2018
Ramírez, David; Santamaría, Ignacio; Vaerenbergh, Steven Van; Scharf, L L
An alternating optimization algorithm for two-channel factor analysis with common and uncommon factors Artículo en actas
En: Proc. Asilomar Conf. Signals, Syst. and Computers, Pacific Grove, USA, 2018.
Resumen | Enlaces | BibTeX | Etiquetas:
@inproceedings{RamirezSantamariaVaerenbergh-2018,
title = {An alternating optimization algorithm for two-channel factor analysis with common and uncommon factors},
author = {David Ram\'{i}rez and Ignacio Santamar\'{i}a and Steven Van Vaerenbergh and L L Scharf},
doi = {10.1109/ACSSC.2018.8645457},
year = {2018},
date = {2018-00-01},
booktitle = {Proc. Asilomar Conf. Signals, Syst. and Computers},
address = {Pacific Grove, USA},
abstract = {An alternating optimization algorithm is presented and analyzed for identifying low-rank signal components, known in factor analysis terminology as common factors, that are correlated across two multiple-input multiple-output (MIMO) channels. The additive noise model at each of the MIMO channels consists of white uncorrelated noises of unequal variances plus a low-rank structured interference that is not correlated across the two channels. The low-rank components at each channel represent uncommon or channel-specific factors.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
An alternating optimization algorithm is presented and analyzed for identifying low-rank signal components, known in factor analysis terminology as common factors, that are correlated across two multiple-input multiple-output (MIMO) channels. The additive noise model at each of the MIMO channels consists of white uncorrelated noises of unequal variances plus a low-rank structured interference that is not correlated across the two channels. The low-rank components at each channel represent uncommon or channel-specific factors.
Eguizabal, Alma; Schreier, Peter J; Ramírez, David
Model-order selection in statistical shape models Artículo en actas
En: Proc. IEEE Int. Work. Machine Learning for Signal Process., Aalborg, Denmark, 2018.
Resumen | Enlaces | BibTeX | Etiquetas:
@inproceedings{EguizabalSchreierRamirez-2018,
title = {Model-order selection in statistical shape models},
author = {Alma Eguizabal and Peter J Schreier and David Ram\'{i}rez},
doi = {10.1109/MLSP.2018.8516941},
year = {2018},
date = {2018-00-01},
booktitle = {Proc. IEEE Int. Work. Machine Learning for Signal Process.},
address = {Aalborg, Denmark},
abstract = {Statistical shape models enhance machine learning algorithms providing prior information about deformation. A Point Distribution Model (PDM) is a popular landmark-based statistical shape model for segmentation. It requires choosing a model order, which determines how much of the variation seen in the training data is accounted for by the PDM. A good choice of the model order depends on the number of training samples and the noise level in the training data set. Yet the most common approach for choosing the model order simply keeps a predetermined percentage of the total shape variation. In this paper, we present a technique for choosing the model order based on information-theoretic criteria, and we show empirical evidence that the model order chosen by this technique provides a good trade-off between over- and underfitting.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Statistical shape models enhance machine learning algorithms providing prior information about deformation. A Point Distribution Model (PDM) is a popular landmark-based statistical shape model for segmentation. It requires choosing a model order, which determines how much of the variation seen in the training data is accounted for by the PDM. A good choice of the model order depends on the number of training samples and the noise level in the training data set. Yet the most common approach for choosing the model order simply keeps a predetermined percentage of the total shape variation. In this paper, we present a technique for choosing the model order based on information-theoretic criteria, and we show empirical evidence that the model order chosen by this technique provides a good trade-off between over- and underfitting.
Santamaría, Ignacio; Ramírez, David; Scharf, L L
Subspace averaging for source enumeration in large arrays Artículo en actas
En: Proc. IEEE Work. Stat. Signal Process., Freiburg, Germany, 2018.
Resumen | Enlaces | BibTeX | Etiquetas:
@inproceedings{SantamariaRamirezScharf-2018,
title = {Subspace averaging for source enumeration in large arrays},
author = {Ignacio Santamar\'{i}a and David Ram\'{i}rez and L L Scharf},
doi = {10.1109/SSP.2018.8450837},
year = {2018},
date = {2018-00-01},
booktitle = {Proc. IEEE Work. Stat. Signal Process.},
address = {Freiburg, Germany},
abstract = {Subspace averaging is proposed and examined as a method of enumerating sources in large linear arrays, under conditions of low sample support. The key idea is to exploit shift invariance as a way of extracting many subspaces, which may then be approximated by a single extrinsic average. An automatic order determination rule for this extrinsic average is then the rule for determining the number of sources. Experimental results are presented for cases where the number of array snapshots is roughly half the number of array elements, and sources are well separated with respect to the Rayleigh limit.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Subspace averaging is proposed and examined as a method of enumerating sources in large linear arrays, under conditions of low sample support. The key idea is to exploit shift invariance as a way of extracting many subspaces, which may then be approximated by a single extrinsic average. An automatic order determination rule for this extrinsic average is then the rule for determining the number of sources. Experimental results are presented for cases where the number of array snapshots is roughly half the number of array elements, and sources are well separated with respect to the Rayleigh limit.
2017
Akyildiz, O. D.; Mariño, Inés P.; Míguez, Joaquín
Adaptive noisy importance sampling for stochastic optimization Artículo en actas
En: 2017 IEEE 7th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP), 2017.
Enlaces | BibTeX | Etiquetas: adaptive noisy importance sampler, Optimization
@inproceedings{JMiguez17c,
title = {Adaptive noisy importance sampling for stochastic optimization},
author = {O. D. Akyildiz and In\'{e}s P. Mari\~{n}o and Joaqu\'{i}n M\'{i}guez},
doi = {10.1109/CAMSAP.2017.8313215},
year = {2017},
date = {2017-12-10},
booktitle = {2017 IEEE 7th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP)},
keywords = {adaptive noisy importance sampler, Optimization},
pubstate = {published},
tppubtype = {inproceedings}
}
Horstmann, S; Ramírez, David; Schreier, Peter J
Detection of almost-cyclostationarity: An approach based on a multiple hypothesis test Artículo en actas
En: Proc. Asilomar Conf. Signals Syst. Computers, Pacific Grove, USA, 2017.
@inproceedings{HorstmannRamirezSchreier:2017,
title = {Detection of almost-cyclostationarity: An approach based on a multiple hypothesis test},
author = {S Horstmann and David Ram\'{i}rez and Peter J Schreier},
year = {2017},
date = {2017-10-01},
booktitle = {Proc. Asilomar Conf. Signals Syst. Computers},
address = {Pacific Grove, USA},
abstract = {This work presents a technique to detect whether a signal is almost cyclostationary (ACS) or wide-sense stationary (WSS). Commonly, ACS (and also CS) detectors require a priori knowledge of the cycle period, which in the ACS case is not an integer. To tackle the case of unknown cycle period, we propose an approach that combines a resampling technique, which handles the fractional part of the cycle period and allows the use of the generalized likelihood ratio test (GLRT), with a multiple hypothesis test, which handles the integer part of the cycle period. We control the probability of false alarm based on the known distribution of the individual GLRT statistic, results from order statistics, and the Holm multiple test procedure. To evaluate the performance of the proposed detector we consider a communications example, where simulation results show that the proposed technique outperforms state-of-the-art competitors.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
This work presents a technique to detect whether a signal is almost cyclostationary (ACS) or wide-sense stationary (WSS). Commonly, ACS (and also CS) detectors require a priori knowledge of the cycle period, which in the ACS case is not an integer. To tackle the case of unknown cycle period, we propose an approach that combines a resampling technique, which handles the fractional part of the cycle period and allows the use of the generalized likelihood ratio test (GLRT), with a multiple hypothesis test, which handles the integer part of the cycle period. We control the probability of false alarm based on the known distribution of the individual GLRT statistic, results from order statistics, and the Holm multiple test procedure. To evaluate the performance of the proposed detector we consider a communications example, where simulation results show that the proposed technique outperforms state-of-the-art competitors.
Míguez, Joaquín
On the performance of nonlinear importance samplers and population Monte Carlo schemes Artículo en actas
En: 2017 22nd International Conference on Digital Signal Processing (DSP), 2017, ISSN: 2165-3577.
Enlaces | BibTeX | Etiquetas: Monte Carlo algorithm
@inproceedings{JMiguez17b,
title = {On the performance of nonlinear importance samplers and population Monte Carlo schemes},
author = {Joaqu\'{i}n M\'{i}guez},
doi = {10.1109/ICDSP.2017.8096057},
issn = {2165-3577},
year = {2017},
date = {2017-08-23},
booktitle = {2017 22nd International Conference on Digital Signal Processing (DSP)},
keywords = {Monte Carlo algorithm},
pubstate = {published},
tppubtype = {inproceedings}
}
Liu, Yanfang; Olmos, Pablo M; Koch, Tobias
On LDPC code ensembles with generalized constraints Artículo en actas
En: 2017 IEEE International Symposium on Information Theory (ISIT), pp. 371-375, 2017.
@inproceedings{8006552,
title = {On LDPC code ensembles with generalized constraints},
author = {Yanfang Liu and Pablo M Olmos and Tobias Koch},
doi = {10.1109/ISIT.2017.8006552},
year = {2017},
date = {2017-06-01},
booktitle = {2017 IEEE International Symposium on Information Theory (ISIT)},
pages = {371-375},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Geiger, Bernhard C; Koch, Tobias
On the information dimension rate of stochastic processes Artículo en actas
En: 2017 IEEE International Symposium on Information Theory (ISIT), pp. 888-892, 2017.
@inproceedings{8006656,
title = {On the information dimension rate of stochastic processes},
author = {Bernhard C Geiger and Tobias Koch},
doi = {10.1109/ISIT.2017.8006656},
year = {2017},
date = {2017-06-01},
booktitle = {2017 IEEE International Symposium on Information Theory (ISIT)},
pages = {888-892},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Lancho, Alejandro; Koch, Tobias; Durisi, Giuseppe
A high-SNR normal approximation for single-antenna Rayleigh block-fading channels Artículo en actas
En: 2017 IEEE International Symposium on Information Theory (ISIT), pp. 1773-1777, 2017.
@inproceedings{8006834,
title = {A high-SNR normal approximation for single-antenna Rayleigh block-fading channels},
author = {Alejandro Lancho and Tobias Koch and Giuseppe Durisi},
doi = {10.1109/ISIT.2017.8006834},
year = {2017},
date = {2017-06-01},
booktitle = {2017 IEEE International Symposium on Information Theory (ISIT)},
pages = {1773-1777},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Ramírez, David; Marques, Antonio G; Segarra, Santiago
Graph-signal reconstruction and blind deconvolution for diffused sparse inputs Artículo en actas
En: Proc. IEEE Int. Conf. Acoustics, Speech and Signal Process., New Orleans, USA, 2017.
@inproceedings{RamirezMarquesSegarra:2017,
title = {Graph-signal reconstruction and blind deconvolution for diffused sparse inputs},
author = {David Ram\'{i}rez and Antonio G Marques and Santiago Segarra},
year = {2017},
date = {2017-03-01},
booktitle = {Proc. IEEE Int. Conf. Acoustics, Speech and Signal Process.},
address = {New Orleans, USA},
abstract = {This paper investigates the problems of signal reconstruction and blind deconvolution for graph signals that have been generated by an originally sparse input diffused through the network via the application of a graph filter operator. Assuming that the support of the sparse input signal is unknown, and that the diffused signal is observed only at a subset of nodes, we address the related problems of: 1) identifying the input and 2) interpolating the values of the diffused signal at the non-sampled nodes. We first consider the more tractable case where the coefficients of the diffusing graph filter are known and then address the problem of joint input and filter identification. The corresponding blind identification problems are formulated, novel convex relaxations are discussed, and modifications to incorporate a priori information on the sparse inputs are provided.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
This paper investigates the problems of signal reconstruction and blind deconvolution for graph signals that have been generated by an originally sparse input diffused through the network via the application of a graph filter operator. Assuming that the support of the sparse input signal is unknown, and that the diffused signal is observed only at a subset of nodes, we address the related problems of: 1) identifying the input and 2) interpolating the values of the diffused signal at the non-sampled nodes. We first consider the more tractable case where the coefficients of the diffusing graph filter are known and then address the problem of joint input and filter identification. The corresponding blind identification problems are formulated, novel convex relaxations are discussed, and modifications to incorporate a priori information on the sparse inputs are provided.
Ríos-Muñoz, Gonzalo; Hernandez, Pablo Ruiz M; Castellanos, Evaristo; Ávila, Pablo; Loughlin, Gerard; Fernandez-Aviles, Francisco; Artés-Rodríguez, Antonio; Arenal, Ángel
Substrate Characterization of Rotational Activity Sites in Persistent Atrial Fibrillation Patients Artículo en actas
En: CNIC Conference Atrial Fibrillation: From Mechanisms to Population Science, Madrid, 2017.
@inproceedings{RiosMunoz2017a,
title = {Substrate Characterization of Rotational Activity Sites in Persistent Atrial Fibrillation Patients},
author = {Gonzalo R\'{i}os-Mu\~{n}oz and Pablo Ruiz M Hernandez and Evaristo Castellanos and Pablo \'{A}vila and Gerard Loughlin and Francisco Fernandez-Aviles and Antonio Art\'{e}s-Rodr\'{i}guez and \'{A}ngel Arenal},
year = {2017},
date = {2017-01-01},
booktitle = {CNIC Conference Atrial Fibrillation: From Mechanisms to Population Science},
address = {Madrid},
abstract = {The underlying mechanisms initiating and sustaining atrial fibrillation (AF) are still under debate, and an optimal treatment for AF is not been well established. Spatiotemporal stable sources (rotors) have been proposed as maintenance mechanism of AF. The use of high density electroanatomical mapping with microelectrodes (HDEMM) and a novel rotational activity detection system we can detect rotors and characterize the tissue where rotors are located. The analysis shows evidence of voltage values related to rotational activity beyond bipolar voltage range 0.1-0.5 mV, classically considered for scar definitions. Functional assessment may add incremental value to invasive treatment of AF.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The underlying mechanisms initiating and sustaining atrial fibrillation (AF) are still under debate, and an optimal treatment for AF is not been well established. Spatiotemporal stable sources (rotors) have been proposed as maintenance mechanism of AF. The use of high density electroanatomical mapping with microelectrodes (HDEMM) and a novel rotational activity detection system we can detect rotors and characterize the tissue where rotors are located. The analysis shows evidence of voltage values related to rotational activity beyond bipolar voltage range 0.1-0.5 mV, classically considered for scar definitions. Functional assessment may add incremental value to invasive treatment of AF.
Ríos-Muñoz, Gonzalo; Hernandez, Pablo Ruiz M; Castellanos, Evaristo; Ávila, Pablo; Loughlin, Gerard; Fernandez-Aviles, Francisco; Artés-Rodríguez, Antonio; Arenal, Ángel
Presence and Voltage Characterization of Rotational Activity in Atrial Fibrillation Patients Artículo en actas
En: Atrial Signals 2017, Valencia, 2017.
@inproceedings{RiosMunoz2017b,
title = {Presence and Voltage Characterization of Rotational Activity in Atrial Fibrillation Patients},
author = {Gonzalo R\'{i}os-Mu\~{n}oz and Pablo Ruiz M Hernandez and Evaristo Castellanos and Pablo \'{A}vila and Gerard Loughlin and Francisco Fernandez-Aviles and Antonio Art\'{e}s-Rodr\'{i}guez and \'{A}ngel Arenal},
year = {2017},
date = {2017-01-01},
booktitle = {Atrial Signals 2017},
address = {Valencia},
abstract = {The underlying mechanisms initiating and sustaining atrial fibrillation (AF) are still under debate, and an optimal treatment for AF is not yet established. Spatiotemporal stable sources (rotors) have been proposed as maintenance mechanism of AF. Using high density electroanatomical mapping with microelectrodes (HDEMM) and a novel rotational activity detection system we are able to detect rotors and characterize the tissue where rotors are located. The analysis shows evidence of voltage values related to rotational activity beyond the bipolar voltage range 0.1-0.5 mV, classically considered for scar definitions. Functional assessment may add incremental value to invasive treatment of AF.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The underlying mechanisms initiating and sustaining atrial fibrillation (AF) are still under debate, and an optimal treatment for AF is not yet established. Spatiotemporal stable sources (rotors) have been proposed as maintenance mechanism of AF. Using high density electroanatomical mapping with microelectrodes (HDEMM) and a novel rotational activity detection system we are able to detect rotors and characterize the tissue where rotors are located. The analysis shows evidence of voltage values related to rotational activity beyond the bipolar voltage range 0.1-0.5 mV, classically considered for scar definitions. Functional assessment may add incremental value to invasive treatment of AF.
Hernandez, Pablo Ruiz M; Ríos-Muñoz, Gonzalo; Castellanos, Evaristo; Ávila, Pablo; Torrecilla, Esteban G; Loughlin, Gerard; Datino, Tomas; Atienza, Felipe; Fernandez-Aviles, Francisco; Artés-Rodríguez, Antonio; Arenal, Ángel
Presence and Distribution of Rotational Conduction Points and Its Association With Scar in Patients With Persistent Atrial Fibrillation Artículo en actas
En: Hear. Rhythm, pp. S235––S236, Elsevier 2017.
@inproceedings{Ruiz2017b,
title = {Presence and Distribution of Rotational Conduction Points and Its Association With Scar in Patients With Persistent Atrial Fibrillation},
author = {Pablo Ruiz M Hernandez and Gonzalo R\'{i}os-Mu\~{n}oz and Evaristo Castellanos and Pablo \'{A}vila and Esteban G Torrecilla and Gerard Loughlin and Tomas Datino and Felipe Atienza and Francisco Fernandez-Aviles and Antonio Art\'{e}s-Rodr\'{i}guez and \'{A}ngel Arenal},
year = {2017},
date = {2017-01-01},
booktitle = {Hear. Rhythm},
volume = {14},
number = {5},
pages = {S235----S236},
organization = {Elsevier},
abstract = {Background: Persistent AF has shown high post-ablation recurrence rates, and left atrial (LA) fibrosis is a possible cause. Extrapulmonary rotational activation (rotors), could be related to fibrosis. High density electroanatomical mapping with microelectrodes (HDEMM) can allow better LA tissue characterization and rotor identification. Objetive: To assess the presence and distribution of rotors, and their relationship with areas of LA scar in patients with persistent AF. Conclusion: Rotational conduction is observed, to some degree, in the majority of sites evaluated in patients with persistent AF. Gyre complexity was greater in areas of higher voltage, and does not seem to be associated with any particular LA location.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Background: Persistent AF has shown high post-ablation recurrence rates, and left atrial (LA) fibrosis is a possible cause. Extrapulmonary rotational activation (rotors), could be related to fibrosis. High density electroanatomical mapping with microelectrodes (HDEMM) can allow better LA tissue characterization and rotor identification. Objetive: To assess the presence and distribution of rotors, and their relationship with areas of LA scar in patients with persistent AF. Conclusion: Rotational conduction is observed, to some degree, in the majority of sites evaluated in patients with persistent AF. Gyre complexity was greater in areas of higher voltage, and does not seem to be associated with any particular LA location.
Lancho, Alejandro; Koch, Tobias; Durisi, Giuseppe
A high-SNR normal approximation for single-antenna Rayleigh block-fading channels Artículo en actas
En: 2017 IEEE International Symposium on Information Theory (ISIT), pp. 1773-1777, 2017.
@inproceedings{8006834b,
title = {A high-SNR normal approximation for single-antenna Rayleigh block-fading channels},
author = {Alejandro Lancho and Tobias Koch and Giuseppe Durisi},
doi = {10.1109/ISIT.2017.8006834},
year = {2017},
date = {2017-01-01},
urldate = {2017-01-01},
booktitle = {2017 IEEE International Symposium on Information Theory (ISIT)},
pages = {1773-1777},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
2016
Villacrés, Grace; Koch, Tobias
Wireless networks of bounded capacity Artículo en actas
En: 2016 IEEE International Symposium on Information Theory (ISIT), pp. 2584-2588, 2016.
@inproceedings{7541766,
title = {Wireless networks of bounded capacity},
author = {Grace Villacr\'{e}s and Tobias Koch},
doi = {10.1109/ISIT.2016.7541766},
year = {2016},
date = {2016-07-01},
booktitle = {2016 IEEE International Symposium on Information Theory (ISIT)},
pages = {2584-2588},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Koch, Tobias; Vazquez-Vilar, Gonzalo
A general rate-distortion converse bound for entropy-constrained scalar quantization Artículo en actas
En: 2016 IEEE International Symposium on Information Theory (ISIT), pp. 735-739, 2016.
@inproceedings{7541396,
title = {A general rate-distortion converse bound for entropy-constrained scalar quantization},
author = {Tobias Koch and Gonzalo Vazquez-Vilar},
doi = {10.1109/ISIT.2016.7541396},
year = {2016},
date = {2016-07-01},
booktitle = {2016 IEEE International Symposium on Information Theory (ISIT)},
pages = {735-739},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Vazquez-Vilar, Gonzalo
Multiple quantum hypothesis testing expressions and classical-quantum channel converse bounds Artículo en actas
En: 2016 IEEE International Symposium on Information Theory (ISIT 2016), Barcelona, Spain, 2016.
BibTeX | Etiquetas:
@inproceedings{gvazquez-isit2016a,
title = {Multiple quantum hypothesis testing expressions and classical-quantum channel converse bounds},
author = {Gonzalo Vazquez-Vilar},
year = {2016},
date = {2016-07-01},
booktitle = {2016 IEEE International Symposium on Information Theory (ISIT 2016)},
address = {Barcelona, Spain},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Pries, Aaron; Ramírez, David; Schreier, Peter J
Detection of Cyclostationarity in the Presence of Temporal or Spatial Structure with Applications to Cognitive Radio Artículo en actas
En: Proc. IEEE Int. Conf. Acoustics, Speech and Signal Process., (ICASSP 2016), Shanghai, 2016.
@inproceedings{Pries2016,
title = {Detection of Cyclostationarity in the Presence of Temporal or Spatial Structure with Applications to Cognitive Radio},
author = {Aaron Pries and David Ram\'{i}rez and Peter J Schreier},
url = {http://www.icassp2016.org/Papers/ViewPapers.asp?PaperNum=1789},
year = {2016},
date = {2016-03-01},
booktitle = {Proc. IEEE Int. Conf. Acoustics, Speech and Signal Process., (ICASSP 2016)},
address = {Shanghai},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Koch, Tobias
A Necessary and Sufficient Condition for the Asymptotic Tightness of the Shannon Lower Bound Artículo en actas
En: International Zurich Seminar on Communications, Zurich, 2016.
BibTeX | Etiquetas:
@inproceedings{Koch2016bb,
title = {A Necessary and Sufficient Condition for the Asymptotic Tightness of the Shannon Lower Bound},
author = {Tobias Koch},
year = {2016},
date = {2016-03-01},
booktitle = {International Zurich Seminar on Communications},
address = {Zurich},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Elvira, Victor; Miguez, Joaquin; Djuric, Petar M
Online Adaptation of the Number of Particles of Sequential Monte Carlo Methods Artículo en actas
En: IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2016), Shanghai, 2016.
BibTeX | Etiquetas:
@inproceedings{Elvira2016,
title = {Online Adaptation of the Number of Particles of Sequential Monte Carlo Methods},
author = {Victor Elvira and Joaquin Miguez and Petar M Djuric},
year = {2016},
date = {2016-03-01},
booktitle = {IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2016)},
address = {Shanghai},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Martino, Luca; Elvira, Victor; Luengo, David; Louzada, Francisco
Parallel Metropolis Chains with Cooperative Adaptation Artículo en actas
En: IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2016), Shanghai, 2016.
@inproceedings{Martino2016bb,
title = {Parallel Metropolis Chains with Cooperative Adaptation},
author = {Luca Martino and Victor Elvira and David Luengo and Francisco Louzada},
url = {http://www.icassp2016.org/Papers/ViewPapers.asp?PaperNum=3747},
year = {2016},
date = {2016-03-01},
booktitle = {IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2016)},
address = {Shanghai},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Vazquez-Vilar, Gonzalo; i Fàbregas, Albert Guillén; Verdú, Sergio
Hypothesis testing and quasi-perfect codes Artículo en actas
En: 2016 International Zurich Seminar on Communications (IZS 2016), Zurich, Switzerland, 2016.
BibTeX | Etiquetas:
@inproceedings{gvazquez-izs2016,
title = {Hypothesis testing and quasi-perfect codes},
author = {Gonzalo Vazquez-Vilar and Albert Guill\'{e}n i F\`{a}bregas and Sergio Verd\'{u}},
year = {2016},
date = {2016-03-01},
booktitle = {2016 International Zurich Seminar on Communications (IZS 2016)},
address = {Zurich, Switzerland},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Luengo, David; Ríos-Muñoz, Gonzalo; Elvira, Victor; Artés-Rodríguez, Antonio
A hierarchical algorithm for causality discovery among atrial fibrillation electrograms Artículo en actas
En: 2016 IEEE Int. Conf. Acoust. Speech Signal Process., pp. 774–778, IEEE, 2016, ISBN: 978-1-4799-9988-0.
Resumen | Enlaces | BibTeX | Etiquetas:
@inproceedings{Luengo2016b,
title = {A hierarchical algorithm for causality discovery among atrial fibrillation electrograms},
author = {David Luengo and Gonzalo R\'{i}os-Mu\~{n}oz and Victor Elvira and Antonio Art\'{e}s-Rodr\'{i}guez},
url = {http://ieeexplore.ieee.org/document/7471780/},
doi = {10.1109/ICASSP.2016.7471780},
isbn = {978-1-4799-9988-0},
year = {2016},
date = {2016-03-01},
booktitle = {2016 IEEE Int. Conf. Acoust. Speech Signal Process.},
pages = {774--778},
publisher = {IEEE},
abstract = {Multi-channel intracardiac electrocardiograms (electrograms) are sequentially acquired, at the electrophysiology laboratory, in order to guide radio frequency catheter ablation during heart surgery performed on patients with sustained atrial fibrillation (AF). These electrograms are used by cardiologists to determine candidate areas for ablation (e.g., areas corresponding to high dominant frequencies or complex fractionated electrograms). In this paper, we introduce a novel hierarchical algorithm for causality discovery among these multi-output sequentially acquired electrograms. The causal model obtained provides important information about the propagation of the electrical signals inside the heart, uncovering wavefronts and activation patterns that will serve to increase our knowledge about AF and guide cardiologists towards candidate areas for catheter ablation. Numerical results on synthetic signals, generated using the FitzHugh-Nagumo model, show the good performance of the proposed approach.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Multi-channel intracardiac electrocardiograms (electrograms) are sequentially acquired, at the electrophysiology laboratory, in order to guide radio frequency catheter ablation during heart surgery performed on patients with sustained atrial fibrillation (AF). These electrograms are used by cardiologists to determine candidate areas for ablation (e.g., areas corresponding to high dominant frequencies or complex fractionated electrograms). In this paper, we introduce a novel hierarchical algorithm for causality discovery among these multi-output sequentially acquired electrograms. The causal model obtained provides important information about the propagation of the electrical signals inside the heart, uncovering wavefronts and activation patterns that will serve to increase our knowledge about AF and guide cardiologists towards candidate areas for catheter ablation. Numerical results on synthetic signals, generated using the FitzHugh-Nagumo model, show the good performance of the proposed approach.
Castellanos, Evaristo; Hernandez, Pablo Ruiz M; Ríos-Muñoz, Gonzalo; Ávila, Pablo; Datino, Tomás; Atienza, Felipe; Fernández-Avilés, Francisco; Arenal, Ángel
Influencia del Ritmo en la Identificación de Islotes de Escara Auricular en Pacientes con FA Persistente sin Disfunción Ventricular Izquierda Detectada con Catéter de Mapeo Multielectrodo de 1mm Artículo en actas
En: SEC 2016 - El Congr. las Enfermedades Cardiovasc., 2016.
@inproceedings{Castellanos2016,
title = {Influencia del Ritmo en la Identificaci\'{o}n de Islotes de Escara Auricular en Pacientes con FA Persistente sin Disfunci\'{o}n Ventricular Izquierda Detectada con Cat\'{e}ter de Mapeo Multielectrodo de 1mm},
author = {Evaristo Castellanos and Pablo Ruiz M Hernandez and Gonzalo R\'{i}os-Mu\~{n}oz and Pablo \'{A}vila and Tom\'{a}s Datino and Felipe Atienza and Francisco Fern\'{a}ndez-Avil\'{e}s and \'{A}ngel Arenal},
year = {2016},
date = {2016-01-01},
booktitle = {SEC 2016 - El Congr. las Enfermedades Cardiovasc.},
number = {6002-38},
abstract = {En la fibrilaci\'{o}n auricular persistente (FA-Per), el aislamiento de las venas pulmonares (VVPP) presenta una mayor tasa de recidiva que en FA parox\'{i}stica. La FA-Per induce una remodelaci\'{o}n estructural caracterizada por fibrosis y formaci\'{o}n de tejido cicatricial en la aur\'{i}cula. La remodelaci\'{o}n estructural se asocia con una mayor tasa de recurrencia tras la ablaci\'{o}n. El mapeo electroanat\'{o}mico del tejido cicatricial no asociado a las VVPP podr\'{i}a facilitar la identificaci\'{o}n del sustrato espec\'{i}fico de la FA-Per. Los cat\'{e}teres de mapeo multielectrodo proporcionan una alta definici\'{o}n de tejido fibr\'{o}tico en pacientes con taquicardias auriculares. El objetivo fue evaluar y cuantificar la presencia de islotes de tejido cicatricial (durante el ritmo sinusal (RS) y FA) en los pacientes con FA-Per sin disfunci\'{o}n ventricular izquierda},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
En la fibrilación auricular persistente (FA-Per), el aislamiento de las venas pulmonares (VVPP) presenta una mayor tasa de recidiva que en FA paroxística. La FA-Per induce una remodelación estructural caracterizada por fibrosis y formación de tejido cicatricial en la aurícula. La remodelación estructural se asocia con una mayor tasa de recurrencia tras la ablación. El mapeo electroanatómico del tejido cicatricial no asociado a las VVPP podría facilitar la identificación del sustrato específico de la FA-Per. Los catéteres de mapeo multielectrodo proporcionan una alta definición de tejido fibrótico en pacientes con taquicardias auriculares. El objetivo fue evaluar y cuantificar la presencia de islotes de tejido cicatricial (durante el ritmo sinusal (RS) y FA) en los pacientes con FA-Per sin disfunción ventricular izquierda
2015
Valera, Isabel; Ruiz, Francisco J R; Svensson, Lennart; Perez-Cruz, Fernando
Infinite Factorial Dynamical Model Artículo en actas
En: Advances in Neural Information Processing Systems, pp. 1657–1665, Montreal, 2015.
Resumen | Enlaces | BibTeX | Etiquetas: CASI CAM CM, GAMMA-L+ UC3M
@inproceedings{Valera2015a,
title = {Infinite Factorial Dynamical Model},
author = {Isabel Valera and Francisco J R Ruiz and Lennart Svensson and Fernando Perez-Cruz},
url = {http://papers.nips.cc/paper/5667-infinite-factorial-dynamical-model},
year = {2015},
date = {2015-12-01},
booktitle = {Advances in Neural Information Processing Systems},
pages = {1657--1665},
address = {Montreal},
abstract = {We propose the infinite factorial dynamic model (iFDM), a general Bayesian nonparametric model for source separation. Our model builds on the Markov Indian buffet process to consider a potentially unbounded number of hidden Markov chains (sources) that evolve independently according to some dynamics, in which the state space can be either discrete or continuous. For posterior inference, we develop an algorithm based on particle Gibbs with ancestor sampling that can be efficiently applied to a wide range of source separation problems. We evaluate the performance of our iFDM on four well-known applications: multitarget tracking, cocktail party, power disaggregation, and multiuser detection. Our experimental results show that our approach for source separation does not only outperform previous approaches, but it can also handle problems that were computationally intractable for existing approaches.},
keywords = {CASI CAM CM, GAMMA-L+ UC3M},
pubstate = {published},
tppubtype = {inproceedings}
}
We propose the infinite factorial dynamic model (iFDM), a general Bayesian nonparametric model for source separation. Our model builds on the Markov Indian buffet process to consider a potentially unbounded number of hidden Markov chains (sources) that evolve independently according to some dynamics, in which the state space can be either discrete or continuous. For posterior inference, we develop an algorithm based on particle Gibbs with ancestor sampling that can be efficiently applied to a wide range of source separation problems. We evaluate the performance of our iFDM on four well-known applications: multitarget tracking, cocktail party, power disaggregation, and multiuser detection. Our experimental results show that our approach for source separation does not only outperform previous approaches, but it can also handle problems that were computationally intractable for existing approaches.
Luengo, David; Martino, Luca; Elvira, Victor; Bugallo, Monica F
Bias correction for distributed Bayesian estimators Artículo en actas
En: 2015 IEEE 6th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP), pp. 253–256, IEEE, Cancun, 2015, ISBN: 978-1-4799-1963-5.
Resumen | Enlaces | BibTeX | Etiquetas: Bayes methods, Big data, Distributed databases, Estimation, Probability density function, Wireless Sensor Networks
@inproceedings{Luengo2015a,
title = {Bias correction for distributed Bayesian estimators},
author = {David Luengo and Luca Martino and Victor Elvira and Monica F Bugallo},
url = {http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=7383784},
doi = {10.1109/CAMSAP.2015.7383784},
isbn = {978-1-4799-1963-5},
year = {2015},
date = {2015-12-01},
booktitle = {2015 IEEE 6th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP)},
pages = {253--256},
publisher = {IEEE},
address = {Cancun},
abstract = {Dealing with the whole dataset in big data estimation problems is usually unfeasible. A common solution then consists of dividing the data into several smaller sets, performing distributed Bayesian estimation and combining these partial estimates to obtain a global estimate. A major problem of this approach is the presence of a non-negligible bias in the partial estimators, due to the mismatch between the unknown true prior and the prior assumed in the estimation. A simple method to mitigate the effect of this bias is proposed in this paper. Essentially, the approach is based on using a reference data set to obtain a rough estimation of the parameter of interest, i.e., a reference parameter. This information is then communicated to the partial filters that handle the smaller data sets, which can thus use a refined prior centered around this parameter. Simulation results confirm the good performance of this scheme.},
keywords = {Bayes methods, Big data, Distributed databases, Estimation, Probability density function, Wireless Sensor Networks},
pubstate = {published},
tppubtype = {inproceedings}
}
Dealing with the whole dataset in big data estimation problems is usually unfeasible. A common solution then consists of dividing the data into several smaller sets, performing distributed Bayesian estimation and combining these partial estimates to obtain a global estimate. A major problem of this approach is the presence of a non-negligible bias in the partial estimators, due to the mismatch between the unknown true prior and the prior assumed in the estimation. A simple method to mitigate the effect of this bias is proposed in this paper. Essentially, the approach is based on using a reference data set to obtain a rough estimation of the parameter of interest, i.e., a reference parameter. This information is then communicated to the partial filters that handle the smaller data sets, which can thus use a refined prior centered around this parameter. Simulation results confirm the good performance of this scheme.
Dashti, Marzieh; Yiu, Simon; Yousefi, Siamak; Perez-Cruz, Fernando; Claussen, Holger
RSSI Localization with Gaussian Processes and Tracking Artículo en actas
En: IEEE Globecom, San Diego, 2015.
@inproceedings{Dashi2015,
title = {RSSI Localization with Gaussian Processes and Tracking},
author = {Marzieh Dashti and Simon Yiu and Siamak Yousefi and Fernando Perez-Cruz and Holger Claussen},
url = {http://globecom2015.ieee-globecom.org/},
year = {2015},
date = {2015-12-01},
booktitle = {IEEE Globecom},
address = {San Diego},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Elvira, Victor; Martino, Luca; Luengo, David; Bugallo, Monica F
On Sample Generation and Weight Calculation in Multiple Importance Sampling Artículo en actas
En: IEEE Conference on Signals, Systems, and Computers (ASILOMAR 2015), Pacific Groove, 2015.
Resumen | Enlaces | BibTeX | Etiquetas:
@inproceedings{Elvira2015b,
title = {On Sample Generation and Weight Calculation in Multiple Importance Sampling},
author = {Victor Elvira and Luca Martino and David Luengo and Monica F Bugallo},
url = {http://www.asilomarsscconf.org/webpage/asil15/Asilomar 2015 Book of Abstracts v005.pdf},
year = {2015},
date = {2015-11-01},
booktitle = {IEEE Conference on Signals, Systems, and Computers (ASILOMAR 2015)},
address = {Pacific Groove},
abstract = {We investigate various sampling and weight updating techniques, which are the two crucial steps of importance sampling. We discuss the standard mixture sampling that randomly draws samples from a set of proposals and the deterministic mixture sampling, where exactly one sample is drawn from each proposal. For weight calculation, we either compute the weights by considering the particular proposal used for each sample or by interpreting the proposal as a mixture formed by all available proposals. All combinations of sampling and weight calculation and some modifications that improve the performance and/or reduce the computational complexity are examined through computer simulations},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
We investigate various sampling and weight updating techniques, which are the two crucial steps of importance sampling. We discuss the standard mixture sampling that randomly draws samples from a set of proposals and the deterministic mixture sampling, where exactly one sample is drawn from each proposal. For weight calculation, we either compute the weights by considering the particular proposal used for each sample or by interpreting the proposal as a mixture formed by all available proposals. All combinations of sampling and weight calculation and some modifications that improve the performance and/or reduce the computational complexity are examined through computer simulations
Acer, Utku Gunay; Boran, Aidan; Forlivesi, Claudio; Liekens, Werner; Perez-cruz, Fernando; Kawsar, Fahim
Sensing WiFi Network for Personal IoT Analytics Artículo en actas
En: 2015 5th International Conference on the Internet of Things (IOT), pp. 104–111, IEEE, Seoul, 2015, ISBN: 978-1-4673-8056-0.
Resumen | Enlaces | BibTeX | Etiquetas: Accelerometers, cloud based query server, cloud computing, data transport mechanism, digital signatures, Distance measurement, Internet of Things, internetworking, IoT analytic, Logic gates, Mobile communication, motion signatures, network servers, Probes, proximity ranging algorithm, Search problems, telecommunication network management, WiFi gateway captures, WiFi management probes, WiFi network, wireless LAN
@inproceedings{Acer2015,
title = {Sensing WiFi Network for Personal IoT Analytics},
author = {Utku Gunay Acer and Aidan Boran and Claudio Forlivesi and Werner Liekens and Fernando Perez-cruz and Fahim Kawsar},
url = {http://ieeexplore.ieee.org/articleDetails.jsp?arnumber=7356554},
doi = {10.1109/IOT.2015.7356554},
isbn = {978-1-4673-8056-0},
year = {2015},
date = {2015-10-01},
booktitle = {2015 5th International Conference on the Internet of Things (IOT)},
pages = {104--111},
publisher = {IEEE},
address = {Seoul},
abstract = {We present the design, implementation and evaluation of an enabling platform for locating and querying physical objects using existing WiFi network. We propose the use of WiFi management probes as a data transport mechanism for physical objects that are tagged with WiFi-enabled accelerometers and are capable of determining their state-of-use based on motion signatures. A local WiFi gateway captures these probes emitted from the connected objects and stores them locally after annotating them with a coarse grained location estimate using a proximity ranging algorithm. External applications can query the aggregated views of state-of-use and location traces of connected objects through a cloud-based query server. We present the technical architecture and algorithms of the proposed platform together with a prototype personal object analytics application and assess the feasibility of our different design decisions. This work makes important contributions by demonstrating that it is possible to build a pure network-based IoT analytics platform with only location and motion signatures of connected objects, and that the WiFi network is the key enabler for the future IoT applications.},
keywords = {Accelerometers, cloud based query server, cloud computing, data transport mechanism, digital signatures, Distance measurement, Internet of Things, internetworking, IoT analytic, Logic gates, Mobile communication, motion signatures, network servers, Probes, proximity ranging algorithm, Search problems, telecommunication network management, WiFi gateway captures, WiFi management probes, WiFi network, wireless LAN},
pubstate = {published},
tppubtype = {inproceedings}
}
We present the design, implementation and evaluation of an enabling platform for locating and querying physical objects using existing WiFi network. We propose the use of WiFi management probes as a data transport mechanism for physical objects that are tagged with WiFi-enabled accelerometers and are capable of determining their state-of-use based on motion signatures. A local WiFi gateway captures these probes emitted from the connected objects and stores them locally after annotating them with a coarse grained location estimate using a proximity ranging algorithm. External applications can query the aggregated views of state-of-use and location traces of connected objects through a cloud-based query server. We present the technical architecture and algorithms of the proposed platform together with a prototype personal object analytics application and assess the feasibility of our different design decisions. This work makes important contributions by demonstrating that it is possible to build a pure network-based IoT analytics platform with only location and motion signatures of connected objects, and that the WiFi network is the key enabler for the future IoT applications.
Luengo, David; Ríos-Muñoz, Gonzalo; Elvira, Victor
Causality analysis of atrial fibrillation electrograms Artículo en actas
En: 2015 Comput. Cardiol. Conf., pp. 585–588, IEEE, 2015, ISBN: 978-1-5090-0685-4.
Resumen | Enlaces | BibTeX | Etiquetas:
@inproceedings{Luengo2015b,
title = {Causality analysis of atrial fibrillation electrograms},
author = {David Luengo and Gonzalo R\'{i}os-Mu\~{n}oz and Victor Elvira},
url = {http://ieeexplore.ieee.org/document/7410978/},
doi = {10.1109/CIC.2015.7410978},
isbn = {978-1-5090-0685-4},
year = {2015},
date = {2015-09-01},
booktitle = {2015 Comput. Cardiol. Conf.},
pages = {585--588},
publisher = {IEEE},
abstract = {Multi-channel intracavitary electrograms (EGMs) are acquired at the electrophysiology laboratory to guide radio frequency catheter ablation of patients suffering from atrial fibrillation (AF). These EGMs are used by cardiologists to determine candidate areas for ablation (e.g., areas corresponding to high dominant frequencies or complex fractionated electrograms). In this paper, we introduce two hierarchical algorithms to retrieve the causal interactions among these multiple EGMs. Both algorithms are based on Granger causality, but other causality measures can be easily incorporated. In both cases, they start by selecting a root node, but they differ on the way in which they explore the set of signals to determine their causeeffect relationships: either testing the full set of unexplored signals (GS-CaRe) or performing a local search only among the set of neighbor EGMs (LS-CaRe). The ensuing causal model provides important information about the propagation of the electrical signals inside the atria, uncovering wavefronts and activation patterns that can guide cardiologists towards candidate areas for catheter ablation. Numerical experiments, on both synthetic signals and annotated real-world signals, show the good performance of the two proposed approaches.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Multi-channel intracavitary electrograms (EGMs) are acquired at the electrophysiology laboratory to guide radio frequency catheter ablation of patients suffering from atrial fibrillation (AF). These EGMs are used by cardiologists to determine candidate areas for ablation (e.g., areas corresponding to high dominant frequencies or complex fractionated electrograms). In this paper, we introduce two hierarchical algorithms to retrieve the causal interactions among these multiple EGMs. Both algorithms are based on Granger causality, but other causality measures can be easily incorporated. In both cases, they start by selecting a root node, but they differ on the way in which they explore the set of signals to determine their causeeffect relationships: either testing the full set of unexplored signals (GS-CaRe) or performing a local search only among the set of neighbor EGMs (LS-CaRe). The ensuing causal model provides important information about the propagation of the electrical signals inside the atria, uncovering wavefronts and activation patterns that can guide cardiologists towards candidate areas for catheter ablation. Numerical experiments, on both synthetic signals and annotated real-world signals, show the good performance of the two proposed approaches.
Luengo, David; Ríos-Muñoz, Gonzalo; Elvira, Victor
Causality Analysis of Atrial Fibrillation Electrograms Artículo en actas
En: Computing in Cardiology, Nice, 2015.
@inproceedings{Luengo2015c,
title = {Causality Analysis of Atrial Fibrillation Electrograms},
author = {David Luengo and Gonzalo R\'{i}os-Mu\~{n}oz and Victor Elvira},
url = {http://www.cinc2015.org/},
year = {2015},
date = {2015-09-01},
booktitle = {Computing in Cardiology},
address = {Nice},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Valera, Isabel; Ruiz, Francisco J R; Svensson, Lennart; Perez-Cruz, Fernando
A Bayesian Nonparametric Approach for Blind Multiuser Channel Estimation Artículo en actas
En: 2015 23rd European Signal Processing Conference (EUSIPCO), pp. 2766–2770, IEEE, Nice, 2015, ISBN: 978-0-9928-6263-3.
Resumen | Enlaces | BibTeX | Etiquetas: Bayes methods, Bayesian nonparametric, communication systems, factorial HMM, Hidden Markov models, machine-to-machine, multiuser communication, Receiving antennas, Signal to noise ratio, Transmitters
@inproceedings{Valera2015b,
title = {A Bayesian Nonparametric Approach for Blind Multiuser Channel Estimation},
author = {Isabel Valera and Francisco J R Ruiz and Lennart Svensson and Fernando Perez-Cruz},
url = {http://ieeexplore.ieee.org/articleDetails.jsp?arnumber=7362888 http://www.eurasip.org/Proceedings/Eusipco/Eusipco2015/papers/1570096659.pdf},
doi = {10.1109/EUSIPCO.2015.7362888},
isbn = {978-0-9928-6263-3},
year = {2015},
date = {2015-08-01},
booktitle = {2015 23rd European Signal Processing Conference (EUSIPCO)},
pages = {2766--2770},
publisher = {IEEE},
address = {Nice},
abstract = {In many modern multiuser communication systems, users are allowed to enter and leave the system at any given time. Thus, the number of active users is an unknown and time-varying parameter, and the performance of the system depends on how accurately this parameter is estimated over time. We address the problem of blind joint channel parameter and data estimation in a multiuser communication channel in which the number of transmitters is not known. For that purpose, we develop a Bayesian nonparametric model based on the Markov Indian buffet process and an inference algorithm that makes use of slice sampling and particle Gibbs with ancestor sampling. Our experimental results show that the proposed approach can effectively recover the data-generating process for a wide range of scenarios.},
keywords = {Bayes methods, Bayesian nonparametric, communication systems, factorial HMM, Hidden Markov models, machine-to-machine, multiuser communication, Receiving antennas, Signal to noise ratio, Transmitters},
pubstate = {published},
tppubtype = {inproceedings}
}
In many modern multiuser communication systems, users are allowed to enter and leave the system at any given time. Thus, the number of active users is an unknown and time-varying parameter, and the performance of the system depends on how accurately this parameter is estimated over time. We address the problem of blind joint channel parameter and data estimation in a multiuser communication channel in which the number of transmitters is not known. For that purpose, we develop a Bayesian nonparametric model based on the Markov Indian buffet process and an inference algorithm that makes use of slice sampling and particle Gibbs with ancestor sampling. Our experimental results show that the proposed approach can effectively recover the data-generating process for a wide range of scenarios.
Santos, Irene; Murillo-Fuentes, Juan Jose; Olmos, Pablo M
Block Expectation Propagation Equalization for ISI Channels Artículo en actas
En: 2015 23rd European Signal Processing Conference (EUSIPCO), pp. 379–383, IEEE, Nice, 2015, ISBN: 978-0-9928-6263-3.
Resumen | Enlaces | BibTeX | Etiquetas: Approximation algorithms, Approximation methods, BCJR algorithm, channel equalization, Complexity theory, Decoding, Equalizers, expectation propagation, ISI, low complexity, Signal processing algorithms
@inproceedings{Santos2015,
title = {Block Expectation Propagation Equalization for ISI Channels},
author = {Irene Santos and Juan Jose Murillo-Fuentes and Pablo M Olmos},
url = {http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=7362409},
doi = {10.1109/EUSIPCO.2015.7362409},
isbn = {978-0-9928-6263-3},
year = {2015},
date = {2015-08-01},
booktitle = {2015 23rd European Signal Processing Conference (EUSIPCO)},
pages = {379--383},
publisher = {IEEE},
address = {Nice},
abstract = {Actual communications systems use high-order modulations and channels with memory. However, as the memory of the channels and the order of the constellations grow, optimal equalization such as BCJR algorithm is computationally intractable, as their complexity increases exponentially with the number of taps and size of modulation. In this paper, we propose a novel low-complexity hard and soft output equalizer based on the Expectation Propagation (EP) algorithm that provides high-accuracy posterior probability estimations at the input of the channel decoder with similar computational complexity than the linear MMSE. We experimentally show that this quasi-optimal solution outperforms classical solutions reducing the bit error probability with low complexity when LDPC channel decoding is used, avoiding the curse of dimensionality with channel memory and constellation size.},
keywords = {Approximation algorithms, Approximation methods, BCJR algorithm, channel equalization, Complexity theory, Decoding, Equalizers, expectation propagation, ISI, low complexity, Signal processing algorithms},
pubstate = {published},
tppubtype = {inproceedings}
}
Actual communications systems use high-order modulations and channels with memory. However, as the memory of the channels and the order of the constellations grow, optimal equalization such as BCJR algorithm is computationally intractable, as their complexity increases exponentially with the number of taps and size of modulation. In this paper, we propose a novel low-complexity hard and soft output equalizer based on the Expectation Propagation (EP) algorithm that provides high-accuracy posterior probability estimations at the input of the channel decoder with similar computational complexity than the linear MMSE. We experimentally show that this quasi-optimal solution outperforms classical solutions reducing the bit error probability with low complexity when LDPC channel decoding is used, avoiding the curse of dimensionality with channel memory and constellation size.
Martino, Luca; Elvira, Victor; Luengo, David; Corander, Jukka
Parallel interacting Markov adaptive importance sampling Artículo en actas
En: 2015 23rd European Signal Processing Conference (EUSIPCO), pp. 499–503, IEEE, Nice, 2015, ISBN: 978-0-9928-6263-3.
Resumen | Enlaces | BibTeX | Etiquetas: Adaptive importance sampling, Bayesian inference, MCMC methods, Monte Carlo methods, Parallel Chains, Probability density function, Proposals, Signal processing, Signal processing algorithms, Sociology
@inproceedings{Martino2015bb,
title = {Parallel interacting Markov adaptive importance sampling},
author = {Luca Martino and Victor Elvira and David Luengo and Jukka Corander},
url = {http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=7362433 http://www.eurasip.org/Proceedings/Eusipco/Eusipco2015/papers/1570111267.pdf},
doi = {10.1109/EUSIPCO.2015.7362433},
isbn = {978-0-9928-6263-3},
year = {2015},
date = {2015-08-01},
booktitle = {2015 23rd European Signal Processing Conference (EUSIPCO)},
pages = {499--503},
publisher = {IEEE},
address = {Nice},
abstract = {Monte Carlo (MC) methods are widely used for statistical inference in signal processing applications. A well-known class of MC methods is importance sampling (IS) and its adaptive extensions. In this work, we introduce an iterated importance sampler using a population of proposal densities, which are adapted according to an MCMC technique over the population of location parameters. The novel algorithm provides a global estimation of the variables of interest iteratively, using all the samples weighted according to the deterministic mixture scheme. Numerical results, on a multi-modal example and a localization problem in wireless sensor networks, show the advantages of the proposed schemes.},
keywords = {Adaptive importance sampling, Bayesian inference, MCMC methods, Monte Carlo methods, Parallel Chains, Probability density function, Proposals, Signal processing, Signal processing algorithms, Sociology},
pubstate = {published},
tppubtype = {inproceedings}
}
Monte Carlo (MC) methods are widely used for statistical inference in signal processing applications. A well-known class of MC methods is importance sampling (IS) and its adaptive extensions. In this work, we introduce an iterated importance sampler using a population of proposal densities, which are adapted according to an MCMC technique over the population of location parameters. The novel algorithm provides a global estimation of the variables of interest iteratively, using all the samples weighted according to the deterministic mixture scheme. Numerical results, on a multi-modal example and a localization problem in wireless sensor networks, show the advantages of the proposed schemes.
Olmos, Pablo M; Mitchell, David G M; Costello, Daniel J
Analyzing the Finite-Length Performance of Generalized LDPC Codes Artículo en actas
En: 2015 IEEE International Symposium on Information Theory (ISIT), pp. 2683–2687, IEEE, Hong Kong, 2015, ISBN: 978-1-4673-7704-1.
Resumen | Enlaces | BibTeX | Etiquetas: BEC, binary codes, binary erasure channel, Block codes, Codes on graphs, Decoding, Differential equations, error probability, finite-length generalized LDPC block codes, finite-length performance analysis, generalized LDPC codes, generalized peeling decoder, GLDPC block codes, graph degree distribution, graph theory, Iterative decoding, parity check codes, protographs
@inproceedings{Olmos2015b,
title = {Analyzing the Finite-Length Performance of Generalized LDPC Codes},
author = {Pablo M Olmos and David G M Mitchell and Daniel J Costello},
url = {http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=7282943},
doi = {10.1109/ISIT.2015.7282943},
isbn = {978-1-4673-7704-1},
year = {2015},
date = {2015-06-01},
booktitle = {2015 IEEE International Symposium on Information Theory (ISIT)},
pages = {2683--2687},
publisher = {IEEE},
address = {Hong Kong},
abstract = {In this paper, we analyze the performance of finite-length generalized LDPC (GLDPC) block codes constructed from protographs when transmission takes place over the binary erasure channel (BEC). A generalized peeling decoder is proposed and we derive a system of differential equations that gives the expected evolution of the graph degree distribution during decoding. We then show that the finite-length performance of a GLDPC code can be estimated by means of a simple scaling law, where a single scaling parameter represents the finite-length properties of the code. We also show that, as we consider stronger component codes, both the asymptotic threshold and the finite-length scaling parameter are improved.},
keywords = {BEC, binary codes, binary erasure channel, Block codes, Codes on graphs, Decoding, Differential equations, error probability, finite-length generalized LDPC block codes, finite-length performance analysis, generalized LDPC codes, generalized peeling decoder, GLDPC block codes, graph degree distribution, graph theory, Iterative decoding, parity check codes, protographs},
pubstate = {published},
tppubtype = {inproceedings}
}
In this paper, we analyze the performance of finite-length generalized LDPC (GLDPC) block codes constructed from protographs when transmission takes place over the binary erasure channel (BEC). A generalized peeling decoder is proposed and we derive a system of differential equations that gives the expected evolution of the graph degree distribution during decoding. We then show that the finite-length performance of a GLDPC code can be estimated by means of a simple scaling law, where a single scaling parameter represents the finite-length properties of the code. We also show that, as we consider stronger component codes, both the asymptotic threshold and the finite-length scaling parameter are improved.
Stinner, Markus; Olmos, Pablo M
Finite-Length Performance of Multi-Edge Protograph-Based Spatially Coupled LDPC Codes Artículo en actas
En: 2015 IEEE International Symposium on Information Theory (ISIT), pp. 889–893, IEEE, Hong Kong, 2015, ISBN: 978-1-4673-7704-1.
Resumen | Enlaces | BibTeX | Etiquetas: binary erasure channel, Block codes, Couplings, Decoding, Error analysis, finite length performance, finite-length performance, graph theory, Iterative decoding, low density parity check codes, multiedge protograph, parity check codes, spatially coupled LDPC codes, spatially-coupled LDPC codes, Steady-state
@inproceedings{Stinner2015,
title = {Finite-Length Performance of Multi-Edge Protograph-Based Spatially Coupled LDPC Codes},
author = {Markus Stinner and Pablo M Olmos},
url = {http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=7282583},
doi = {10.1109/ISIT.2015.7282583},
isbn = {978-1-4673-7704-1},
year = {2015},
date = {2015-06-01},
booktitle = {2015 IEEE International Symposium on Information Theory (ISIT)},
pages = {889--893},
publisher = {IEEE},
address = {Hong Kong},
abstract = {The finite-length performance of multi-edge spatially coupled low-density parity-check (SC-LDPC) codes over the binary erasure channel (BEC) is analyzed. Existing scaling laws are extended to arbitrary protograph base matrices that include puncturing patterns and multiple edges between nodes. A regular protograph-based SC-LDPC construction based on the (4; 8)-regular LDPC block code works well in the waterfall region compared to more involved rate-1/2 structures proposed to improve the threshold to minimum distance trade-off. Scaling laws are also used for code design and to estimate the block length of a given SC-LDPC code ensemble to match the performance of some other code. Estimates on the performance degradation are developed if the chain length varies.},
keywords = {binary erasure channel, Block codes, Couplings, Decoding, Error analysis, finite length performance, finite-length performance, graph theory, Iterative decoding, low density parity check codes, multiedge protograph, parity check codes, spatially coupled LDPC codes, spatially-coupled LDPC codes, Steady-state},
pubstate = {published},
tppubtype = {inproceedings}
}
The finite-length performance of multi-edge spatially coupled low-density parity-check (SC-LDPC) codes over the binary erasure channel (BEC) is analyzed. Existing scaling laws are extended to arbitrary protograph base matrices that include puncturing patterns and multiple edges between nodes. A regular protograph-based SC-LDPC construction based on the (4; 8)-regular LDPC block code works well in the waterfall region compared to more involved rate-1/2 structures proposed to improve the threshold to minimum distance trade-off. Scaling laws are also used for code design and to estimate the block length of a given SC-LDPC code ensemble to match the performance of some other code. Estimates on the performance degradation are developed if the chain length varies.
Vazquez-Vilar, Gonzalo; Martinez, Alfonso; i Fabregas, Albert Guillen
A derivation of the Cost-Constrained Sphere-Packing Exponent Artículo en actas
En: 2015 IEEE International Symposium on Information Theory (ISIT), pp. 929–933, IEEE, Hong Kong, 2015, ISBN: 978-1-4673-7704-1.
Enlaces | BibTeX | Etiquetas: Channel Coding, channel-coding cost-constrained sphere-packing exp, continuous channel, continuous memoryless channel, cost constraint, error probability, hypothesis testing, Lead, Memoryless systems, Optimization, per-codeword cost constraint, reliability function, spherepacking exponent, Testing
@inproceedings{Vazquez-Vilar2015,
title = {A derivation of the Cost-Constrained Sphere-Packing Exponent},
author = {Gonzalo Vazquez-Vilar and Alfonso Martinez and Albert Guillen i Fabregas},
url = {http://ieeexplore.ieee.org/articleDetails.jsp?arnumber=7282591},
doi = {10.1109/ISIT.2015.7282591},
isbn = {978-1-4673-7704-1},
year = {2015},
date = {2015-06-01},
booktitle = {2015 IEEE International Symposium on Information Theory (ISIT)},
pages = {929--933},
publisher = {IEEE},
address = {Hong Kong},
keywords = {Channel Coding, channel-coding cost-constrained sphere-packing exp, continuous channel, continuous memoryless channel, cost constraint, error probability, hypothesis testing, Lead, Memoryless systems, Optimization, per-codeword cost constraint, reliability function, spherepacking exponent, Testing},
pubstate = {published},
tppubtype = {inproceedings}
}
Elvira, Victor; Martino, Luca; Luengo, David; Corander, Jukka
A Gradient Adaptive Population Importance Sampler Artículo en actas
En: 2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 4075–4079, IEEE, Brisbane, 2015, ISBN: 978-1-4673-6997-8.
Resumen | Enlaces | BibTeX | Etiquetas: adaptive extensions, adaptive importance sampler, Adaptive importance sampling, gradient adaptive population, gradient matrix, Hamiltonian Monte Carlo, Hessian matrices, Hessian matrix, learning (artificial intelligence), Machine learning, MC methods, Monte Carlo, Monte Carlo methods, population Monte Carlo (PMC), proposal densities, Signal processing, Sociology, statistics, target distribution
@inproceedings{Elvira2015a,
title = {A Gradient Adaptive Population Importance Sampler},
author = {Victor Elvira and Luca Martino and David Luengo and Jukka Corander},
url = {http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=7178737 http://www.tsc.uc3m.es/~velvira/papers/ICASSP2015_elvira.pdf},
doi = {10.1109/ICASSP.2015.7178737},
isbn = {978-1-4673-6997-8},
year = {2015},
date = {2015-04-01},
booktitle = {2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
pages = {4075--4079},
publisher = {IEEE},
address = {Brisbane},
abstract = {Monte Carlo (MC) methods are widely used in signal processing and machine learning. A well-known class of MC methods is composed of importance sampling and its adaptive extensions (e.g., population Monte Carlo). In this paper, we introduce an adaptive importance sampler using a population of proposal densities. The novel algorithm dynamically optimizes the cloud of proposals, adapting them using information about the gradient and Hessian matrix of the target distribution. Moreover, a new kind of interaction in the adaptation of the proposal densities is introduced, establishing a trade-off between attaining a good performance in terms of mean square error and robustness to initialization.},
keywords = {adaptive extensions, adaptive importance sampler, Adaptive importance sampling, gradient adaptive population, gradient matrix, Hamiltonian Monte Carlo, Hessian matrices, Hessian matrix, learning (artificial intelligence), Machine learning, MC methods, Monte Carlo, Monte Carlo methods, population Monte Carlo (PMC), proposal densities, Signal processing, Sociology, statistics, target distribution},
pubstate = {published},
tppubtype = {inproceedings}
}
Monte Carlo (MC) methods are widely used in signal processing and machine learning. A well-known class of MC methods is composed of importance sampling and its adaptive extensions (e.g., population Monte Carlo). In this paper, we introduce an adaptive importance sampler using a population of proposal densities. The novel algorithm dynamically optimizes the cloud of proposals, adapting them using information about the gradient and Hessian matrix of the target distribution. Moreover, a new kind of interaction in the adaptation of the proposal densities is introduced, establishing a trade-off between attaining a good performance in terms of mean square error and robustness to initialization.
Fernandez-Bes, Jesus; Elvira, Victor; Vaerenbergh, Steven Van
A Probabilistic Least-Mean-Squares Filter Artículo en actas
En: 2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 2199–2203, IEEE, Brisbane, 2015, ISBN: 978-1-4673-6997-8.
Resumen | Enlaces | BibTeX | Etiquetas: adaptable step size LMS algorithm, Adaptation models, adaptive filtering, Approximation algorithms, Bayesian machine learning techniques, efficient approximation algorithm, filtering theory, Least squares approximations, least-mean-squares, probabilistic least mean squares filter, Probabilistic logic, probabilisticmodels, Probability, Signal processing algorithms, Standards, state-space models
@inproceedings{Fernandez-Bes2015,
title = {A Probabilistic Least-Mean-Squares Filter},
author = {Jesus Fernandez-Bes and Victor Elvira and Steven Van Vaerenbergh},
url = {http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=7178361 http://www.tsc.uc3m.es/~velvira/papers/ICASSP2015_bes.pdf},
doi = {10.1109/ICASSP.2015.7178361},
isbn = {978-1-4673-6997-8},
year = {2015},
date = {2015-04-01},
booktitle = {2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
pages = {2199--2203},
publisher = {IEEE},
address = {Brisbane},
abstract = {We introduce a probabilistic approach to the LMS filter. By means of an efficient approximation, this approach provides an adaptable step-size LMS algorithm together with a measure of uncertainty about the estimation. In addition, the proposed approximation preserves the linear complexity of the standard LMS. Numerical results show the improved performance of the algorithm with respect to standard LMS and state-of-the-art algorithms with similar complexity. The goal of this work, therefore, is to open the door to bring somemore Bayesian machine learning techniques to adaptive filtering.},
keywords = {adaptable step size LMS algorithm, Adaptation models, adaptive filtering, Approximation algorithms, Bayesian machine learning techniques, efficient approximation algorithm, filtering theory, Least squares approximations, least-mean-squares, probabilistic least mean squares filter, Probabilistic logic, probabilisticmodels, Probability, Signal processing algorithms, Standards, state-space models},
pubstate = {published},
tppubtype = {inproceedings}
}
We introduce a probabilistic approach to the LMS filter. By means of an efficient approximation, this approach provides an adaptable step-size LMS algorithm together with a measure of uncertainty about the estimation. In addition, the proposed approximation preserves the linear complexity of the standard LMS. Numerical results show the improved performance of the algorithm with respect to standard LMS and state-of-the-art algorithms with similar complexity. The goal of this work, therefore, is to open the door to bring somemore Bayesian machine learning techniques to adaptive filtering.
Luengo, David; Martino, Luca; Elvira, Victor; Bugallo, Monica F
Efficient Linear Combination of Partial Monte Carlo Estimators Artículo en actas
En: 2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 4100–4104, IEEE, Brisbane, 2015, ISBN: 978-1-4673-6997-8.
Resumen | Enlaces | BibTeX | Etiquetas: covariance matrices, efficient linear combination, Estimation, fusion, Global estimator, global estimators, least mean squares methods, linear combination, minimum mean squared error estimators, Monte Carlo estimation, Monte Carlo methods, partial estimator, partial Monte Carlo estimators, Xenon
@inproceedings{Luengo2015bb,
title = {Efficient Linear Combination of Partial Monte Carlo Estimators},
author = {David Luengo and Luca Martino and Victor Elvira and Monica F Bugallo},
url = {http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=7178742 http://www.tsc.uc3m.es/~velvira/papers/ICASSP2015_luengo.pdf},
doi = {10.1109/ICASSP.2015.7178742},
isbn = {978-1-4673-6997-8},
year = {2015},
date = {2015-04-01},
booktitle = {2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
pages = {4100--4104},
publisher = {IEEE},
address = {Brisbane},
abstract = {In many practical scenarios, including those dealing with large data sets, calculating global estimators of unknown variables of interest becomes unfeasible. A common solution is obtaining partial estimators and combining them to approximate the global one. In this paper, we focus on minimum mean squared error (MMSE) estimators, introducing two efficient linear schemes for the fusion of partial estimators. The proposed approaches are valid for any type of partial estimators, although in the simulated scenarios we concentrate on the combination of Monte Carlo estimators due to the nature of the problem addressed. Numerical results show the good performance of the novel fusion methods with only a fraction of the cost of the asymptotically optimal solution.},
keywords = {covariance matrices, efficient linear combination, Estimation, fusion, Global estimator, global estimators, least mean squares methods, linear combination, minimum mean squared error estimators, Monte Carlo estimation, Monte Carlo methods, partial estimator, partial Monte Carlo estimators, Xenon},
pubstate = {published},
tppubtype = {inproceedings}
}
In many practical scenarios, including those dealing with large data sets, calculating global estimators of unknown variables of interest becomes unfeasible. A common solution is obtaining partial estimators and combining them to approximate the global one. In this paper, we focus on minimum mean squared error (MMSE) estimators, introducing two efficient linear schemes for the fusion of partial estimators. The proposed approaches are valid for any type of partial estimators, although in the simulated scenarios we concentrate on the combination of Monte Carlo estimators due to the nature of the problem addressed. Numerical results show the good performance of the novel fusion methods with only a fraction of the cost of the asymptotically optimal solution.
Nazabal, Alfredo; Artés-Rodríguez, Antonio
Discriminative spectral learning of hidden markov models for human activity recognition Artículo en actas
En: 2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 1966–1970, IEEE, Brisbane, 2015, ISBN: 978-1-4673-6997-8.
Resumen | Enlaces | BibTeX | Etiquetas: Accuracy, Bayesian estimation, classify sequential data, Data models, Databases, Discriminative learning, discriminative spectral learning, Hidden Markov models, HMM parameters, Human activity recognition, learning (artificial intelligence), maximum likelihood, maximum likelihood estimation, ML, moment matching learning technique, Observable operator models, sensors, Spectral algorithm, spectral learning, Speech recognition, Training
@inproceedings{Nazabal2015,
title = {Discriminative spectral learning of hidden markov models for human activity recognition},
author = {Alfredo Nazabal and Antonio Art\'{e}s-Rodr\'{i}guez},
url = {http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=7178314},
doi = {10.1109/ICASSP.2015.7178314},
isbn = {978-1-4673-6997-8},
year = {2015},
date = {2015-04-01},
booktitle = {2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
pages = {1966--1970},
publisher = {IEEE},
address = {Brisbane},
abstract = {Hidden Markov Models (HMMs) are one of the most important techniques to model and classify sequential data. Maximum Likelihood (ML) and (parametric and non-parametric) Bayesian estimation of the HMM parameters suffers from local maxima and in massive datasets they can be specially time consuming. In this paper, we extend the spectral learning of HMMs, a moment matching learning technique free from local maxima, to discriminative HMMs. The resulting method provides the posterior probabilities of the classes without explicitly determining the HMM parameters, and is able to deal with missing labels. We apply the method to Human Activity Recognition (HAR) using two different types of sensors: portable inertial sensors, and fixed, wireless binary sensor networks. Our algorithm outperforms the standard discriminative HMM learning in both complexity and accuracy.},
keywords = {Accuracy, Bayesian estimation, classify sequential data, Data models, Databases, Discriminative learning, discriminative spectral learning, Hidden Markov models, HMM parameters, Human activity recognition, learning (artificial intelligence), maximum likelihood, maximum likelihood estimation, ML, moment matching learning technique, Observable operator models, sensors, Spectral algorithm, spectral learning, Speech recognition, Training},
pubstate = {published},
tppubtype = {inproceedings}
}
Hidden Markov Models (HMMs) are one of the most important techniques to model and classify sequential data. Maximum Likelihood (ML) and (parametric and non-parametric) Bayesian estimation of the HMM parameters suffers from local maxima and in massive datasets they can be specially time consuming. In this paper, we extend the spectral learning of HMMs, a moment matching learning technique free from local maxima, to discriminative HMMs. The resulting method provides the posterior probabilities of the classes without explicitly determining the HMM parameters, and is able to deal with missing labels. We apply the method to Human Activity Recognition (HAR) using two different types of sensors: portable inertial sensors, and fixed, wireless binary sensor networks. Our algorithm outperforms the standard discriminative HMM learning in both complexity and accuracy.
Martino, Luca; Elvira, Victor; Luengo, David; Artés-Rodríguez, Antonio; Corander, Jukka
Smelly Parallel MCMC Chains Artículo en actas
En: 2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 4070–4074, IEEE, Brisbane, 2015, ISBN: 978-1-4673-6997-8.
Resumen | Enlaces | BibTeX | Etiquetas: Bayesian inference, learning (artificial intelligence), Machine learning, Markov chain Monte Carlo, Markov chain Monte Carlo algorithms, Markov processes, MC methods, MCMC algorithms, MCMC scheme, mean square error, mean square error methods, Monte Carlo methods, optimisation, parallel and interacting chains, Probability density function, Proposals, robustness, Sampling methods, Signal processing, Signal processing algorithms, signal sampling, smelly parallel chains, smelly parallel MCMC chains, Stochastic optimization
@inproceedings{Martino2015a,
title = {Smelly Parallel MCMC Chains},
author = {Luca Martino and Victor Elvira and David Luengo and Antonio Art\'{e}s-Rodr\'{i}guez and Jukka Corander},
url = {http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=7178736 http://www.tsc.uc3m.es/~velvira/papers/ICASSP2015_martino.pdf},
doi = {10.1109/ICASSP.2015.7178736},
isbn = {978-1-4673-6997-8},
year = {2015},
date = {2015-04-01},
booktitle = {2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
pages = {4070--4074},
publisher = {IEEE},
address = {Brisbane},
abstract = {Monte Carlo (MC) methods are useful tools for Bayesian inference and stochastic optimization that have been widely applied in signal processing and machine learning. A well-known class of MC methods are Markov Chain Monte Carlo (MCMC) algorithms. In this work, we introduce a novel parallel interacting MCMC scheme, where the parallel chains share information, thus yielding a faster exploration of the state space. The interaction is carried out generating a dynamic repulsion among the “smelly” parallel chains that takes into account the entire population of current states. The ergodicity of the scheme and its relationship with other sampling methods are discussed. Numerical results show the advantages of the proposed approach in terms of mean square error, robustness w.r.t. to initial values and parameter choice.},
keywords = {Bayesian inference, learning (artificial intelligence), Machine learning, Markov chain Monte Carlo, Markov chain Monte Carlo algorithms, Markov processes, MC methods, MCMC algorithms, MCMC scheme, mean square error, mean square error methods, Monte Carlo methods, optimisation, parallel and interacting chains, Probability density function, Proposals, robustness, Sampling methods, Signal processing, Signal processing algorithms, signal sampling, smelly parallel chains, smelly parallel MCMC chains, Stochastic optimization},
pubstate = {published},
tppubtype = {inproceedings}
}
Monte Carlo (MC) methods are useful tools for Bayesian inference and stochastic optimization that have been widely applied in signal processing and machine learning. A well-known class of MC methods are Markov Chain Monte Carlo (MCMC) algorithms. In this work, we introduce a novel parallel interacting MCMC scheme, where the parallel chains share information, thus yielding a faster exploration of the state space. The interaction is carried out generating a dynamic repulsion among the “smelly” parallel chains that takes into account the entire population of current states. The ergodicity of the scheme and its relationship with other sampling methods are discussed. Numerical results show the advantages of the proposed approach in terms of mean square error, robustness w.r.t. to initial values and parameter choice.
Perez-Cruz, Fernando; Huang, Howard
A Blind Nonparametric Non-line of Sight Bias Model for Accurate Localization Artículo en actas
En: Information Theory and Applications (ITA), San Diego, 2015.
Resumen | Enlaces | BibTeX | Etiquetas:
@inproceedings{Perez-Cruz2015,
title = {A Blind Nonparametric Non-line of Sight Bias Model for Accurate Localization},
author = {Fernando Perez-Cruz and Howard Huang},
url = {http://ita.ucsd.edu/workshop/15/files/abstract/abstract_1462.txt},
year = {2015},
date = {2015-02-01},
booktitle = {Information Theory and Applications (ITA)},
address = {San Diego},
abstract = {One of the most promising solutions for accurate localization services is estimating the Time Difference of Arrival (TDoA) with a cellular infrastructure and triangulating the position of the sought device. There are three different elements that limit the accuracy of TDoA: bandwidth/snr, clock accuracy and non-line-of-sight (NLOS) bias. The Cramer-Rao lower bound is well known and can be made sufficiently low (centimeters) with existing technologies. GPS clock accuracy is below 15ns (less than 5 meters). NLOS is difficult to characterize and depends heavily on the environment. We cannot rely on simple distributions to model it and we should not expect it to follow a few typical scenarios. In this talk, we present a nonparametric model for estimating the NLOS bias and an algorithm that learns the model on the fly without feedback on the true position. This procedure allows getting accurate localization in any environment and without needing to fine-tune a priori de NLOS for each base station. The actual accuracy depends on the number of bases that hear the device, but uncontrolled outliers no longer limit it. For a dense infrastructure, we show that the localization error can be measured in a few meters.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
One of the most promising solutions for accurate localization services is estimating the Time Difference of Arrival (TDoA) with a cellular infrastructure and triangulating the position of the sought device. There are three different elements that limit the accuracy of TDoA: bandwidth/snr, clock accuracy and non-line-of-sight (NLOS) bias. The Cramer-Rao lower bound is well known and can be made sufficiently low (centimeters) with existing technologies. GPS clock accuracy is below 15ns (less than 5 meters). NLOS is difficult to characterize and depends heavily on the environment. We cannot rely on simple distributions to model it and we should not expect it to follow a few typical scenarios. In this talk, we present a nonparametric model for estimating the NLOS bias and an algorithm that learns the model on the fly without feedback on the true position. This procedure allows getting accurate localization in any environment and without needing to fine-tune a priori de NLOS for each base station. The actual accuracy depends on the number of bases that hear the device, but uncontrolled outliers no longer limit it. For a dense infrastructure, we show that the localization error can be measured in a few meters.
2014
Farajtabar, Mehrdad; Du, Nan; Gomez-rodriguez, Manuel; Valera, Isabel; Zha, Hongyuan; Song, Le
Shaping Social Activity by Incentivizing Users Artículo en actas
En: Advances in Neural Information Processing Systems, pp. 2474–2482, Montreal, 2014.
Resumen | Enlaces | BibTeX | Etiquetas:
@inproceedings{Farajtabar2014,
title = {Shaping Social Activity by Incentivizing Users},
author = {Mehrdad Farajtabar and Nan Du and Manuel Gomez-rodriguez and Isabel Valera and Hongyuan Zha and Le Song},
url = {http://papers.nips.cc/paper/5365-shaping-social-activity-by-incentivizing-users.pdf},
year = {2014},
date = {2014-12-01},
booktitle = {Advances in Neural Information Processing Systems},
volume = {December},
pages = {2474--2482},
address = {Montreal},
abstract = {Events in an online social network can be categorized roughly into endogenous events, where users just respond to the actions of their neighbors within the network, or exogenous events, where users take actions due to drives external to the network. How much external drive should be provided to each user, such that the network activity can be steered towards a target state? In this paper, we model social events using multivariate Hawkes processes, which can capture both endogenous and exogenous event intensities, and derive a time dependent linear relation between the intensity of exogenous events and the overall network activity. Exploiting this connection, we develop a convex optimization framework for determining the required level of external drive in order for the network to reach a desired activity level. We experimented with event data gathered from Twitter, and show that our method can steer the activity of the network more accurately than alternatives.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Events in an online social network can be categorized roughly into endogenous events, where users just respond to the actions of their neighbors within the network, or exogenous events, where users take actions due to drives external to the network. How much external drive should be provided to each user, such that the network activity can be steered towards a target state? In this paper, we model social events using multivariate Hawkes processes, which can capture both endogenous and exogenous event intensities, and derive a time dependent linear relation between the intensity of exogenous events and the overall network activity. Exploiting this connection, we develop a convex optimization framework for determining the required level of external drive in order for the network to reach a desired activity level. We experimented with event data gathered from Twitter, and show that our method can steer the activity of the network more accurately than alternatives.
Bocharova, Irina E; i Fàbregas, Albert Guillén; Kudryashov, Boris D; Martinez, Alfonso; Campo, Adria Tauste; Vazquez-Vilar, Gonzalo
Source-Channel Coding with Multiple Classes Artículo en actas
En: 2014 IEEE International Symposium on Information Theory (ISIT 2014), Honolulu, HI, USA, 2014.
BibTeX | Etiquetas:
@inproceedings{gvazquez-isit2014,
title = {Source-Channel Coding with Multiple Classes},
author = {Irina E Bocharova and Albert Guill\'{e}n i F\`{a}bregas and Boris D Kudryashov and Alfonso Martinez and Adria Tauste Campo and Gonzalo Vazquez-Vilar},
year = {2014},
date = {2014-06-01},
booktitle = {2014 IEEE International Symposium on Information Theory (ISIT 2014)},
address = {Honolulu, HI, USA},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Taborda, Camilo G; Perez-Cruz, Fernando; Guo, Dongning
New Information-Estimation Results for Poisson, Binomial and Negative Binomial Models Artículo en actas
En: 2014 IEEE International Symposium on Information Theory, pp. 2207–2211, IEEE, Honolulu, 2014, ISBN: 978-1-4799-5186-4.
Resumen | Enlaces | BibTeX | Etiquetas: Bregman divergence, Estimation, estimation measures, Gaussian models, Gaussian processes, information measures, information theory, information-estimation results, negative binomial models, Poisson models, Stochastic processes
@inproceedings{Taborda2014,
title = {New Information-Estimation Results for Poisson, Binomial and Negative Binomial Models},
author = {Camilo G Taborda and Fernando Perez-Cruz and Dongning Guo},
url = {http://ieeexplore.ieee.org/articleDetails.jsp?arnumber=6875225},
doi = {10.1109/ISIT.2014.6875225},
isbn = {978-1-4799-5186-4},
year = {2014},
date = {2014-06-01},
booktitle = {2014 IEEE International Symposium on Information Theory},
pages = {2207--2211},
publisher = {IEEE},
address = {Honolulu},
abstract = {In recent years, a number of mathematical relationships have been established between information measures and estimation measures for various models, including Gaussian, Poisson and binomial models. In this paper, it is shown that the second derivative of the input-output mutual information with respect to the input scaling can be expressed as the expectation of a certain Bregman divergence pertaining to the conditional expectations of the input and the input power. This result is similar to that found for the Gaussian model where the Bregman divergence therein is the square distance. In addition, the Poisson, binomial and negative binomial models are shown to be similar in the small scaling regime in the sense that the derivative of the mutual information and the derivative of the relative entropy converge to the same value.},
keywords = {Bregman divergence, Estimation, estimation measures, Gaussian models, Gaussian processes, information measures, information theory, information-estimation results, negative binomial models, Poisson models, Stochastic processes},
pubstate = {published},
tppubtype = {inproceedings}
}
In recent years, a number of mathematical relationships have been established between information measures and estimation measures for various models, including Gaussian, Poisson and binomial models. In this paper, it is shown that the second derivative of the input-output mutual information with respect to the input scaling can be expressed as the expectation of a certain Bregman divergence pertaining to the conditional expectations of the input and the input power. This result is similar to that found for the Gaussian model where the Bregman divergence therein is the square distance. In addition, the Poisson, binomial and negative binomial models are shown to be similar in the small scaling regime in the sense that the derivative of the mutual information and the derivative of the relative entropy converge to the same value.
Miguez, Joaquin
On the uniform asymptotic convergence of a distributed particle filter Artículo en actas
En: 2014 IEEE 8th Sensor Array and Multichannel Signal Processing Workshop (SAM), pp. 241–244, IEEE, A Coruña, 2014, ISBN: 978-1-4799-1481-4.
Resumen | Enlaces | BibTeX | Etiquetas: ad hoc networks, Approximation algorithms, approximation errors, Approximation methods, classical convergence theorems, Convergence, convergence of numerical methods, distributed particle filter scheme, distributed signal processing algorithms, Monte Carlo methods, parallel computing systems, particle filtering (numerical methods), Signal processing, Signal processing algorithms, stability assumptions, uniform asymptotic convergence, Wireless Sensor Networks, WSNs
@inproceedings{Miguez2014,
title = {On the uniform asymptotic convergence of a distributed particle filter},
author = {Joaquin Miguez},
url = {http://ieeexplore.ieee.org/articleDetails.jsp?arnumber=6882385},
doi = {10.1109/SAM.2014.6882385},
isbn = {978-1-4799-1481-4},
year = {2014},
date = {2014-06-01},
booktitle = {2014 IEEE 8th Sensor Array and Multichannel Signal Processing Workshop (SAM)},
pages = {241--244},
publisher = {IEEE},
address = {A Coru\~{n}a},
abstract = {Distributed signal processing algorithms suitable for their implementation over wireless sensor networks (WSNs) and ad hoc networks with communications and computing capabilities have become a hot topic during the past years. One class of algorithms that have received special attention are particles filters. However, most distributed versions of this type of methods involve various heuristic or simplifying approximations and, as a consequence, classical convergence theorems for standard particle filters do not hold for their distributed counterparts. In this paper, we look into a distributed particle filter scheme that has been proposed for implementation in both parallel computing systems and WSNs, and prove that, under certain stability assumptions regarding the physical system of interest, its asymptotic convergence is guaranteed. Moreover, we show that convergence is attained uniformly over time. This means that approximation errors can be kept bounded for an arbitrarily long period of time without having to progressively increase the computational effort.},
keywords = {ad hoc networks, Approximation algorithms, approximation errors, Approximation methods, classical convergence theorems, Convergence, convergence of numerical methods, distributed particle filter scheme, distributed signal processing algorithms, Monte Carlo methods, parallel computing systems, particle filtering (numerical methods), Signal processing, Signal processing algorithms, stability assumptions, uniform asymptotic convergence, Wireless Sensor Networks, WSNs},
pubstate = {published},
tppubtype = {inproceedings}
}
Distributed signal processing algorithms suitable for their implementation over wireless sensor networks (WSNs) and ad hoc networks with communications and computing capabilities have become a hot topic during the past years. One class of algorithms that have received special attention are particles filters. However, most distributed versions of this type of methods involve various heuristic or simplifying approximations and, as a consequence, classical convergence theorems for standard particle filters do not hold for their distributed counterparts. In this paper, we look into a distributed particle filter scheme that has been proposed for implementation in both parallel computing systems and WSNs, and prove that, under certain stability assumptions regarding the physical system of interest, its asymptotic convergence is guaranteed. Moreover, we show that convergence is attained uniformly over time. This means that approximation errors can be kept bounded for an arbitrarily long period of time without having to progressively increase the computational effort.
Crisan, Dan; Miguez, Joaquin
Nested Particle Filters for Sequential Parameter Estimation in Discrete-time State-space Models Artículo en actas
En: SIAM 2014 Conference on Uncertainty Quantification, Savannah, 2014.
@inproceedings{Crisan2014b,
title = {Nested Particle Filters for Sequential Parameter Estimation in Discrete-time State-space Models},
author = {Dan Crisan and Joaquin Miguez},
year = {2014},
date = {2014-03-01},
booktitle = {SIAM 2014 Conference on Uncertainty Quantification},
address = {Savannah},
abstract = {The problem of estimating the parameters of nonlinear, possibly non-Gaussian discrete-time state models has drawn considerable attention during the past few years, leading to the appearance of general methodologies (SMC2, particle MCMC, recursive ML) that have improved on earlier, simpler extensions of the standard particle filter. However, there is still a lack of recursive (online) methods that can provide a theoretically-grounded approximation of the joint posterior probability distribution of the parameters and the dynamic state variables of the model. In the talk, we will describe a two-layer particle filtering scheme that addresses this problem. Both a recursive algorithm, suitable for online implementation, and some results regarding its asymptotic convergence will be presented.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The problem of estimating the parameters of nonlinear, possibly non-Gaussian discrete-time state models has drawn considerable attention during the past few years, leading to the appearance of general methodologies (SMC2, particle MCMC, recursive ML) that have improved on earlier, simpler extensions of the standard particle filter. However, there is still a lack of recursive (online) methods that can provide a theoretically-grounded approximation of the joint posterior probability distribution of the parameters and the dynamic state variables of the model. In the talk, we will describe a two-layer particle filtering scheme that addresses this problem. Both a recursive algorithm, suitable for online implementation, and some results regarding its asymptotic convergence will be presented.
Martino, Luca; Elvira, Víctor; Luengo, David
An Adaptive Population Importance Sampler Artículo en actas
En: IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2014), Florencia, 2014.
Enlaces | BibTeX | Etiquetas: ALCIT, COMPREHENSION
@inproceedings{Martino2014,
title = {An Adaptive Population Importance Sampler},
author = {Luca Martino and V\'{i}ctor Elvira and David Luengo},
url = {http://www.icassp2014.org/home.html},
year = {2014},
date = {2014-01-01},
booktitle = {IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2014)},
address = {Florencia},
keywords = {ALCIT, COMPREHENSION},
pubstate = {published},
tppubtype = {inproceedings}
}
Pastore, Adriano; Koch, Tobias; Fonollosa, Javier Rodriguez
A Rate-Splitting Approach to Fading Multiple-Access Channels with Imperfect Channel-State Information Artículo en actas
En: International Zurich Seminar on Communications (IZS), Zurich, 2014.
Resumen | Enlaces | BibTeX | Etiquetas: ALCIT
@inproceedings{Pastore2014,
title = {A Rate-Splitting Approach to Fading Multiple-Access Channels with Imperfect Channel-State Information},
author = {Adriano Pastore and Tobias Koch and Javier Rodriguez Fonollosa},
url = {http://www.tsc.uc3m.es/~koch/files/IZS_2014_009-012.pdf http://e-collection.library.ethz.ch/eserv/eth:8192/eth-8192-01.pdf},
year = {2014},
date = {2014-01-01},
booktitle = {International Zurich Seminar on Communications (IZS)},
address = {Zurich},
abstract = {As shown by M´edard, the capacity of fading channels with imperfect channel-state information (CSI) can be lowerbounded by assuming a Gaussian channel input and by treating the unknown portion of the channel multiplied by the channel input as independent worst-case (Gaussian) noise. Recently, we have demonstrated that this lower bound can be sharpened by a rate-splitting approach: by expressing the channel input as the sum of two independent Gaussian random variables (referred to as layers), say X = X1+X2, and by applying M´edard’s bounding technique to first lower-bound the capacity of the virtual channel from X1 to the channel output Y (while treating X2 as noise), and then lower-bound the capacity of the virtual channel from X2 to Y (while assuming X1 to be known), one obtains a lower bound that is strictly larger than M´edard’s bound. This ratesplitting approach is reminiscent of an approach used by Rimoldi and Urbanke to achieve points on the capacity region of the Gaussian multiple-access channel (MAC). Here we blend these two rate-splitting approaches to derive a novel inner bound on the capacity region of the memoryless fading MAC with imperfect CSI. Generalizing the above rate-splitting approach to more than two layers, we show that, irrespective of how we assign powers to each layer, the supremum of all rate-splitting bounds is approached as the number of layers tends to infinity, and we derive an integral expression for this supremum. We further derive an expression for the vertices of the best inner bound, maximized over the number of layers and over all power assignments.},
keywords = {ALCIT},
pubstate = {published},
tppubtype = {inproceedings}
}
As shown by M´edard, the capacity of fading channels with imperfect channel-state information (CSI) can be lowerbounded by assuming a Gaussian channel input and by treating the unknown portion of the channel multiplied by the channel input as independent worst-case (Gaussian) noise. Recently, we have demonstrated that this lower bound can be sharpened by a rate-splitting approach: by expressing the channel input as the sum of two independent Gaussian random variables (referred to as layers), say X = X1+X2, and by applying M´edard’s bounding technique to first lower-bound the capacity of the virtual channel from X1 to the channel output Y (while treating X2 as noise), and then lower-bound the capacity of the virtual channel from X2 to Y (while assuming X1 to be known), one obtains a lower bound that is strictly larger than M´edard’s bound. This ratesplitting approach is reminiscent of an approach used by Rimoldi and Urbanke to achieve points on the capacity region of the Gaussian multiple-access channel (MAC). Here we blend these two rate-splitting approaches to derive a novel inner bound on the capacity region of the memoryless fading MAC with imperfect CSI. Generalizing the above rate-splitting approach to more than two layers, we show that, irrespective of how we assign powers to each layer, the supremum of all rate-splitting bounds is approached as the number of layers tends to infinity, and we derive an integral expression for this supremum. We further derive an expression for the vertices of the best inner bound, maximized over the number of layers and over all power assignments.