Jesus Cid-sueiro - Profile on Academia.edu (original) (raw)

Papers by Jesus Cid-sueiro

2022 IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events (PerCom Workshops)

Many ways of annotating a dataset for machine learning classification tasks that go beyond the us... more Many ways of annotating a dataset for machine learning classification tasks that go beyond the usual class labels exist in practice. These are of interest as they can simplify or facilitate the collection of annotations, while not greatly affecting the resulting machine learning model. Many of these fall under the umbrella term of weak labels or annotations. However, it is not always clear how different alternatives are related. In this paper we propose a framework for categorising weak supervision settings with the aim of: (1) helping the dataset owner or annotator navigate through the available options within weak supervision when prescribing an annotation process, and (2) describing existing annotations for a dataset to machine learning practitioners so that we allow them to understand the implications for the learning process. To this end, we identify the key elements that characterise weak supervision and devise a series of dimensions that categorise most of the existing approaches. We show how common settings in the literature fit within the framework and discuss its possible uses in practice.

Publication in the conference proceedings of EUSIPCO, Trieste, Italy, 1996

This Letter discusses the application of gradient-based methods to train a single layer perceptro... more This Letter discusses the application of gradient-based methods to train a single layer perceptron subject to the constraint that the saturation degree of the sigmoid activation function (measured as its maximum slope in the sample space) is ¢xed to a given value. From a theoretical standpoint, we show that, if the training set is not linearly separable, the minimization of an L p error norm provides an approximation to the minimum error classi¢er, provided that the perceptron is highly saturated. Moreover, if data are linearly separable, the perceptron approximates the maximum margin classi¢er.

Content Understanding for Smart Multimedia Systems

EUROCON 2005 - The International Conference on "Computer as a Tool", 2005

Abstract The paper is presenting a core idea of the research line, dedicated to tools and utiliti... more Abstract The paper is presenting a core idea of the research line, dedicated to tools and utilities for a more friendly information access. In order to make future multimedia systems smarter, new mechanisms for high-level data and user understanding need to be embedded in multimedia communication systems

The IEEE Region 8 EUROCON 2003. Computer as a Tool.

Fuzzy student model in InterMediActor platform

International Conference on Information Technology Interfaces, 2004

The paper deals with personalization of navigation in the educational content, introduced in a co... more The paper deals with personalization of navigation in the educational content, introduced in a competence-based instructional design system InterMediActor. The system constructs an individualized navigation graph for each student and thus suggests the learning objectives the student is most prepared to attain. The navigation tools rely on the graph of dependencies between competences, and the student model. We use fuzzy

Optimal Double Route Search in a Telecommunication Network: A Telecontrol Network Application

IEEE Transactions on Reliability, 2011

High availability communication networks with very low failure rates are often designed by using ... more High availability communication networks with very low failure rates are often designed by using physical diversity, i.e., the traffic between a given pair of nodes is routed by using several physically disjoint paths. The selection of the pair of routes that maximizes the connectivity of a node is not an easy problem, be- causesuchconnectivitycannotbeexpressedasanadditivefunction of the availability of links and

Se propone, al amparo del proyecto “Nuevos Algoritmos para la Gestión Eficiente de Contenidos Mul... more Se propone, al amparo del proyecto “Nuevos Algoritmos para la Gestión Eficiente de Contenidos Multimedia en Redes de Comunicaciones Móviles”(NAGEC), un nuevo mecanismo para la búsqueda y recuperación de imágenes basado en realimentación de relevancia. La arquitectura propuesta se compone de una red neuronal y un tesauro. La red neuronal extrae de las imágenes dos parámetros: textura y color. El tesauro recoge las relaciones semánticas existentes entre los términos descriptores de las imágenes de la ...

WSEAS Transactions on Computers. 2004; 1 (3): 115-121, 2004

This paper describes a failure alert system and a methodology for content reuse in a new instruct... more This paper describes a failure alert system and a methodology for content reuse in a new instructional design system called InterMediActor (IMA). IMA provides an environment for instructional content design, production and reuse, and for students' evaluation based in content specification through a hierarchical structure of competences. The student assessment process and information extraction process for content reuse are explained.

Computer based training or distance education are facing dramatic changes with the advent of stan... more Computer based training or distance education are facing dramatic changes with the advent of standardization efforts, some of them concentrating in maximal reuse. This is of paramount importance for a sustainable-cost affordable-production of educational materials. Reuse in itself should not be a goal, though, since many methodological aspects might be lost. In this paper we propose two content production approaches for the InterMediActor platform under a competence-based methodology: either a bottom-up approach where ...

Improving Classification under Changes in Class and Within-Class Distributions

Lecture Notes in Computer Science, 2009

Abstract. The fundamental assumption that training and operational data come from the same probab... more Abstract. The fundamental assumption that training and operational data come from the same probability distribution, which is the basis of most learning algorithms, is often not satisfied in practice. Several algorithms have been proposed to cope with classification problems where the class priors may change after training, but they can show a poor performance when the class conditional data densities also change. In this paper, we propose a re-estimation algorithm that makes use of unlabeled operational data to adapt ...

Improving Conventional Equalizers with Neural Networks

Proceedings of the International Workshop on Applications of Neural Networks to Telecommunications

Abstract In this paper we show that the low detection capabilities of conventional equalizers (li... more Abstract In this paper we show that the low detection capabilities of conventional equalizers (linear and decision feedback equalizers) and the excessive complexity of those based on neural networks can be avoided by means of mixed schemes. Linear equalizers aided by controid selection (as in Radial Basis Function networks) improve performance over the standard linear FIR equalization approach, and modified DFE based on bi-layer perceptron avoids error propagation, outperforming conventional schemes. 1

Recycling weak labels for multiclass classification

Neurocomputing

Abstract This paper explores the mechanisms to efficiently combine annotations of different quali... more Abstract This paper explores the mechanisms to efficiently combine annotations of different quality for multiclass classification datasets, as we argue that it is easier to obtain large collections of weak labels as opposed to true labels. Since labels come from different sources, their annotations may have different degrees of reliability (e.g., noisy labels, supersets of labels, complementary labels or annotations performed by domain experts), and we must make sure that the addition of potentially inaccurate labels does not degrade the performance achieved when using only true labels. For this reason, we consider each group of annotations as being weakly supervised and pose the problem as finding the optimal combination of such collections. We propose an efficient algorithm based on expectation-maximization and show its performance in both synthetic and real-world classification tasks in a variety of weak label scenarios.

IEEE Transactions on Circuits and Systems for Video Technology

In this work we describe and optimize a general scheme based on lifting transforms on graphs for ... more In this work we describe and optimize a general scheme based on lifting transforms on graphs for video coding. A graph is constructed to represent the video signal. Each pixel becomes a node in the graph and links between nodes represent similarity between them. Therefore, spatial neighbors and temporal motion-related pixels can be linked, while nonsimilar pixels (e.g., pixels across an edge) may not be. Then, a lifting-based transform, in which filterin operations are performed using linked nodes, is applied to this graph, leading to a 3-dimensional (spatio-temporal) directional transform which can be viewed as an extension of wavelet transforms for video. The design of the proposed scheme requires four main steps: (i) graph construction, (ii) graph splitting, (iii) filte design, and (iv) extension of the transform to different levels of decomposition. We focus on the optimization of these steps in order to obtain an effective transform for video coding. Furthermore, based on this scheme, we propose a coefficien reordering method and an entropy coder leading to a complete video encoder that achieves better coding performance than a motion compensated temporal filterin wavelet-based encoder and a simple encoder derived from H.264/AVC that makes use of similar tools as our proposed encoder (reference software JM15.1 configu ed to use 1 reference frame, no subpixel motion estimation, 16 × 16 inter and 4 × 4 intra modes).

2015 IEEE 6th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP), 2015

In this paper, we analyze energy-harvesting adaptive diffusion networks for a distributed estimat... more In this paper, we analyze energy-harvesting adaptive diffusion networks for a distributed estimation problem. In order to wisely manage the available energy resources, we propose a scheme where a censoring algorithm is jointly applied over the diffusion strategy. An energy-aware variation of a diffusion algorithm is used, and a new way of measuring the relevance of the estimates in diffusion networks is proposed in order to apply a subsequent censoring mechanism. Simulation results show the potential benefit of integrating censoring schemes in energy-constrained diffusion networks.

The usual way of testing neurd networks in equalization problems is carrying out computer simulat... more The usual way of testing neurd networks in equalization problems is carrying out computer simulations using high-level programming languages. At this level, neural equalizers have demonstrated their ability to reduce the error rates of conventional detectors. However, the red implementation of this kind of schemes state many problems that may be critical t o many of the neural structures proposed in the literature: quantization effects, real-time processing, limited training sequence, time varying channels, and so on. In this paper we analyze the performance of soft-decision tree structures in a DSP processor, under conditions much nearer to a redistic environment.

Journal of Machine Learning Research

The design of a minimum risk classifier based on data usually stems from the stationarity assumpt... more The design of a minimum risk classifier based on data usually stems from the stationarity assumption that the conditions during training and test are the same: the misclassification costs assumed during training must be in agreement with real costs, and the same statistical process must have generated both training and test data. Unfortunately, in real world applications, these assumptions may not hold. This paper deals with the problem of training a classifier when prior probabilities cannot be reliably induced from training data. Some strategies based on optimizing the worst possible case (conventional minimax) have been proposed previously in the literature, but they may achieve a robust classification at the expense of a severe performance degradation. In this paper we propose a minimax regret (minimax deviation) approach, that seeks to minimize the maximum deviation from the performance of the optimal risk classifier. A neural-based minimax regret classifier for general multi-class decision problems is presented. Experimental results show its robustness and the advantages in relation to other approaches.

Exchanging profiles to connect peers in distributed systems

Designing for Emergence in Multi-Agent Systems