University of Graz

About: University of Graz is a education organization based out in Graz, Steiermark, Austria. It is known for research contribution in the topics: Population & Quantum chromodynamics. The organization has 17934 authors who have published 37489 publications receiving 1110980 citations. The organization is also known as: Carolo Franciscea Graecensis & Karl Franzens Universität.

Causability and explainability of artificial intelligence in medicine.

Abstract: Explainable artificial intelligence (AI) is attracting much interest in medicine. Technically, the problem of explainability is as old as AI itself and classic AI represented comprehensible retraceable approaches. However, their weakness was in dealing with uncertainties of the real world. Through the introduction of probabilistic learning, applications became increasingly successful, but increasingly opaque. Explainable AI deals with the implementation of transparency and traceability of statistical black-box machine learning methods, particularly deep learning (DL). We argue that there is a need to go beyond explainable AI. To reach a level of explainable medicine we need causability. In the same way that usability encompasses measurements for the quality of use, causability encompasses measurements for the quality of explanations. In this article, we provide some necessary definitions to discriminate between explainability and causability as well as a use-case of DL interpretation and of human explanation in histopathology. The main contribution of this article is the notion of causability, which is differentiated from explainability in that causability is a property of a person, while explainability is a property of a system This article is categorized under: Fundamental Concepts of Data and Knowledge > Human Centricity and User Interaction.

Cardioprotection and lifespan extension by the natural polyamine spermidine

Abstract: Aging is associated with an increased risk of cardiovascular disease and death. Here we show that oral supplementation of the natural polyamine spermidine extends the lifespan of mice and exerts cardioprotective effects, reducing cardiac hypertrophy and preserving diastolic function in old mice. Spermidine feeding enhanced cardiac autophagy, mitophagy and mitochondrial respiration, and it also improved the mechano-elastical properties of cardiomyocytes in vivo, coinciding with increased titin phosphorylation and suppressed subclinical inflammation. Spermidine feeding failed to provide cardioprotection in mice that lack the autophagy-related protein Atg5 in cardiomyocytes. In Dahl salt-sensitive rats that were fed a high-salt diet, a model for hypertension-induced congestive heart failure, spermidine feeding reduced systemic blood pressure, increased titin phosphorylation and prevented cardiac hypertrophy and a decline in diastolic function, thus delaying the progression to heart failure. In humans, high levels of dietary spermidine, as assessed from food questionnaires, correlated with reduced blood pressure and a lower incidence of cardiovascular disease. Our results suggest a new and feasible strategy for protection against cardiovascular disease.

A new dynamic approach for statistical optimization of GNSS radio occultation bending angles for optimal climate monitoring utility

Abstract: [1] Global Navigation Satellite System (GNSS)-based radio occultation (RO) is a satellite remote sensing technique providing accurate profiles of the Earth’s atmosphere for weather and climate applications. Above about 30km altitude, however, statistical optimization is a critical process for initializing the RO bending angles in order to optimize the climate monitoring utility of the retrieved atmospheric profiles. Here we introduce an advanced dynamic statistical optimization algorithm, which uses bending angles from multiple days of European Centre for Medium-range Weather Forecasts (ECMWF) short-range forecast and analysis fields, together with averaged-observed bending angles, to obtain background profiles and associated error covariance matrices with geographically varying background uncertainty estimates on a daily updated basis. The new algorithm is evaluated against the existing Wegener Center Occultation Processing System version 5.4 (OPSv5.4) algorithm, using several days of simulated MetOp and observed CHAMP and COSMIC data, for January and July conditions. We find the following for the new method’s performance compared to OPSv5.4: 1.) it significantly reduces random errors (standard deviations), down to about half their size, and leaves less or about equal residual systematic errors (biases) in the optimized bending angles; 2.) the dynamic (daily) estimate of the background error correlation matrix alone already improves the optimized bending angles; 3.) the subsequently retrievedrefractivityprofilesandatmospheric(temperature)profilesbenefit by improvederror characteristics,especiallyabove about 30km. Based on theseencouraging results, we work to employ similar dynamic error covariance estimation also for the observed bending angles and to apply the method to full months and subsequently to entire climate data records.