A5019831662- Statistical Methods and Inference
- Bayesian Methods and Mixture Models
- Bayesian Modeling and Causal Inference
- Statistical Methods and Bayesian Inference
- T-cell and B-cell Immunology
- Immune Cell Function and Interaction
- Soil Geostatistics and Mapping
- Spine and Intervertebral Disc Pathology
- Anomaly Detection Techniques and Applications
- Hematopoietic Stem Cell Transplantation
- Cerebrovascular and Carotid Artery Diseases
- Musculoskeletal pain and rehabilitation
- Scoliosis diagnosis and treatment
- Adversarial Robustness in Machine Learning
- Intracranial Aneurysms: Treatment and Complications
- Single-cell and spatial transcriptomics
- Health, Environment, Cognitive Aging
- CAR-T cell therapy research
- Musicians’ Health and Performance
- Cardiovascular Health and Disease Prevention
- Gaussian Processes and Bayesian Inference
- Renal and related cancers
- Pregnancy-related medical research
- Radiomics and Machine Learning in Medical Imaging
- Extracellular vesicles in disease
Dalle Molle Institute for Artificial Intelligence Research
2015-2025
University of Applied Sciences and Arts of Southern Switzerland
2015-2025
Università della Svizzera italiana
2017-2021
Politecnico di Milano
2011-2014
According to existing literature, musicians are at risk of experiencing a range painful musculoskeletal conditions. Recently, novel digital technology was developed investigate pain location and extent. The aim this study describe extent in using method for drawing (PD) analysis. Additionally, the association between PD variables clinical features were explored with pain.One hundred fifty-eight (90 women 68 men; aged 22.4 ± 3.6 years) recruited from Swiss U.K. conservatories. Participants...
Abstract Immune reconstitution plays a crucial role on the outcome of patients given T cell-depleted HLA-haploidentical hematopoietic stem cell transplantation (hHSCT) for hematological malignancies. CD1d-restricted invariant NKT (iNKT) cells are innate-like, lipid-reactive lymphocytes controlling infections, cancer, and autoimmunity. Adult mature iNKT divided in two functionally distinct CD4+ CD4− subsets that express NK receptor CD161 derive from thymic CD4+CD161− precursors. We...
Adoptive immunotherapy with T cells engineered tumor-specific cell receptors (TCRs) holds promise for cancer treatment. However, suppressive cues generated in the tumor microenvironment (TME) can hinder efficacy of these therapies, prompting search strategies to overcome detrimental conditions and improve cellular therapeutic approaches. CD1d-restricted invariant natural killer (iNKT) actively participate immunosurveillance by restricting myeloid populations TME. Here, we showed that...
Single-cell sequencing provides rich information; however, its clinical use is limited due to high costs and complex data output. Here, we present a protocol for extracting single-cell-related information from bulk RNA-sequencing (RNA-seq) using the pathway-level extractor (PLIER) algorithm. We describe steps single-cell signatures literature, training PLIER model based on (named CLIER), applying it new dataset. This produces latent variables that are interpretable in context of specific...
We propose an innovative method for the accurate estimation of surfaces and spatial fields when prior knowledge phenomenon under study is available. The included in model derives from physics, physiology, or mechanics problem at hand, formalized terms a partial differential equation governing behavior, as well conditions that has to satisfy boundary domain. proposed models exploit advanced scientific computing techniques specifically make use finite element method. estimators have penalized...
The application of single-cell technologies in clinical nephrology remains elusive. We generated an atlas transcriptionally defined cell types and states human kidney disease by integrating signatures reported the literature with newly obtained from 5 patients acute injury. used this information to develop kidney-specific cell-level ExtractoR (K-CLIER), a transfer learning approach specifically tailored evaluate role types/states on bulk RNAseq data. validated K-CLIER as reliable...
Information theoretic feature selection methods quantify the importance of each by estimating mutual information terms to capture: relevancy, redundancy and complementarity. These are commonly estimated maximum likelihood, while an under-explored area research is how use shrinkage instead. Our work suggests a novel method for data-efficient estimation terms. The small sample behaviour makes it particularly suitable discrete distributions with large number categories (bins). Using our...
We study a class of models at the interface between statistics and numerical analysis. Specifically, we consider nonparametric regression for estimation spatial fields from pointwise noisy observations, which account problem-specific prior information, described in terms partial differential equation governing phenomenon under study. The information is incorporated model via roughness term using penalized framework. prove well-posedness problem, resort to mixed equal order finite element...
Background: In Switzerland, Aedes albopictus is well established in Ticino, south of the Alps, where surveillance and control are implemented. The mosquito has also been observed Swiss cities north Alps. Decision-making tools urgently needed by local authorities order to optimize control. Methods: A regularized logistic regression was used link long-term dataset Ae. occurrence Ticino with socioenvironmental predictors. probability establishment extrapolated Switzerland more finely Basel...
The study aims to create a preoperative model from baseline demographic and health-related quality of life scores (HRQOL) predict good excellent early clinical outcome using machine learning (ML) approach. A single spine surgery center retrospective review prospectively collected data January 2016 December 2020 the institutional registry (SpineREG) was performed. inclusion criteria were age ≥ 18 years, both sexes, lumbar arthrodesis procedure, complete follow up assessment (Oswestry...
We present a novel approach for estimating conditional probability tables, based on joint, rather than independent, estimate of the distributions belonging to same table. derive exact analytical expressions estimators and we analyse their properties both analytically via simulation. then apply this method estimation parameters in Bayesian network. Given structure network, proposed better estimates joint distribution significantly improves classification performance with respect traditional...
Abstract Aims To create, using a machine learning (ML) approach, preoperative model from baseline demographic and health-related quality of life scores (HRQOL) to predict good excellent early clinical outcome. Patients Methods A single spine surgery center retrospective review prospectively collected data January 2016 December 2020 the institutional registry (SpineREG) was performed. The inclusion criteria were age ≥ 18 years, both sexes, lumbar arthrodesis procedure, complete follow up...
We present a novel strategy for detecting global outliers in federated learning setting, targeting particular cross-silo scenarios. Our approach involves the use of two servers and transmission masked local data from clients to one servers. The masking prevents disclosure sensitive information while still permitting identification outliers. Moreover, further safeguard privacy, permutation mechanism is implemented so that server does not know which client owns any point. performs outlier...
The impact of physical activity on a person's progression to type 2 diabetes is multifaceted. Systems ordinary differential equations have been crucial in simulating this progression. However, such models often operate multiple timescales, making them computationally expensive when long-term effects. To overcome this, we propose homogenized version two-timescale model that captures the short- and effects blood glucose regulation. By invoking contribution session into effects, reduce full...
Score functions for learning the structure of Bayesian networks in literature assume that data are a homogeneous set observations; whereas it is often case they comprise different related, but not homogeneous, sets collected ways. In this paper we propose new Dirichlet score, which call Hierarchical (BHD). The proposed score based on hierarchical model pools information across to learn single encompassing network structure, while taking into account differences their probabilistic...