2014/035 and the Centro Singular de Investigación de Galicia, accreditation 2016–2019, Ref. ED431G/01) and by the European Union (FEDER/ERDF)., Xunta de Galicia; GRC2014/035, Xunta de Galicia; ED431G/01
RUC. Repositorio da Universidade da Coruña
Publication Contribución de congreso (ConferenceOutput)

Can data placement be effective for Neural Networks classification tasks? Introducing the Orthogonal Loss

  • Cancela, Brais
  • Bolón-Canedo, Verónica
  • Alonso-Betanzos, Amparo
Presented at: 5th International Conference on Pattern Recognition, ICPR 2020V, Virtual, Milan, 10-15 January 2021, This version of the paper has been accepted for publication. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. The final published paper is available online at: https://doi.org/10.1109/ICPR48806.2021.9412704, [Abstract]: Traditionally, a Neural Network classification training loss function follows the same principle: minimizing the distance between samples that belong to the same class, while maximizing the distance to the other classes. There are no restrictions on the spatial placement of deep features (last layer input). This paper addresses this issue when dealing with Neural Networks, providing a set of loss functions that are able to train a classifier by forcing the deep features to be projected over a predefined orthogonal basis. Experimental results shows that these `data placement' functions can overcome the training accuracy provided by the classic cross-entropy loss function., This research has been financially supported in part by European Union ERDF funds, by the Spanish Ministerio de Economía y Competitividad (research projects TIN2015-65069-C2 and PID2019-109238GB-C22), and by the Xunta de Galicia (research projects GRC2014/035 and ED431G/01). We also gratefully acknowledge the support of NVIDIA Corporation with the donation of the Titan Xp GPU used for this research. Brais Cancela acknowledges the support of the Xunta de Galicia under its postdoctoral program., Xunta de Galicia; GRC2014/035, Xunta de Galicia; ED431G/01
RUC. Repositorio da Universidade da Coruña
Publication Artículo científico (JournalArticle)

On developing an automatic threshold applied to feature selection ensembles

  • Seijo Pardo, Borja
  • Bolón-Canedo, Verónica
  • Alonso-Betanzos, Amparo
© 2019. This manuscript version is made available under the CC-BY-NC-ND 4.0 license https://creativecommons.org/licenses/by-nc-nd/4.0/. This version of the article "R.-J. Palma-Mendoza, L. de-Marcos, D. Rodriguez, y A. Alonso-Betanzos, «B. Seijo-Pardo, V. Bolón-Canedo, y A. Alonso-Betanzos, «On developing an automatic threshold applied to feature selection ensembles», Information Fusion, vol. 45, pp. 227-245, ene. 2019" has been accepted for publication in Information Fusion. The Version of Record is available online at https://doi.org/10.1016/j.inffus.2018.02.007, [Abstract]: Feature selection ensemble methods are a recent approach aiming at adding diversity in sets of selected features, improving performance and obtaining more robust and stable results. However, using an ensemble introduces the need for an aggregation step to combine all the output methods that confirm the ensemble. Besides, when trying to improve computational efficiency, ranking methods that order all initial features are preferred, and so an additional thresholding step is also mandatory. In this work two different ensemble designs based on ranking methods are described. The main difference between them is the order in which the combination and thresholding steps are performed. In addition, a new automatic threshold based on the combination of three data complexity measures is proposed and compared with traditional thresholding approaches based on retaining a fixed percentage of features. The behavior of these methods was tested, according to the SVM classification accuracy, with satisfactory results, for three different scenarios: synthetic datasets and two types of real datasets (where sample size is much higher than feature size, and where feature size is much higher than sample size)., This research has been financially supported in part by the Spanish Ministerio de Economa y Competitividad (research project TIN 2015-65069-C2-1-R), by the Xunta de Galicia (research projects GRC2014/035 and the Centro Singular de Investigación de Galicia, accreditation 2016–2019) and by the European Union (FEDER/ERDF)., Xunta de Galicia; GRC2014/035
RUC. Repositorio da Universidade da Coruña
Publication Artículo científico (JournalArticle)

DSVD-autoencoder: A scalable distributed privacy-preserving method for one-class classification

  • Fontenla-Romero, Óscar
  • Pérez-Sánchez, Beatriz
  • Guijarro-Berdiñas, Bertha
This is the peer reviewed version of the following article: DSVD-autoencoder: A scalable distributed privacy-preserving method for one-class classification, which has been published in final form at https://doi.org/10.1002/int.22296. This article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Use of Self-Archived Versions. This article may not be enhanced, enriched or otherwise transformed into a derivative work, without express permission from Wiley or by statutory rights under applicable legislation. Copyright notices must not be removed, obscured or modified. The article must be linked to Wiley’s version of record on Wiley Online Library and any embedding, framing or otherwise making available the article or pages thereof by third parties from platforms, services and websites other than Wiley Online Library must be prohibited., [Abstract]: One-class classification has gained interest as a solution to certain kinds of problems typical in a wide variety of real environments like anomaly or novelty detection. Autoencoder is the type of neural network that has been widely applied in these one-class problems. In the Big Data era, new challenges have arisen, mainly related with the data volume. Another main concern derives from Privacy issues when data is distributed and cannot be shared among locations. These two conditions make many of the classic and brilliant methods not applicable. In this paper, we present distributed singular value decomposition (DSVD-autoencoder), a method for autoencoders that allows learning in distributed scenarios without sharing raw data. Additionally, to guarantee privacy, it is noniterative and hyperparameter-free, two interesting characteristics when dealing with Big Data. In comparison with the state of the art, results demonstrate that DSVD-autoencoder provides a highly competitive solution to deal with very large data sets by reducing training from several hours to seconds while maintaining good accuracy., This work has been supported by the grant Machine Learning on the Edge - Ayudas Fundación BBVA a Equipos de Investigación Científica 2019. It has also been possible thanks to the support received by theNational Plan for Scientific and Technical Research and Innovation of the Spanish Government (Grants TIN2015-65069-C2-1-R and PID2019-109238GB-C2), and by the Xunta de Galicia (Grant ED431C 2018/34) with the European Union ERDF funds. CITIC, as a Research Center of the University System of Galicia, is funded by Consellería de Educación, Universidade e Formación Profesional of the Xunta de Galicia through the European Regional Development Fund (ERDF) and the Secretaría Xeral de Universidades (Ref. ED431G 2019/01)., Xunta de Galicia; ED431C 2018/34, Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publication Artículo científico (JournalArticle)

Distributed classification based on distances between probability distributions in feature space

  • Montero Manso, Pablo
  • Morán-Fernández, Laura
  • Bolón-Canedo, Verónica
  • Vilar, José
  • Alonso-Betanzos, Amparo
© 2017. This manuscript version is made available under the CC-BY-NC-ND 4.0 license https://creativecommons.org/licenses/by-nc-nd/4.0/. This version of the article: Montero-Manso, P., Morán-Fernández, L., Bolón-Canedo, V., Vilar, J. A., & Alonso-Betanzos, A. (2019). ‘Distributed classification based on distances between probability distributions in feature space’ has been accepted for publication in: Information Sciences, 496, 431–450. The Version of Record is available online at https://doi.org/10.1016/j.ins.2018.12.044., [Abstract]: We consider a distributed framework where training and test samples drawn from the same distribution are available, with the training instances spread across disjoint nodes. In this setting, a novel learning algorithm based on combining with different weights the outputs of classifiers trained at each node is proposed. The weights depend on the distributional distance between each node and the test set in the feature space. Two different weighting approaches are introduced, which are referred to as per-Node Weighting (pNW) and per-Instance Weighting (pIW). While pNW assigns the same weight to all test instances at each node, pIW allows distinct weights for test instances differently represented at the node. By construction, our approach is particularly useful to deal with unbalanced nodes. Our methods require no communication between nodes, allowing for data privacy, independence of the kind of trained classifier at each node and maximum training speedup. In fact, our methods do not require retraining of the node’s classifiers if available. Although a range of different combination rules are considered to ensemble the single classifiers, theoretical support for the optimality of using the sum rule is provided. Our experiments illustrate all of these properties and show that pIW produces the highest classification accuracies compared with pNW and the standard unweighted approaches., This research has been supported by Spanish Ministerio de Economía y Competitividad (grants TIN2015-65069-C2-1-R, MTM2014-52876-R and MTM2017-82724-R), and by the Consellería de Industria of the Xunta de Galicia (projects GRC2014/035, Grupos de Referencia Competitiva ED431D-R2016/045 and ED431C-2016-015, and Centro Singular de Investigación de Galicia ED431G/01), all of them through the European Regional Development Fund, ERDF., Xunta de Galicia; GRC2014/035, Xunta de Galicia; ED431D-R2016/045, Xunta de Galicia; ED431C-2016-015, Xunta de Galicia; ED431G/01
RUC. Repositorio da Universidade da Coruña
Publication Artículo científico (JournalArticle)

Community detection and social network analysis based on the Italian wars of the 15th century

  • Fumanal-Idocin, Javier
  • Alonso-Betanzos, Amparo
  • Cordón, Oscar
  • Bustince, Humberto
  • Minárová, Mária
© 2020. This manuscript version is made available under the CC-BY-NC-ND 4.0 license https://creativecommons.org/licenses/by-nc-nd/4.0/. This version of the article: Fumanal-Idocin, J., Alonso-Betanzos, A., Cordón, O., Bustince, H., & Minárová, M. (2020). ‘Community detection and social network analysis based on the Italian wars of the 15th century’ has been accepted for publication in: Future Generation Computer Systems, 113, 25–40. The Version of Record is available online at https://doi.org/10.1016/j.future.2020.06.030, [Abstract]: In this contribution we study social network modelling by using human interaction as a basis. To do so, we propose a new set of functions, affinities, designed to capture the nature of the local interactions among each pair of actors in a network. By using these functions, we develop a new community detection algorithm, the Borgia Clustering, where communities naturally arise from the multi-agent interaction in the network. We also discuss the effects of size and scale for communities regarding this case, as well as how we cope with the additional complexity present when big communities arise. Finally, we compare our community detection solution with other representative algorithms, finding favourable results., Javier Fumanal Idocin’s and Humberto Bustince’s research has been supported by the project TIN2016-77356-P (AEI/FEDER,UE). Oscar Cordón’s research was supported by the Spanish Ministry of Science, Innovation and Universities under grant EX-ASOCO (PGC2018-101216-B-I00), including, European Regional Development Funds (ERDF). Amparo Alonso-Betanzos’ research has been financially supported in part by the Spanish Ministerio de Economía y Competitividad (research project TIN2015-65069-C2-1-R), by European Union FEDER funds and by the Consellería de Industria of the Xunta de Galicia, Spain (research project GRC2014 /035). M. Minárová’s research has been funded by the project work was supported by the project APVV-17-0066., Xunta de Galicia; GRC2014 /035, Slovak Republic. Slovak Research and Development Agency; APVV-17-0066
RUC. Repositorio da Universidade da Coruña
Publication Artículo científico (JournalArticle)

CUDA-JMI: Acceleration of feature selection on heterogeneous systems

  • González-Domínguez, Jorge
  • Expósito, Roberto R.
  • Bolón-Canedo, Verónica
©2019 Elsevier B.V. All rights reserved. This manuscript version is made available under the CC-BY-NC-ND 4.0 license https://creativecommons.org/licenses/bync-nd/4.0/. This version of the article has been accepted for publication in Future Generation Computer Systems. The Version of Record is available online at https://doi.org/10.1016/j.future.2019.08.031, Versión final aceptada de: J. González-Domínguez, R. R. Expósito, and V. Bolón-Canedo, "CUDA-JMI: Acceleration of feature selection on heterogeneous systemss", Future Generation Computer Systems, Vol. 102, pp. 426-436, Jan. 2020, https://doi.org/10.1016/j.future.2019.08.031, [Abstract]: Feature selection is a crucial step nowadays in machine learning and data analytics to remove irrelevant and redundant characteristics and thus to provide fast and reliable analyses. Many research works have focused on developing new methods that increase the global relevance of the subset of selected features while reducing the redundancy of information. However, those methods that select features with high relevance and low redundancy are extremely time-consuming when processing large datasets. In this work we present CUDA-JMI, a tool based on Joint Mutual Information that accelerates feature selection by exploiting the computational capabilities of modern heterogeneous systems that contain several CPU cores and GPU devices. The experimental evaluation has been carried out in three systems with different type and amount of CPUs and GPUs using five publicly available datasets from different fields. These results show that CUDA-JMI is significantly faster than its original sequential counterpart for all systems and input datasets. For instance, the runtime of CUDA-JMI is up to 52 times faster than an existing sequential JMI-based implementation in a machine with 24 CPU cores and two NVIDIA M60 boards (four GPUs). CUDA-JMI is publicly available to download from https://sourceforge.net/projects/cuda-jmi, This research has been partially funded by projects TIN2016-75845-P and TIN-2015-65069-C2-1-R of the Ministry of Economy, Industry and Competitiveness of Spain, as well as by Xunta de Galicia, Spain projects ED431D R2016/045, ED431G/01 and GRC2014/035, all of them partially funded by FEDER, Spain funds of the European Union., Xunta de Galicia; ED431D R2016/045, Xunta de Galicia; ED431G/01, Xunta de Galicia; GRC2014/035
RUC. Repositorio da Universidade da Coruña
Publication Contribución de congreso (ConferenceOutput)

Adaptive Real-Time Method for Anomaly Detection Using Machine Learning

  • Novoa-Paradela, David
  • Fontenla-Romero, Óscar
  • Guijarro-Berdiñas, Bertha
[Abstract] Anomaly detection is a sub-area of machine learning that deals with the development of methods to distinguish among normal and anomalous data. Due to the frequent use of anomaly-detection systems in monitoring and the lack of methods capable of learning in real time, this research presents a new method that provides such online adaptability. The method bases its operation on the properties of scaled convex hulls. It begins building a convex hull, using a minimum set of data, that is adapted and subdivided along time to accurately fit the boundary of the normal class data. The model has online learning ability and its execution can be carried out in a distributed and parallel way, all of them interesting advantages when dealing with big datasets. The method has been compared to other state-of-the-art algorithms demonstrating its effectiveness., This work has been supported by Spanish Government’s Secretaría de Estado de Investigación (Grant TIN2015-65069-C2-1-R), Xunta de Galicia (Grants ED431C 2018/34 and ED431G/01) and EU FEDER funds., Xunta de Galicia; ED431C 2018/34, Xunta de Galicia; ED431G/01
RUC. Repositorio da Universidade da Coruña
Publication Contribución de congreso (ConferenceOutput)

A delayed elastic-net approach for performing adversarial attacks

  • Cancela, Brais
  • Bolón-Canedo, Verónica
  • Alonso-Betanzos, Amparo
Presented at: 5th International Conference on Pattern Recognition, ICPR 2020, Virtual, Milan, 10-15 January 2021, This version of the paper has been accepted for publication. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. The final published paper is available online at: https://doi.org/10.1109/ICPR48806.2021.9413170, [Abstract]: With the rise of the so-called Adversarial Attacks, there is an increased concern on model security. In this paper we present two different contributions: novel measures of robustness (based on adversarial attacks) and a novel adversarial attack. The key idea behind these metrics is to obtain a measure that could compare different architectures, with independence of how the input is preprocessed (robustness against different input sizes and value ranges). To do so, a novel adversarial attack is presented, performing a delayed elastic-net adversarial attack (constraints are only used whenever a successful adversarial attack is obtained). Experimental results show that our approach obtains state-of-the-art adversarial samples, in terms of minimal perturbation distance. Finally, a benchmark of ImageNet pretrained models is used to conduct experiments aiming to shed some light about which model should be selected whenever security is a role factor., This research has been financially supported in part by European Union ERDF funds, by the Spanish Ministerio de Economía y Competitividad (research project TIN2015-65069-C2), and by the Xunta de Galicia (research projects GRC2014/035 and ED431G/01). We also gratefully acknowledge the support of NVIDIA Corporation with the donation of the Titan Xp GPU used for this research. Brais Cancela acknowledges the support of the Xunta de Galicia under its postdoctoral program., Xunta de Galicia; GRC2014/035, Xunta de Galicia; ED431G/01
RUC. Repositorio da Universidade da Coruña
Publication Artículo científico (JournalArticle)

Distributed correlation-based feature selection in spark

  • Palma Mendoza, Raúl José
  • Marcos, Luis de
  • Rodríguez, Daniel
  • Alonso-Betanzos, Amparo
© 2019. This manuscript version is made available under the CC-BY-NC-ND 4.0 license https://creativecommons.org/licenses/by-nc-nd/4.0/. This version of the article "R.-J. Palma-Mendoza, L. de-Marcos, D. Rodriguez, y A. Alonso-Betanzos, «Distributed correlation-based feature selection in spark», Information Sciences, vol. 496, pp. 287-299, sep. 2019" has been accepted for publication in Information Sciences. The Version of Record is available online at doi: 10.1016/j.ins.2018.10.052., [Abstract]: Feature selection (FS) is a key preprocessing step in data mining. CFS (Correlation-Based Feature Selection) is an FS algorithm that has been successfully applied to classification problems in many domains. We describe Distributed CFS (DiCFS) as a completely redesigned, scalable, parallel and distributed version of the CFS algorithm, capable of dealing with the large volumes of data typical of big data applications. Two versions of the algorithm were implemented and compared using the Apache Spark cluster computing model, currently gaining popularity due to its much faster processing times than Hadoop’s MapReduce model. We tested our algorithms on four publicly available datasets, each consisting of a large number of instances and two also consisting of a large number of features. The results show that our algorithms were superior in terms of both time-efficiency and scalability. In leveraging a computer cluster, they were able to handle larger datasets than the non-distributed WEKA version while maintaining the quality of the results, i.e., exactly the same features were returned by our algorithms when compared to the original algorithm available in WEKA., The authors thank CESGA for use of their supercomputing resources. This research has been partially supported by the Spanish Ministerio de Economía y Competitividad (research projects TIN 2015-65069-C2-1R, TIN2016-76956-C3-3-R), the Xunta de Galicia (Grants GRC2014/035 and ED431G/01) and the European Union Regional Development Funds. R. Palma-Mendoza holds a scholarship from the Spanish Fundación Carolina and the National Autonomous University of Honduras., Xunta de Galicia; GRC2014/035, Xunta de Galicia; ED431G/01
RUC. Repositorio da Universidade da Coruña
Publication Artículo científico (JournalArticle)

Insights into distributed feature ranking

  • Bolón-Canedo, Verónica
  • Sechidis, Konstantinos
  • Sánchez-Maroño, Noelia
  • Alonso-Betanzos, Amparo
  • Brown, Gavin
This version of the article: Bolón-Canedo, V., Sechidis, K., Sánchez-Maroño, N., Alonso-Betanzos, A., & Brown, G. (2019). ‘Insights into distributed feature ranking’ has been accepted for publication in: Information Sciences, 496, 378–398. The Version of Record is available online at https://doi.org/10.1016/j.ins.2018.09.045., [Abstract]: In an era in which the volume and complexity of datasets is continuously growing, feature selection techniques have become indispensable to extract useful information from huge amounts of data. However, existing algorithms may not scale well when dealing with huge datasets, and a possible solution is to distribute the data in several nodes. In this work we explore the different ways of distributing the data (by features and by samples) and we evaluate to what extent it is possible to obtain similar results as those obtained with the whole dataset. Trying to deal with the challenge of distributing the feature ranking process, we have performed experiments with different aggregation methods and feature rankers, and also evaluated the effect of distributing the feature ranking process in the subsequent classification performance., This research has been economically supported in part by the Spanish Ministerio de Economía y Competitividad and FEDER funds of the European Union through the research project TIN2015-65069-C2-1-R; and by the Consellería de Industria of the Xunta de Galicia through the research project GRC2014/035. Financial support from the Xunta de Galicia (Centro singular de investigación de Galicia accreditation 2016-2019) and the European Union (European Regional Development Fund - ERDF), is gratefully acknowledged (research project ED431G/01). V. Bolón-Canedo acknowledges support of the Xunta de Galicia under postdoctoral Grant code ED481B 2014/164-0., Xunta de Galicia; GRC2014/035, Xunta de Galicia; ED431G/01, Xunta de Galicia; ED481B 2014/164-0
RUC. Repositorio da Universidade da Coruña
Publication Artículo científico (JournalArticle)

Feature selection with limited bit depth mutual information for portable embedded systems

  • Morán-Fernández, Laura
  • Sechidis, Konstantinos
  • Bolón-Canedo, Verónica
  • Alonso-Betanzos, Amparo
  • Brown, Gavin
This version of the article: Morán-Fernández, L., Sechidis, K., Bolón-Canedo, V., Alonso-Betanzos, A., & Brown, G. (2020). ‘Feature selection with limited bit depth mutual information for portable embedded systems’ has been accepted for publication in: Knowledge-Based Systems, 197, 105885. The Version of Record is available online at https://doi.org/10.1016/j.knosys.2020.105885., [Abstract]: Since wearable computing systems have grown in importance in the last years, there is an increased interest in implementing machine learning algorithms with reduced precision parameters/computations. Not only learning, also feature selection, most of the times a mandatory preprocessing step in machine learning, is often constrained by the available computational resources. This work considers mutual information – one of the most common measures of dependence used in feature selection algorithms – with a limited number of bits. In order to test the procedure designed, we have implemented it in several well-known feature selection algorithms. Experimental results over several synthetic and real datasets demonstrate that low bit representations are sufficient to achieve performances close to that of double precision parameters and thus open the door for the use of feature selection in embedded platforms that minimize the energy consumption and carbon emissions., This research has been financially supported in part by the Spanish Ministerio de Economía y Competitividad (research project TIN2015-65069-C2-1-R), by European Union FEDER funds and by the Consellería de Industria of the Xunta de Galicia (research project GRC2014 /035). Financial sup-port from the Xunta de Galicia (Centro singular de investigación de Galicia accreditation 2016-2019) and the European Union (European Regional Development Fund - ERDF), is gratefully acknowledged (research project ED431G/01). Project supported by a 2018 Leonardo Grant for Researchers and Cultural Creators, BBVA Foundation. Laura Morán-Fernández acknowledges predoctoral stay grant by INDITEX-UDC 2015., Xunta de Galicia; ED431G/01, Xunta de Galicia; GRC2014 /035
RUC. Repositorio da Universidade da Coruña
Publication Artículo científico (JournalArticle)

Large scale anomaly detection in mixed numerical and categorical input spaces

  • Eiras-Franco, Carlos
  • Martínez Rego, David
  • Guijarro-Berdiñas, Bertha
  • Alonso-Betanzos, Amparo
  • Bahamonde, Antonio
© 2019. This manuscript version is made available under the CC-BY-NC-ND 4.0 license https://creativecommons.org/licenses/by-nc-nd/4.0/. This version of the article Eiras-Franco, C., Martínez-Rego, D., Guijarro-Berdiñas, B., Alonso-Betanzos, A., Bahamonde, A. (2019) ‘Large scale anomaly detection in mixed numerical and categorical input spaces’ has been accepted for publication in: Information Sciences, 487, pp. 115-127. The Version of Record is available online at https://doi.org/10.1016/j.ins.2019.03.013., [Abstract]: This work presents the ADMNC method, designed to tackle anomaly detection for large-scale problems with a mixture of categorical and numerical input variables. A flexible parametric probability measure is adjusted to input data, allowing low likelihood values to be tracked as anomalies. The main contribution of this method is that, to cope with the variable nature of the variables, we factorize the joint probability measure into two parts, namely, the marginal density of the continuous variables and the conditional probability of the categorical variables given the continuous part of the feature vector. The result is a model trained through a maximum likelihood objective function optimized with stochastic gradient descent that yields an effective and scalable algorithm. Compared with other well-known anomaly detection algorithms over several datasets, ADMNC is observed to both offer top level accuracy in datasets that are out of reach for the most effective existing methods and to scale up well to processing very large datasets. This makes it a powerful tool for solving a problem growing in popularity that currently lacks suitable scalable algorithms., This research has been financially supported in part by the Spanish Ministerio de EconomÍa y Competitividad (research projects TIN 2015-65069-C2, both 1-R and 2-R), by the Xunta de Galicia (Grants GRC2014/035 and ED431G/01) and the European Union Regional Development Funds., Xunta de Galicia; GRC2014/035, Xunta de Galicia; ED431G/01
RUC. Repositorio da Universidade da Coruña
Publication Artículo científico (JournalArticle)

A scalable decision-tree-based method to explain interactions in dyadic data

  • Eiras-Franco, Carlos
  • Guijarro-Berdiñas, Bertha
  • Alonso-Betanzos, Amparo
  • Bahamonde, Antonio
This version of the article: Eiras-Franco, C., Guijarro-Berdiñas, B., Alonso-Betanzos, A., & Bahamonde, A. (2019). ‘A scalable decision-tree-based method to explain interactions in dyadic data’ has been accepted for publication in: Decision Support Systems, 127, 113141. The Version of Record is available online at https://doi.org/10.1016/j.dss.2019.113141., [Abstract]: Gaining relevant insight from a dyadic dataset, which describes interactions between two entities, is an open problem that has sparked the interest of researchers and industry data scientists alike. However, the existing methods have poor explainability, a quality that is becoming essential in certain applications. We describe an explainable and scalable method that, operating on dyadic datasets, obtains an easily interpretable high-level summary of the relationship between entities. To do this, we propose a quality measure, which can be configured to a level that suits the user, that factors in the explainability of the model. We report experiments that confirm better results for the proposed method over alternatives, in terms of both explainability and accuracy. We also analyse the method's capacity to extract relevant actionable information and to handle large datasets., This research has been supported in part by the Spanish Ministerio de Economía y Competitividad (projects TIN 2015-65069-C2-1-R and 2-R), partially funded by FEDER funds of the EU and by the Xunta de Galicia (projects ED431C 2018/34 and Centro Singular de Investigación de Galicia, accreditation 2016-2019). The authors wish to thank the Fundación Pública Galega Centro Tecnolóxico de Supercomputación de Galicia (CESGA) for the use of their computing resources. We also thank Jorge Díez, Pablo Pérez and Oscar Luaces for providing the data needed to construct the (Outbrain DAY1) dataset and for their assistance in an early stage of this project., Xunta de Galicia; ED431C 2018/34
RUC. Repositorio da Universidade da Coruña
Publication Artículo científico (JournalArticle)

On the scalability of feature selection methods on high-dimensional data

  • Bolón-Canedo, Verónica
  • Rego-Fernández, Diego
  • Peteiro Barral, Diego
  • Alonso-Betanzos, Amparo
  • Guijarro-Berdiñas, Bertha
  • Sánchez-Maroño, Noelia
This version of the article has been accepted for publication, after peer review and is subject to Springer Nature’s AM terms of use, but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record is available online at: https://doi.org/10.1007/s10115-017-1140-3., [Abstract]: Lately, derived from the explosion of high dimensionality, researchers in machine learning became interested not only in accuracy, but also in scalability. Although scalability of learning methods is a trending issue, scalability of feature selection methods has not received the same amount of attention. This research analyzes the scalability of state-of-the-art feature selection methods, belonging to filter, embedded and wrapper approaches. For this purpose, several new measures are presented, based not only on accuracy but also on execution time and stability. The results on seven classical artificial datasets are presented and discussed, as well as two cases study analyzing the particularities of microarray data and the effect of redundancy. Trying to check whether the results can be generalized, we included some experiments with two real datasets. As expected, filters are the most scalable feature selection approach, being INTERACT, ReliefF and mRMR the most accurate methods., This research has been economically supported in part by the Ministerio de Economía y Competitividad of the Spanish Government (research project TIN2015-65069-C2-1-R), by European Union FEDER funds and by the Consellería de Industria of the Xunta de Galicia (research project GRC2014/035). Financial support from the Xunta de Galicia (Centro singular de investigación de Galicia accreditation 2016-2019) and the European Union (European Regional Development Fund - ERDF), is gratefully acknowledged (research project ED431G/01)., Xunta de Galicia; GRC2014/035, Xunta de Galicia; ED431G/01
RUC. Repositorio da Universidade da Coruña
Publication Artículo científico (JournalArticle)

Dealing with heterogeneity in the context of distributed feature selection for classification

  • Morillo-Salas, José Luis
  • Bolón-Canedo, Verónica
  • Alonso-Betanzos, Amparo
This version of the article has been accepted for publication, after peer review and is subject to Springer Nature’s AM terms of use, but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record is available online at: https://doi.org/10.1007/s10115-020-01526-4., [Abstract]: Advances in the information technologies have greatly contributed to the advent of larger datasets. These datasets often come from distributed sites, but even so, their large size usually means they cannot be handled in a centralized manner. A possible solution to this problem is to distribute the data over several processors and combine the different results. We propose a methodology to distribute feature selection processes based on selecting relevant and discarding irrelevant features. This preprocessing step is essential for current high-dimensional sets, since it allows the input dimension to be reduced. We pay particular attention to the problem of data imbalance, which occurs because the original dataset is unbalanced or because the dataset becomes unbalanced after data partitioning. Most works approach unbalanced scenarios by oversampling, while our proposal tests both over- and undersampling strategies. Experimental results demonstrate that our distributed approach to classification obtains comparable accuracy results to a centralized approach, while reducing computational time and efficiently dealing with data imbalance., This research has been financially supported in part by the Spanish Ministerio de Economía y Competitividad (research projects TIN2015-65069-C2-1-R and PID2019-109238GB-C22), by European Union FEDER funds and by the Consellería de Industria of the Xunta de Galicia (research project ED431C 2018/34). Financial support from the Xunta de Galicia (Centro singular de investigación de Galicia accreditation 2016–2019) and the European Union (European Regional Development Fund—ERDF), is gratefully acknowledged (research project ED431G 2019/01)., Xunta de Galicia; ED431C 2018/34, Xunta de Galicia; ED431G 2019/01
Academica-e. Repositorio Institucional de la Universidad Pública de Navarra
Publication Contribución de congreso (ConferenceOutput)

Análisis de redes sociales basado en las conquistas de César Borgia

  • Fumanal Idocin, Javier
  • Cordón, Óscar
  • Alonso Betanzos, Amparo
  • Bustince Sola, Humberto
  • Fernández Fernández, Francisco Javier
En este trabajo presentamos el modelado de redes sociales y detección de comunidades utilizando como base un evento histórico real, las conquistas de César Borgia en el siglo XV. Para ello, proponemos un nuevo conjunto de funciones, llamadas funciones de afinidad, disenadas para capturar la 'naturaleza de las interacciones locales entre cada par de actores en una red. Utilizando estas funciones, desarrollamos un nuevo algoritmo de detección de comunidades, el Borgia Clustering, donde las comunidades surgen naturalmente de un proceso de simulación de interacción de múltiples agentes en la red. También discutimos los efectos del tamaño y la escala de cada comunidad, y como pueden ser tomadas en cuenta en el proceso de simulación. Finalmente, comparamos nuestra detección de comunidades con otros algoritmos representativos, encontrando resultados favorables a nuestra propuesta., El trabajo de Javier Fumanal Idocin y Humberto Bustince ha sido financiado por el proyecto PID2019-108392GB-I00 (AEI/10.13039/ 501100011033). El trabajo de Oscar Cordón ha sido financiado por el Gobierno de España, EXASOCO (PGC2018-101216-B-I00), incluyendo fondos de desarrollo regional europeo (ERDF). La investigación de Amparo Alonso Betanzos ha sido parcialmente financiado por el Ministerio de Economía y Competitividad de España (TIN2015-65069-C2-1-R), por los fondos europeos FEDER y por la Consellería de Industria de la Xunta de Galicia (GRC2014 /035). El trabajo de María Minárová ha sido financiado por los proyectos APVV-17-0066, and APVV-18-0052.
Academica-e. Repositorio Institucional de la Universidad Pública de Navarra
Publication Artículo científico (JournalArticle)

Community detection and social network analysis based on the Italian wars of the 15th century

  • Fumanal Idocin, Javier
  • Alonso Betanzos, Amparo
  • Cordón, Óscar
  • Bustince Sola, Humberto
  • Minárová, María
In this contribution we study social network modelling by using human interaction as a basis. To do so, we propose a new set of functions, affinities, designed to capture the nature of the local interactions among each pair of actors in a network. By using these functions, we develop a new community detection algorithm, the Borgia Clustering, where communities naturally arise from the multi-agent interaction in the network. We also discuss the effects of size and scale for communities regarding this case, as well as how we cope with the additional complexity present when big communities arise. Finally, we compare our community detection solution with other representative algorithms, finding favourable results., Javier Fumanal Idocin’s and Humberto Bustince’s research has been supported by the project TIN2016-77356-P (AEI/FEDER,UE).Oscar Cordón’s research was supported by the Spanish Ministry of Science, Innovation and Universities under grant EX-ASOCO (PGC2018-101216-B-I00), including, European Regional Development Funds (ERDF). Amparo Alonso-Betanzos’ research has been financially supported in part by the Spanish Ministerio de Economía y Competitividad (research project TIN2015-65069-C2-1-R), by European Union FEDER funds and by the Consellería de Industria of the Xunta de Galicia, Spain (research project GRC2014 /035). M. Minárová’s research has been funded by the project work was supported by the project APVV-17-0066.