305,843 research outputs found
Endotelio corneale umano: dagli studi in vitro all’applicazione dell’ ingegneria tissutale
No full textL'endotelio corneale regola lo stato di idratazione stromale necessario per la trasparenza corneale. In età adulta, la densità cellulare endoteliale (ECD) diminuisce annualmente dello 0,6%. Poiché le cellule endoteliali corneali umane hanno ridotta capacità proliferativa in vivo, la loro perdita è compensata dalla migrazione e allargamento delle cellule vicine. Quando l' ECD scende al di sotto del valore soglia di 500 cellule/mm2, in seguito a invecchiamento o trauma o una condizione patologica, l'endotelio non è in grado di garantire una corretta idratazione corneale, causando edema, opacità corneale e disturbi visivi. Il trapianto corneale, con le relative limitazioni, ad oggi è l'unico trattamento efficiente per le malattie endoteliali corneali. Tuttavia, la carenza mondiale di cornee donatrici sta diventando sempre più un problema non trascurabile, con solo 1 cornea disponibile ogni 70 cornee richieste. Questo ha indotto a sviluppare strategie alternative per il trattamento di malattie endoteliali corneali, tra cui gli approcci di ingegneria tissutale.
L'ingegneria tissutale è un approccio terapeutico emergente che combina l'utilizzo di cellule endoteliali corneali con l’utilizzo di un appropriato biomateriale per la coltura ed il trapianto di queste cellule. Il nostro gruppo di ricerca ha precedentemente dimostrato che il legame di un decapeptide contenente il motivo peptidico RGD (Arg-Gly-Asp) allo scaffold di chitina garantisce il mantenimento del comportamento delle cellule epiteliali corneali umane. Le caratteristiche dello scaffold sono state ottimizzate per produrre un substrato con proprietà biomeccaniche simili allo stroma corneale umano per trasparenza, architettura, rigidità e resistenza meccanica.
In questo progetto di ricerca, il nostro obiettivo è quello di studiare l'utilizzo della chitina funzionalizzata con l’ RGD come potenziale substrato anche per l'adesione e l'espansione delle cellule corneali endoteliali umane. Gli esperimenti sono stati condotti al fine di ottenere un tessuto endoteliale ingegnerizzato e, in una prospettiva futura, una cornea umana tridimensionale con tutti i suoi strati (epitelio + stroma + endotelio).
Tuttavia, l'ingegneria tissutale dell’ endotelio corneale è una sfida complessa per diversi motivi: a) le cellule corneali endoteliali hanno una bassa capacità proliferativa che deve essere stimolata finemente in vitro con terreni di coltura appropriati; b) durante la coltura in vitro, le cellule corneali endoteliali vanno incontro a senescenza prematura (in particolare nelle colture cellulari derivate da donatori più anziani) e a una trasformazione fenotipica assumendo un fenotipo mesenchimale, la cosiddetta transizione endoteliale-mesenchimale; e) pochi marcatori molecolari specifici definiscono la qualità delle cellule corneali endoteliali, necessari per controllare le loro funzioni fisiologiche cellulari; d) infine, per l’approccio di ingegneria tissutale dell’ endotelio corneale, non è stato ancora sviluppato un biomateriale in grado di creare un microambiente favorevole all'attività delle cellule corneali endoteliali.
Per questo motivo, in questo progetto di ricerca abbiamo affrontato alcune sfide che rendono difficile l’utilizzo delle cellule corneali endoteliali, in termini di (I) ottimizzazione della tecnica di coltura delle cellule corneali endoteliali umane (Capitolo I), (II) identificazione di marcatori funzionali specifici delle cellule corneali endoteliali (Capitolo II), (III) prevenzione della transizione endotelio-mesenchimale che induce ad un trans-differenziamento cellulare verso un fenotipo mio-fibroblastico che causa una perdita della funzione cellulare (Capitolo III) e (IV) analisi dello scaffold selezionato per coltivare le cellule corneali endoteliali (Capitolo IV).The corneal endothelium (CE) is the innermost layer of the cornea that regulates the stromal hydration state required to maintain corneal transparency. During adulthood, the endothelial cell density (ECD) decreases by 0.6% each year. As human corneal endothelial cells (hCECs) do not proliferate, the loss of aging-induced hCECs is compensated by migration and enlargement of neighbouring cells. When ECD falls below a threshold of 500 cells/mm2, by aging or trauma/disease, the endothelium does not have enough pumping power to guarantee a correct corneal hydration, leading to oedema, corneal opacity, and visual impairment. Corneal transplantation, with related problems, is the only efficient treatment for corneal endothelial diseases up to date. However, the worldwide donor corneas shortage is increasingly becoming a non-negligible issue, with only 1 cornea available for 70 needed. This has led to investigate alternative strategies for treating corneal endothelial diseases, such as tissue engineering approaches.
Corneal endothelial tissue engineering is an emerging therapeutic approach that involves the use of hCECs combined with a biomaterial to create tissue engineered grafts for transplantation. Our research group has previously demonstrated that proper binding of an RGD (Arg-Gly-Asp) peptide to the chitin scaffold guaranteed maintenance of human corneal epithelial cells behaviour. Scaffold characteristics were optimised to produce a substrate with biomechanical properties resembling the human corneal stroma for transparency, architecture, stiffness, and mechanical strength.
In this research project, our aim is to investigate the use of this functionalized biological scaffold as a potential substrate also for hCECs adhesion and expansion. The experiments were carried out in order to obtain a functional tissue engineered endothelial graft and, from a future perspective, a three-dimensional human cornea with all its layers (epithelium + stroma + endothelium). If successful, this elegant approach has the potential to increase access to corneal therapy by treating multiple patients.
However, CE tissue engineering is a major challenge for several reasons: a) the hCECs have a low natural proliferative capacity that must be finely stimulated in vitro with an appropriate mitogen-rich medium; b) during in vitro expansion, hCECs undergo premature senescence (particularly in cultures derived from older donors) and phenotypic transformation to a mesenchymal phenotype, so-called Endothelial-Mesenchymal Transition (EnMT), which must be prevented c) few specific molecular markers define the quality of cultured hCECs, which are needed to control their physiological cell functions; d) finally, to develop a tailored engineered corneal endothelium, a substrate material that is able to create a favourable microenvironment for hCECs activity has not been yet developed.
Thus, in this research project we analysed some challenges faced with hCECs in terms of (I) optimization of hCECs culture techniques (Chapter I), (II) identification of specific hCECs functional markers (Chapter II), (III) prevention of EnMT which leads to a cellular trans-differentiation towards a myofibroblastic phenotype causing a cellular loss of function (Chapter III), and (IV) analysis of the identified scaffold to make bioengineered corneal endothelial grafts (Chapter IV)
DL-FRONT MERRA-2 weather front probability maps over North America, 1980-
No full text<p>DL-FRONT is a Deep Learning Neural Network (DLNN) that was trained to detect weather fronts using spatial grids of near-surface atmospheric variables. The dataset is composed of hourly spatial grids containing probability maps for each of five front-type categories—cold front, warm front, stationary front, occluded front, and no front.</p>
<p>This dataset is the product of processing data from the National Aeronautics and Space Administration (NASA) <a href="https://gmao.gsfc.nasa.gov/reanalysis/MERRA-2/">Modern-Era Retrospective analysis for Research and Applications, Version 2</a> (MERRA-2). DL-FRONT processed MERRA-2 hourly data grids of instantaneous measures of air pressure reduced to mean sea level, air temperature at 2 meters, specific humidity at 2 meters, and wind velocity at 10 meters over the time span 1980 - 2018 to produce this dataset. The original MERRA-2 data were resampled at 1 degree resolution over the spatial range 31W - 171W x 10N - 77N using bicubic interpolation.</p>
<p>At each hourly time step the network produced a set of spatial grids with the same resolution and spatial range as the input, one for each of the five categories mentioned above. Each cell in a spatial grid for a given category records the network-assigned probability (from 0.0 to 1.0) that the cell is in a weather front boundary region of that category (or, for the "no front" category, the probability that the cell is not in any weather front boundary region).</p>
<p>The DLNN was trained using MERRA-2 data and human-identified fronts from the NOAA National Weather Service (NWS) Weather Prediction Center (WPC) <a href="https://www.wpc.ncep.noaa.gov/html/sfc2.shtml">Coded Surface Bulletin</a> dataset. The training datasets covered the years 2003-2007.</p>
<p>The dataset contains two sets of files. The first set contains the original front probability maps. The second set contains "one hot" versions of the front probability maps. In the one hot version the five front-type probabilities for a spatial grid cell for a given time step are replaced by the value 1 for the largest front-type probability, and by 0 for the others.</p>
<p>The front probability files have names that follow the form merra2_merra2-1deg_fronts_<year>.nc. The one hot files have names that follow the form merra2_merra2-1deg_onehot_<year>.nc. Each file contains one year of hourly spatial data grids.</p>
MERRA/AS: The MERRA Analytic Services Project Interim Report
No full textMERRA AS is a cyberinfrastructure resource that will combine iRODS-based Climate Data Server (CDS) capabilities with Coudera MapReduce to serve MERRA analytic products, store the MERRA reanalysis data collection in an HDFS to enable parallel, high-performance, storage-side data reductions, manage storage-side driver, mapper, reducer code sets and realized objects for users, and provide a library of commonly used spatiotemporal operations that can be composed to enable higher-order analyses
MERRA-2: File Specification
No full textThe second Modern-Era Retrospective analysis for Research and Applications (MERRA-2) is a NASA atmospheric reanalysis that begins in 1980. It replaces the original MERRA reanalysis (Rienecker et al., 2011) using an upgraded version of the Goddard Earth Observing System Model, Version 5 (GEOS-5) data assimilation system. The file collections for MERRA-2 are described in detail in this document, including some important changes from those of the MERRA dataset (Lucchesi, 2012)
Land Surface Precipitation and Hydrology in MERRA-2
No full textThe Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2), provides global, 1-hourly estimates of land surface conditions for 1980-present at 50-km resolution. Outside of the high latitudes, MERRA-2 uses observations-based precipitation data products to correct the precipitation falling on the land surface. This paper describes the precipitation correction method and evaluates the MERRA-2 land surface precipitation and hydrology. Compared to monthly GPCPv2.2 observations, the corrected MERRA-2 precipitation (M2CORR) is better than the precipitation generated by the atmospheric models within the cyclingMERRA-2 system and the earlier MERRA reanalysis. Compared to 3-hourlyTRMM observations, the M2CORR diurnal cycle has better amplitude but less realistic phasing than MERRA-2 model-generated precipitation. Because correcting the precipitation within the coupled atmosphere-land modeling system allows the MERRA-2 near-surface air temperature and humidity to respond to the improved precipitation forcing, MERRA-2 provides more self-consistent surface meteorological data than were available from the earlier, offline MERRA-Land reanalysis. Overall, MERRA-2 land hydrology estimates are better than those of MERRA-Land and MERRA. A comparison against GRACE satellite observations of terrestrial water storage demonstrates clear improvements in MERRA-2 over MERRA in South America and Africa but also reflects known errors in the observations used to correct the MERRA-2 precipitation. The MERRA-2 and MERRA-Land surface and root zone soil moisture skill vs. in situ measurements is slightly higher than that of ERA-Interim Land and higher than that of MERRA (significantly for surface soil moisture). Snow amounts from MERRA-2 have lower bias and correlate better against reference data than do those of MERRA-Land and MERRA, with MERRA-2 skill roughly matching that of ERA-Interim Land. Seasonal anomaly R values against naturalized stream flow measurements in the United States are, on balance, highest for MERRA-2 and ERA-Interim Land, somewhat lower for MERRA-Land, and lower still for MERRA
Assessment of MERRA-2 Land Surface Energy Flux Estimates
No full textIn MERRA-2, observed precipitation is inserted in place of model-generated precipitation at the land surface. The use of observed precipitation was originally developed for MERRA-Land(a land-only replay of MERRA with model-generated precipitation replaced with observations).Previously shown that the land hydrology in MERRA-2 and MERRA-Land is better than MERRA. We test whether the improved land surface hydrology in MERRA-2 leads to the expected improvements in the land surface energy fluxes and 2 m air temperatures (T2m)
Adversarial Machine Learning in Recommender Systems
No full textL’attività di ricerca del dottorando Merra Felice Antonio si è sviluppata partendo dallo studio dello stato dell’arte per quanto riguarda i Sistemi di raccomandazione. Contemporaneamente è stato studiato lo stato dell’arte delle tecniche di Adversarial Machine Learning e della sicurezza dei Sistemi di raccomandazione. Sin da subito è emerso che i sistemi di raccomandazione possono essere soggetto a tecniche di “adversarial attacks” con conseguenze molto gravi nelle piattaforme che utilizzano tali sistemi. L’identificazione della mancanza di una risorsa ocmune di ricerca e di possibili nuovi attacchi e difese ha portato alla suddivisione del mio lavoro id ricerca in 4 macro fasi: svolgimento di una revisione della letteratura, sviluppo di tecniche d'interpretazione dell’efficacia degli attacchi, proposta di nuove strategie di attacco, e studio degli effetti delle tecniche di difesa sulle performance globali di un recommender.
Nel primo filone di ricerca è stata svolta una revisione della letteratura che ha riguardato centinaia id articoli e dal cui risultati di analisi è stato pubblicato un survey in cui ho proposta una tassonomia e possibili problematiche future e nuove direzioni di ricerca. Una parte di queste problematiche è stata oggetto di ricerca del mio secondo filone di ricerca. In tale filone ho proposto un modello d'interpretazione della robustezza dei sistemi collaborativi soggetti a tutti gli attacchi allo stato dell’arte. Inoltre, in tale filone, ho proposto una nuova strategia di attacco che inglobasse l’informazione semantica (pubblica).
Nel terzo filone di ricerca ho proposto differenti strategie di attacco che possono distruggere l’affidabilità di un sistema di raccomandazione multimediale con l’inserimento d'immagini “adversarial” le cui perturbazioni sono completamente non visualizzabili dagli esseri umani. A tal proposto e studiato nuove difese che possano proteggere tali sistemi.
Il quarto filone di ricerca è stato dedicato all’analisi dei sistemi di raccomandazione basati su modello (come i sistemi a fattorizzazione di matrici). In questo filone ho proposto una nuova tecnica di attacco di tipo iterativo che ha dimostrato come esistano strategie che rendono il sistema di raccomandazione completamente casuale, perdendo completamente quanto imparato in fase di addestramento. Inoltre, ho studiato matematicamente l’effetto di una tecnica di addestramento adversarial che ha ottenuto molta popolarità nella comunità dei sistemi di raccomandazione constatando che il miglioramento dell’accuratezza dei sistemi è connesso ad una amplificazione della popolarità dei sistemi.Recommender systems are ubiquitous. Our digital lives are influenced by their use when, for instance, we select the news to read, the product to buy, the friend to connect, and the movie to watch. While enormous academic research efforts have been mainly focused on getting high-quality recommendations to reach the maximum customers' satisfaction, little effort has been devoted to studying the integrity and security of these systems. Is there an underlying relationship between the characteristics of the historical user-item interactions and the efficacy of injection of false users/feedback strategies against collaborative models? Can public semantic data be used to perform attacks more potent in raising the recommendability of victim items? Can a malicious user (i.e., the adversary) poison or evade the image data of visual recommenders with adversarial perturbed product images?
What is a possible defensive solution to reduce the effectiveness of test-time adversarial attacks?
Is the family of model-based recommenders more vulnerable to multi-step gradient-based adversarial perturbations? Furthermore, is the adversarial training robustification still effective in the last scenario? Is this training defense influencing the beyond-accuracy and bias performance?
This dissertation intends to pave the way towards more robust recommender systems, beginning with understanding how a model can be made more robust, the cost of robustness in terms of recommendation quality, and the adversarial risks of modern recommenders. This thesis, getting inspiration from the literature on the security of collaborative models against the insertion of hand-engineered fake profiles and the recent advances of adversarial machine learning methods in other research areas like computer vision, contributes to several directions: (i) the proposal of a practical framework to interpret the impact of data characteristics on the robustness of collaborative recommenders, (ii) the design of powerful attack strategies using publicly available semantic data, (iii) the identification of severe adversarial vulnerabilities of visual-based recommender models where adversaries can break the recommendation integrity by pushing products to the highest recommendation positions with a simple and human-imperceptible perturbation of products' images, (iv) the design of a novel defense method to protect visual recommenders against test-time adversarial attacks, (v) the proposal of robust adversarial perturbation methods capable of completely breaking the accuracy of matrix factorization recommenders, and (vi) a formal study that examines the effects of adversarial training in reducing the recommendation quality of state-of-the-art model-based recommenders
- …
