A5001741874- Climate variability and models
- Meteorological Phenomena and Simulations
- Domain Adaptation and Few-Shot Learning
- Metaheuristic Optimization Algorithms Research
- Speech and Audio Processing
- Energy Load and Power Forecasting
- Speech Recognition and Synthesis
- Hydrological Forecasting Using AI
- Neural Networks and Applications
- Geophysics and Gravity Measurements
- Face and Expression Recognition
- Hydrology and Watershed Management Studies
- Statistical Methods and Inference
- Pulsars and Gravitational Waves Research
- Voice and Speech Disorders
- Amyotrophic Lateral Sclerosis Research
- Machine Learning and ELM
- Text and Document Classification Technologies
- Genetics and Physical Performance
- Gamma-ray bursts and supernovae
- Species Distribution and Climate Change
- Catalytic Processes in Materials Science
- Advanced Statistical Methods and Models
- Monoclonal and Polyclonal Antibodies Research
- Time Series Analysis and Forecasting
Lawrence Livermore National Laboratory
2018-2025
National Institute of Amazonian Research
2022
Centre for Research and Development in Telecommunications (Brazil)
2015-2018
Universidade Estadual de Campinas (UNICAMP)
2009-2017
University of Minnesota
2014-2017
Twin Cities Orthopedics
2017
University of Minho
2017
Polytechnic Institute of Bragança
2014-2016
Federal State Budget Scientific Institution Institute of Applied Mathematics and Automation
2015
Instituto Federal Catarinense
2012
Abstract Background Machine learning (ML) has made a significant impact in medicine and cancer research; however, its these areas been undeniably slower more limited than other application domains. A major reason for this the lack of availability patient data to broader ML research community, large part due privacy protection concerns. High-quality, realistic, synthetic datasets can be leveraged accelerate methodological developments medicine. By large, medical is high dimensional often...
Abstract The fidelity of climate projections is often undermined by biases in models due to their simplification or misrepresentation unresolved processes. While various bias correction methods have been developed post‐process model outputs match observations, existing approaches usually focus on limited, low‐order statistics, break either the spatiotemporal consistency target variable, its dependency upon resolved dynamics. We develop a Regularized Adversarial Domain Adaptation (RADA)...
The COVID-19 pandemic underscored the promise of monoclonal antibody-based prophylactic and therapeutic drugs
This technical note addresses the discrete-time Markov jump linear systems <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">H</i> <sub xmlns:xlink="http://www.w3.org/1999/xlink">infin</sub> filtering design problem. First, under assumption that parameter is measurable, main contribution matrix inequality (LMI) characterization of all filters such estimation error remains bounded by a given norm level, yielding complete solution mode-dependent Based...
This paper provides an overview of the Speaker Anti-spoofing Competition organized by Biometric group at Idiap Research Institute for IEEE International Conference on Biometrics: Theory, Applications, and Systems (BTAS 2016). The competition used AVspoof database, which contains a comprehensive set presentation attacks, including, (i) direct replay attacks when genuine data is played back using laptop two phones (Samsung Galaxy S4 iPhone 3G), (ii) synthesized speech replayed with laptop,...
An algorithm to measure the jitter (jitta, jitter, rap and ppq5) shimmer (ShdB, Shim, apq3 apq5) parameters was developed. These can be used in an intelligent system diagnose voice pathologies. The is sensitive fundamental frequency determines based on maximum minimum functions applied each glottal period of signal. previously tested with synthesized speech signals very high accuracy, but several improvements had included for analysis pathologic voices. A comparison using real, control...
A synthesized speech signal was used to measure the accuracy of Jitter and Shimmer parameters calculated by a previously presented algorithm. The formant model synthesis produce signals with controlled glottal periods magnitudes according determined values. (jitta, jitter, rap ppq5) (ShdB, Shim, apq3 apq5) were developed algorithm compared analytic values also measures made Praat software. Experiments different type jitter shimmer perturbations F0 conducted. Also influence variations on experimented.
Multi-task learning (MTL) aims to improve generalization performance by multiple related tasks simultaneously. While sometimes the underlying task relationship structure is known, often needs be estimated from data at hand. In this paper, we present a novel family of models for MTL, applicable regression and classification problems, capable relationship. particular, consider joint estimation problem individual parameters, which solved using alternating minimization. The revealed founded on...
Abstract The path toward realizing the potential of seasonal forecasting and its socioeconomic benefits relies on improving general circulation model (GCM) based dynamical forecast systems. To improve forecasts, it is crucial to set up benchmarks, clarify limitations posed by initialization errors, formulation deficiencies, internal climate variability. With huge costs in generating large ensembles, limited observations for verification, benchmarking diagnosing task proves challenging. Here,...
Este relato de experiência trata ações educativas que fazem parte das atividades promoção à saúde, contempladas pelo Programa Residência Multiprofissional em Saúde da Família Universidade Estadual Santa Cruz. O objetivo deste artigo é descrever as desenvolvidas durante um projeto educação saúde sob uma perspectiva multidisciplinar, visou instrumentalizar escolares acerca dos benefícios alimentação adequada, hábitos vida saudáveis e impactos negativos do excesso alimentos ricos açúcares...
This paper proposes a hierarchical Bayesian multitask learning model that is applicable to the general multi-task binary classification problem where assumes shared sparsity structure across different tasks. We derive computationally efficient inference algorithm based on variational approximate posterior distribution. demonstrate potential of new approach various synthetic datasets and for predicting human health status microbiome profile. Our analysis incorporates data pooled from multiple...
A insuficiência cardíaca com fração de ejeção preservada (ICFEp), a forma mais comum (IC), é definida como sendo uma síndrome clínica disfunção diastólica ventricular esquerda, elevada pressão enchimento esquerdo e reduzida adesão miocárdica. Apesar dos desafios inerentes à abordagem terapêutica da ICFEp, alguns tratamentos são promissores, dentre os quais se destacam terapias farmacológicas, antagonistas do sistema renina-angiotensina-aldosterona, mineralocorticoides inibidores SGLT2, não...
Most previously authorized clinical antibodies against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) have lost neutralizing activity to recent variants due rapid viral evolution. To mitigate such escape, we preemptively enhance AZD3152, an antibody for prophylaxis in immunocompromised individuals. Using deep mutational scanning (DMS) on the SARS-CoV-2 antigen, identify AZD3152 vulnerabilities at antigen positions F456 and D420. Through two iterations of computational design...
Multi-task learning (MTL) aims to improve generalization performance by multiple related tasks simultaneously. While sometimes the underlying task relationship structure is known, often needs be estimated from data at hand. In this paper, we present a novel family of models for MTL, applicable regression and classification problems, capable relationships. particular, consider joint estimation problem individual parameters, which solved using alternating minimization. The component builds on...
Alzheimer’s disease (AD), the most common type of dementia, not only imposes a huge financial burden on health care system, but also psychological and emotional patients their families. There is thus an urgent need to infer trajectories cognitive performance over time identify biomarkers predictive progression. In this article, we propose multi-task learning with fused Laplacian sparse group lasso model, which can closely related measures due its sparsity-inducing property, model progression...
This study analyzed examples of sustainable ecosystem-based agriculture where management methods supported livelihoods smallholders while at the same time local ecosystem services were enhanced in Ethiopia, Brazil, and Philippines. Participation by farmers collective actions found to be a crucial driving force, as specific knowledge "learning doing" main components development. Social cohesion, particularly through associations cooperatives, improved marketing opportunities also important...
Research in the area of automatic speaker verification (ASV) has advanced enough for industry to start using ASV systems practical applications. However, these are highly vulnerable spoofing or presentation attacks(PAs), limiting their wide deployment. Several speech-based attack detection (PAD) methods have been proposed recently but most them based on hand crafted frequency phase-based features. Although convolutional neural networks (CNN) already shown breakthrough results face...
This paper describes presentation attack detection systems developed for the Automatic Speaker Verification Spoofing and Countermeasures Challenge (ASVspoof 2017). The submitted systems, using calibration score fusion techniques, combine different sub-systems (up to 18), which are based on eight state of art features rely Gaussian mixture models feed-forward neural network classifiers. achieved top five performances in competition. We present proposed analyze strategies employed. To assess...
Objective: To assemble and characterize an electronic health record (EHR) dataset for a large cohort of US military Veterans diagnosed with ALS (Amyotrophic Lateral Sclerosis). Methods: An EHR 19,662 between January 1, 2000 to December 31, 2020 was compiled from the Health Administration (VHA) database by query ICD9 diagnosis (335.20) or ICD10 (G12.21) Amyotrophic Sclerosis. Results: The is predominantly male (98.94%) white (72.37%) median age at disease onset 68 years survival date 590...
Future projection of climate is typically obtained by combining outputs from multiple Earth System Models (ESMs) for several variables such as temperature and precipitation. While IPCC has traditionally used a simple model output average, recent work illustrated potential advantages using multitask learning (MTL) framework projections individual variables. In this paper we introduce hierarchical (HMTL) with two levels tasks that each super-task, i.e., task at the top level, itself problem...
The description of the stellar interior compact stars remains as a big challenge for nuclear astrophysics community. consolidated knowledge is restricted to density regions around saturation hadronic matter ρ0=2.8×1014gcm−3, regimes where our models are successfully applied. As one moves towards higher densities and extreme conditions up quark/gluons deconfinement, little can be said about microphysics equation state (EoS). Here, we employ Markov Chain Monte Carlo (MCMC) strategy access...
In this paper, we propose an estimation of distribution algorithm based on inexpensive Gaussian mixture model with online learning, which will be employed in dynamic optimization. Here, the stores a vector sufficient statistics best solutions, is subsequently used to obtain parameters components. This approach able incorporate into current potentially relevant information previous and iterations. The nature proposal desirable context optimization, where prompt reaction new scenarios should...
Multitask learning (MTL) leverages commonalities across related tasks with the aim of improving individual task performance. A key modeling choice in designing MTL models is structure tasks' relatedness, which may not be known. Here we propose a Bayesian multitask model that able to infer relationship directly from data. We present two variations terms priori information relatedness. First, diffuse Wishart prior placed on precision matrix so all are assumed equally priori. Second, graphical...