Pasar al contenido principal

ALGORITMOS ESCALABLES DE APRENDIZAJE COMPUTACIONAL: MAS ALLA DE LA CLASIFICACION Y LA REGRESION

TIN2015-65069-C2-2-R

•
Nombre agencia financiadora Ministerio de Economía y Competitividad
Acrónimo agencia financiadora MINECO
Programa Programa Estatal de I+D+i Orientado a los Retos de la Sociedad
Convocatoria Proyectos I+D+i Retos de Investigación
Año convocatoria 2015
Plan Estatal 2013-2016 Plan Estatal de Investigación Científica y Técnica y de Innovación
Unidad de gestión Dirección General de Investigación Científica y Técnica
Centro beneficiario UNIVERSIDAD DE OVIEDO
Identificador persistente http://dx.doi.org/10.13039/501100003329

Resultados relacionados

Resultados totales (Incluyendo duplicados): 6
Encontrada(s) 1 página(s)

RUC. Repositorio da Universidade da Coruña
Publicación 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
Publicación 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
Publicación 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
Publicación 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
Publicación 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
Publicación 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