Research area:

Itzuli
003

ML

Machine Learning

Our goal is to develop novel and efficient machine learning algorithms able to deal with new data-related practical problems. We also pursue the mathematical modeling of these algorithms in order to provide theoretical guarantees of their performance.

 

In the machine learning research line we deal with data problems coming from different scenarios: industry, biosciences, health, economy, etc. We pursue the developments of new machine learning algorithms that can efficiently tackle these problems. Particularly, we consider problems that account for a variety of data types: from time series, to steaming data or images and speech, and a wide range of modelization techniques and mathematical formalisms such as: probabilistic graphical models, Bayesian approaches, deep learning, etc.

 

The research carried out in the machine learning line is inspired in problems that appear in other scientific, technological or economical disciplines. We develop new machine learning methods and algorithms related with the main data analysis activities such as clustering, supervised classification, feature subset selection, etc. to solve this kind of problems. Based on the specific characteristics of the problem at hand, we design tailored but general algorithms that extract as much information as possible from the available data providing efficient machine learning models that solve the problem.

In addition to that, we also develop mathematical tools able to model the behavior and performance of the algorithms: studying their convergence, the estimation of the performance, the behavior of the algorithms in terms of computational time and memory requirements, etc.

During the last years the machine learning line has worked with different machine learning problems and algorithms. Particularly, we can emphasize the work done in the area of time series mining and data streaming, the adaptation of classical clustering algorithms such as k-means or k-medoids to massive data environments, the probabilistic modeling of permutations and ranked data or the developments in anomaly detection, and the analysis of crowd learning environments.

In terms of formalisms, we strongly rely on probabilistic modeling, using different tools and techniques such as probabilistic graphical models and Gaussian process to name, which in most cases are learned under a Bayesian perspective. We also pursue the use of deep learning when we consider it the most appropriate technique for the problem at hand.

 

ML

Efficient Meta-Heuristics for Spacecraft Trajectory Optimization

Name: Shirazi, Abolfazl
Thesis advisor(s): J.A. Lozano
University: Euskal Herriko Unibertsitatea (UPV/EHU)
ML

Algorithms for large orienteering problems

Name: kobeaga, Gorka
Thesis advisor(s): J.A. Lozano
University: Euskal Herriko Unibertsitatea (UPV/EHU)
ML

Advances on time series analysis using elastic measures of similarity

Name: Oregui, Izaskun
Thesis advisor(s): A. Perez
University: Euskal Herriko Unibertsitatea (UPV/EHU)
ML

K-means for massive data

Name: Capó, Marco
Thesis advisor(s): Jose Antonio Lozano eta Aritz Pérez
University: Euskal Herriko Unibertsitatea (UPV/EHU)
ML

Theoretical and Methodological Advances in semi-supervised learning and the class-imbalance problem

Name: Ortigosa Hernández, Jonathan
Thesis advisor(s): Lozano
University: Euskal Herriko Unibertsitatea (UPV/EHU)

KmeansLandscape

Study the k-means problem from a local optimization perspective

Authors: Aritz Pérez

License: free and open source software

Placement

Local

PGM

Procedures for learning probabilistic graphical models

Authors: Aritz Pérez

License: free and open source software

Placement

Local

On-line Elastic Similarity Measures

Adaptation of the most frequantly used elastic similarity measures: Dynamic Time Warping (DTW), Edit Distance (Edit), Edit Distance for Real Sequences (EDR) and Edit Distance with Real Penalty (ERP) to on-line setting.

Authors: Izaskun Oregi, Aritz Perez, Javier Del Ser, Jose A. Lozano

License: free and open source software

Probabilistic load forecasting based on adaptive online learning

This repository contains code for the paper Probabilistic Load Forecasting based on Adaptive Online Learning

Authors: Veronica Alvarez

License: free and open source software

MRCpy: a library for Minimax Risk Classifiers 

MRCpy library implements minimax risk classifiers (MRCs) that are based on robust risk minimization and can utilize 0-1-loss.

Authors: Kartheek Reddy, Claudia Guerrero, Aritz Perez, Santiago Mazuelas

License: free and open source software

Minimax Classification under Concept Drift with Multidimensional Adaptation and Performance Guarantees (AMRC)

The proposed AMRCs account for multivariate and high-order time changes, provide performance guarantees at specific time instants, and efficiently update classification rules.

Authors: Veronica Alvarez

License: free and open source software

Efficient learning algorithm for Minimax Risk Classifiers (MRCs) in high dimensions

This repository provides efficient learning algorithm for Minimax Risk Classifiers (MRCs) in high dimensions. The presented algorithm utilizes the constraint generation approach for the MRC linear program. 

Authors: Kartheek Reddy

License: free and open source software

efficient learning algorithm for Minimax Risk Classifiers (MRCs) in high dimensions

This repository provides efficient learning algorithm for Minimax Risk Classifiers (MRCs) in covariate shift framework.

Authors: Jose Segovia

License: free and open source software

BayesianTree

Approximating probability distributions with mixtures of decomposable models

Authors: Aritz Pérez

License: free and open source software

Placement

Local

MixtureDecModels

Learning mixture of decomposable models with hidden variables

Authors: Aritz Pérez

License: free and open source software

Placement

Local

FractalTree

Implementation of the procedures presented in A. Pérez, I. Inza and J.A. Lozano (2016). Efficient approximation of probability distributions with k-order decomposable models. International Journal of Approximate Reasoning 74, 58-87.

Authors: Aritz Pérez

License: free and open source software

Minimax Forward and Backward Learning of Evolving Tasks with Performance Guarantees

This repository is the official implementation of Minimax Forward and Backward Learning of Evolving Tasks with Performance Guarantees.

Authors: Veronica Alvarez

License: free and open source software