A5085691944- Target Tracking and Data Fusion in Sensor Networks
- Bayesian Methods and Mixture Models
- Gaussian Processes and Bayesian Inference
- Markov Chains and Monte Carlo Methods
- Statistical Methods and Bayesian Inference
- Statistical Methods and Inference
- Probabilistic and Robust Engineering Design
- Distributed Sensor Networks and Detection Algorithms
- Bayesian Modeling and Causal Inference
- Statistical Distribution Estimation and Applications
- Blind Source Separation Techniques
- Fault Detection and Control Systems
- Control Systems and Identification
- Water Systems and Optimization
- Advanced MIMO Systems Optimization
- Sparse and Compressive Sensing Techniques
- Advanced Wireless Communication Techniques
- Probability and Risk Models
- Advanced Statistical Methods and Models
- Advanced Adaptive Filtering Techniques
- Wireless Communication Networks Research
- Time Series Analysis and Forecasting
- Advanced Statistical Process Monitoring
- Neural Networks and Applications
- Statistical Mechanics and Entropy
University of Edinburgh
2019-2025
IMT Nord Europe
2016-2024
Centre de Recherche en Informatique
2016-2024
Université de Lille
2016-2024
Maxwell Institute for Mathematical Sciences
2023-2024
The Alan Turing Institute
2022-2023
Turing Institute
2022
Institut Mines-Télécom
2016-2019
Edinburgh College
2019
Centre National de la Recherche Scientifique
2016-2018
A fundamental problem in signal processing is the estimation of unknown parameters or functions from noisy observations. Important examples include localization objects wireless sensor networks [1] and Internet Things [2]; multiple source reconstruction electroencephalograms [3]; power spectral density for speech enhancement [4]; inference genomic [5]. Within Bayesian framework, these problems are addressed by constructing posterior probability distributions unknowns. The posteriors combine...
Importance Sampling methods are broadly used to approximate posterior distributions or some of their moments. In its standard approach, samples drawn from a single proposal distribution and weighted properly. However, since the performance depends on mismatch between targeted distributions, several densities often employed for generation samples. Under this Multiple (MIS) scenario, many works have addressed selection adaptation interpreting sampling weighting steps in different ways. paper,...
In this work we address the problem of short-term load forecasting. We propose a generalization linear state-space model where evolution state and observation matrices is unknown. The proposed blind Kalman filter algorithm proceeds via alternating estimation these unknown inference state, within framework expectation-maximization. A mini-batch processing strategy introduced to allow on-the-fly experimental results show that method outperforms state-of-the-art techniques by considerable...
The effective sample size (ESS) is widely used in sample-based simulation methods for assessing the quality of a Monte Carlo approximation given distribution and related integrals. In this paper, we revisit ESS specific context importance sampling (IS). derivation approximation, that will denote as $\widehat{\text{ESS}}$, partially available Kong (1992). This has been last 25 years due to its simplicity practical rule thumb wide variety methods. However, show multiple assumptions...
Monte Carlo (MC) methods are well-known computational techniques, widely used in different fields such as signal processing, communications and machine learning. An important class of MC is composed importance sampling (IS) its adaptive extensions, population (PMC) multiple IS (AMIS). In this paper, we introduce a novel iterated sampler using proposal densities. The proposed algorithm, named (APIS), provides global estimation the variables interest iteratively, making use all samples...
Particle filters are broadly used to approximate posterior distributions of hidden states in state-space models by means sets weighted particles. While the convergence filter is guaranteed when number particles tends infinity, quality approximation usually unknown but strongly dependent on In this paper, we propose a novel method for assessing particle online manner, as well simple scheme adaptation based assessment. The sequential comparison between actual observations and their predictive...
We investigate the problem of distributed multitarget tracking by using a set netted, collaborative sensors with limited sensing range (LSR), where each sensor runs sequential Monte Carlo probability hypothesis density filter and exchanges relevant posterior information its neighbors. The key challenge stems from LSR neighbor whose fields view (FoVs) are partially/non-overlapped, therefore, they may observe different targets at same time. With regard to local common FoVs among sensors,...
Abstract State‐space models are an increasingly common and important tool in the quantitative ecologists’ armoury, particularly for analysis of time‐series data. This is due to both their flexibility intuitive structure, describing different individual processes a complex system, thus simplifying model specification step. composed two (a) system (or state) process that describes dynamics true underlying state over time; (b) observation links observed data with current at time. Specification...
We address the multisensor multitarget tracking problem based on a hierarchical sensor network. In this setup, there is fusion center, several cluster heads, and many sensors. Each runs Gaussian mixture probability hypothesis density (PHD) filter. The sensors send their locally calculated components to local head in presence of false data injection (FDI) denial-of-service (DoS) attackers. propose hybrid PHD averaging framework that consists two parts: one uses arithmetic average (AA)...
Multiple importance sampling (MIS) methods use a set of proposal distributions from which samples are drawn. Each sample is then assigned an weight that can be obtained according to different strategies. This work motivated by the trade-off between variance reduction and computational complexity approaches (classical vs. deterministic mixture) available for calculation. A new method achieves efficient compromise both factors introduced in this paper. It based on forming partition computing...
Importance sampling (IS) and numerical integration methods are usually employed for approximating moments of complicated target distributions. In its basic procedure, the IS methodology randomly draws samples from a proposal distribution weights them accordingly, accounting mismatch between proposal. this work, we present general framework techniques inspired by methodology. The can also be seen as an incorporation deterministic rules into methods, reducing error estimators several orders...
Abstract Statistical signal processing applications usually require the estimation of some parameters interest given a set observed data. These estimates are typically obtained either by solving multi-variate optimization problem, as in maximum likelihood (ML) or posteriori (MAP) estimators, performing multi-dimensional integration, minimum mean squared error (MMSE) estimators. Unfortunately, analytical expressions for these estimators cannot be found most real-world applications, and Monte...
Limited information is available on incidence and outcomes of COVID-19 in patients with multiple sclerosis (MS). This study investigated the risks SARS-CoV-2 infection COVID-19-related MS, compared these general population.A regional registry was created to collect data incidence, hospitalization rates, intensive care unit admission, death MS COVID-19. National government seroprevalence were used for comparison. The conducted at 14 specialist treatment centers Madrid, Spain, between February...
Sequential Monte Carlo methods, also known as particle filtering, have seen an explosion of development both in theory and applications. The publication [1] sparked huge interest the area sequential signal processing, particularly filtering. Ever since, number publications which filtering plays a prominent role has continued to grow. An early reference is [2] later tutorials include [3]-[9]. With we estimate probability density functions (pdfs) by mass functions, whose masses are placed at...
Importance sampling (IS) is a Monte Carlo technique for the approximation of intractable distributions and integrals with respect to them. The origin IS dates from early 1950s. In last decades, rise Bayesian paradigm increase available computational resources have propelled interest in this theoretically sound methodology. paper, we first describe basic algorithm then revisit recent advances We pay particular attention two sophisticated lines. First, focus on multiple (MIS), case where more...
Background Fingolimod is a functional sphingosine-1-phosphate (S1P) antagonist approved for the treatment of multiple sclerosis (MS). affects lymphocyte subpopulations and regulates gene expression in transcriptome. Translational studies are necessary to identify cellular molecular biomarkers that might be used predict clinical response drug. In MS patients, we aimed clarify differential effects fingolimod on T, B natural killer (NK) cell subsets differentially expressed genes responders...