ABSTRACT
Since the onset of the COVID-19 pandemic in 2019, many clinical prognostic scoring tools have been proposed or developed to aid clinicians in the disposition and severity assessment of pneumonia. However, there is limited work that focuses on explaining techniques that are best suited for clinicians in their decision making. In this article, we present a new image explainability method named ensemble AI explainability (XAI), which is based on the SHAP and Grad-CAM++ methods. It provides a visual explanation for a deep learning prognostic model that predicts the mortality risk of community-acquired pneumonia and COVID-19 respiratory infected patients. In addition, we surveyed the existing literature and compiled prevailing quantitative and qualitative metrics to systematically review the efficacy of ensemble XAI, and to make comparisons with several state-of-the-art explainability methods (LIME, SHAP, saliency map, Grad-CAM, Grad-CAM++). Our quantitative experimental results have shown that ensemble XAI has a comparable absence impact (decision impact: 0.72, confident impact: 0.24). Our qualitative experiment, in which a panel of three radiologists were involved to evaluate the degree of concordance and trust in the algorithms, has showed that ensemble XAI has localization effectiveness (mean set accordance precision: 0.52, mean set accordance recall: 0.57, mean set F1: 0.50, mean set IOU: 0.36) and is the most trusted method by the panel of radiologists (mean vote: 70.2%). Finally, the deep learning interpretation dashboard used for the radiologist panel voting will be made available to the community. Our code is available at https://github.com/IHIS-HealthInsights/Interpretation-Methods-Voting-dashboard. © 2020 IEEE.
ABSTRACT
With the continuous iteration of digital technologies Artificial Intelligence AI is emerging as a core strength leading the technological revolution and industrial transformation. Furthermore since the outbreak of COVID-19 epidemic AI has been promoted to a new high. To build new international competitiveness AI has been undoubtedly seen as an important avenue to improve the status of international specializations. Under the opportunity of global value chain GVC reconstruction the traditional comparative advantages such as labor cost are gradually weakened. The textile industry urgently needs to improve the core competitiveness of products through intelligent manufacturing and upgrade from the processing and manufacturing link with the lowest added value to the textile machinery production and product design link with higher added value. Does the current development level of AI have an obvious driving effect on the international specializations status of textile industry in various countries What is the specific mechanism The discussion of these problems has theoretical value and important practical significance for the formulation of AI-related policies the transformation of textile industry and the promotion path selection of the international specializations status. This paper measures the comprehensive development index of AI in 28 countries from 2010 to 2017. Based on the world input-output table a global value chain position index is constructed to measure a country's status of international specializations. The two indicators are connected within a unified framework and the multi-dimensional panel fixed effect model is used for empirical test of textile industry. Then using the intermediary effect model this paper analyzes the influence mechanism of AI on the status of international specializations from three channels technological innovation production efficiency and human capital. There are three possible marginal contributions of this paper First the research on integrated evaluation systems for AI development in 28 economies from 2010 to 2017 allowing for the international comparison and dynamic tracking from multiple dimensions could be initial efforts to break up the one-fold measurement of AI. Second it is the first time to use AI as an emerging influencing factor of the international specializations status in the textile industry. Third through connecting the AI index with international specializations status of the textile industry within the unified accounting framework our study provides a better understanding of mechanisms for AI influence on international specializations from three channels technological innovation production efficiency and human capital. The main conclusions are as follows First AI has significantly improved the position of a country's textile industry in the global value chain. Second through the mechanism test it is confirmed that AI has improved the international specializations status of the textile industry through three channels technological innovation production efficiency and human capital. Among them the kinetic energy of promoting the textile industry to move up the value chain through production efficiency needs to be further stimulated. Therefore in order to seize the opportunity of technological revolution the textile industry should actively develop intelligent manufacturing and use AI technology to help enterprises complete the transformation of automation intelligence and digitization so as to improve the added value and technical content of export products. Meanwhile enterprises and governments should increase R&D research and development investment vigorously promote the transformation of innovative achievements of AI technology in the textile industry and make it gradually occupy the core position of global value chain in international competition. Finally in order to solve the key problems of textile industry such as the over-dependence on low-end labor force administrative departments should speed up the construction of high-end talent team of AI. The emer ing technology represented by AI provides a new path choice for the international specializations status of the textile industry which can not only further promote the AI sustainable development and the effective integration of all links of the textile industry but also break through the dilemma of "low-end locking" of the global value chain so as to achieve the goal of promoting the digital transformation of the textile industry and improving the status of international specializations. It provides a theoretical and factual basis for AI to cultivate new digital kinetic energy in the upgrading of textile industry and to participate in the positioning of a new round of international competition. © 2022 China Silk Association. All rights reserved.
ABSTRACT
Serum amyloid A(SAA) is a novel marker widely used in the acute infection disease, especially viral infection. SAA has shown a cerntain value in assisting the clinical diagnosis, discrimination of severity and monitoring of progress and outcome of COVID-19. This paper introduces the application of SAA structural, function andits dynamic detection in the diagnosis of COVID-19, and the significance of combined detection with COVID-19 antibodies, nucleic acid and other diagnostic indicators.