A5071894546- Artificial Intelligence in Healthcare and Education
- Machine Learning in Healthcare
- Ethics and Social Impacts of AI
- Healthcare Technology and Patient Monitoring
- Semiconductor Lasers and Optical Devices
- Explainable Artificial Intelligence (XAI)
- Digital Holography and Microscopy
- Semiconductor Quantum Structures and Devices
- Translation Studies and Practices
- COVID-19 epidemiological studies
- Photonic and Optical Devices
- Clinical Reasoning and Diagnostic Skills
- Adversarial Robustness in Machine Learning
- Ethics in Clinical Research
- Electronic Health Records Systems
- Optical measurement and interference techniques
- Safety Systems Engineering in Autonomy
- Sepsis Diagnosis and Treatment
- Advanced Sensor and Control Systems
- Text and Document Classification Technologies
- Fungal Biology and Applications
- Anomaly Detection Techniques and Applications
- Spectroscopy and Chemometric Analyses
- Posttraumatic Stress Disorder Research
- Innovative Educational Techniques
University of York
2019-2024
Peng Cheng Laboratory
2024
Shenzhen Institute of Information Technology
2024
Harbin Institute of Technology
2024
University of Copenhagen
2024
Jiangxi University of Finance and Economics
2024
Shanghai University of Engineering Science
2024
National University of Singapore
2024
Tianjin Medical University General Hospital
2024
Nanjing University
2024
In recent years, several new technical methods have been developed to make AI-models more transparent and interpretable. These techniques are often referred collectively as ‘AI explainability’ or ‘XAI’ methods. This paper presents an overview of XAI methods, links them stakeholder purposes for seeking explanation. Because the underlying broadly ethical in nature, we see this analysis a contribution towards bringing together dimensions XAI. We emphasize that use must be linked explanations...
Established approaches to assuring safety-critical systems and software are difficult apply employing ML where there is no clear, pre-defined specification against which assess validity. This problem exacerbated by the "opaque" nature of learnt model not amenable human scrutiny. Explainable AI (XAI) methods have been proposed tackle this issue producing human-interpretable representations models can help users gain confidence build trust in system. However, little work explicitly...
The problemArtificial Intelligence (AI) is often touted as healthcare's saviour, but its potential will only be realised if developers and providers consider the whole clinical context AI's place within it.One of many aspects that question liability.Analysis responsibility attributions in complex, partly automated socio-technical systems has identified risk nearest human operator may bear brunt for overall system malfunctions. 1As we move towards integrating AI into healthcare systems, it...
Abstract Since December 2019, more than 79,000 people have been diagnosed with infection of the Corona Virus Disease 2019 (COVID-19). A large number medical staff were dispersed for Wuhan city and Hubei province to aid COVID-19 control. Psychological stress, especially vicarious traumatization (VT) caused by pandemic, should not be ignored. To address this concern, study employed a total 214 general public (GP) 526 nurses evaluate VT scores via mobile app-based questionnaire. Results showed...
Abstract Levels of cadmium (Cd), arsenic (As), mercury (Hg), lead (Pb), iron (Fe), and zinc (Zn) were investigated in 285 samples 9 species edible fungi ( Lentinus edodes , Auricularia auricula Pleurotus ostreatus Tremella fuciformis Flammulina velutipes Agrocybe chaxinggu Armillaria mellea Agaricus bisporus Pholiota nameko ), which collected from markets Beijing, China. In addition, culture substrates 7 cultivation bases to examine the role substrate trace metal accumulation. Trace...
Machine Learning (ML) is now used in a range of systems with results that are reported to exceed, under certain conditions, human performance. Many these systems, domains such as healthcare , automotive and manufacturing, exhibit high degrees autonomy safety critical. Establishing justified confidence ML forms core part the case for systems. In this document we introduce methodology Assurance use Autonomous Systems (AMLAS). AMLAS comprises set patterns process (1) systematically integrating...
Establishing confidence in the safety of Artificial Intelligence (AI)-based clinical decision support systems is important prior to deployment and regulatory approval for with increasing autonomy. Here, we undertook assurance AI Clinician, a previously published reinforcement learning-based treatment recommendation system sepsis.As part assurance, defined four hazards sepsis resuscitation based on expert opinion existing literature. We then identified set unsafe scenarios, intended limit...
Background: Hemorrhagic fever with renal syndrome (HFRS), a rodent-borne disease caused by different species of hantaviruses, is widely endemic in China. Shandong Province one the most affected areas. This study aims to analyze epidemiological characteristics HFRS, and predict regional risk Province. Methods: Descriptive statistics were used elucidate HFRS cases from 2010 2018. Based on environmental socioeconomic data, boosted regression tree (BRT) model was applied identify important...
Objective This study aimed to evaluate the clinical value of dynamic monitoring neutrophil/lymphocyte ratio (NLR), APACHE II (Acute Physiology and Chronic Health Evaluation II) score, Sequential Organ Failure Assessment (SOFA) score in predicting 28-day prognosis drug resistance patients with bloodstream infection Acinetobacter baumannii–calcoaceticus complex (Abc complex). Patients methods In this research, individuals admitted Tianjin Medical University General Hospital from January 2017...
Weaning patients from ventilation in intensive care units (ICU) is a complex task. There growing desire to build decision-support tools help clinicians during this process, especially those employing Artificial Intelligence (AI). However, built for purpose should fit within and ideally improve the current work environment, ensure they can successfully integrate into clinical practice. To do so, it important identify areas where may aid clinicians, associated design requirements such tools....
Abstract Acoustic metasurfaces have shown indispensable abilities for wave control with subwavelength resolution and enabled exotic sound functions unavailable using naturally occurring materials. To further improve the utilization of field resources enhance coverage acoustic signals, full‐space hold great potential independent tailoring reflected transmitted waves. However, achieving dynamic manipulation full in a programmable fashion has remained inaccessible until now. Here,...
Sepsis, a life-threatening illness, is estimated to be the primary cause of death for 50,000 people year in UK and many more worldwide. Managing treatment sepsis very challenging as it frequently missed at an early stage optimal not yet clear. There are promising attempts use Reinforcement Learning (RL) learn strategies treat patients, especially administration intravenous fluids vasopressors. However, some RL agents only take current state patients into account when recommending dosage This...
Autonomy does not subvert existing safety processes, but they must be supplemented with methods that address autonomy's challenges, especially where perception and decision-making tasks are implemented machine learning.We present an approach to the of autonomous systems, building on complementing established engineering methods.
In the standard interaction model of clinical decision support systems, system makes a recommendation, and clinician decides whether to act on it.However, this can compromise patient-centeredness care level involvement.There is scope develop alternative models, but we need methods for exploring comparing these assess how they may impact decision-making.Through collaborating with clinical, AI safety, HCI experts, patient representatives, co-designed number human-AI models decision-making.We...
Hand-foot-mouth disease (HFMD) is a global public health issues, especially in China. It has threat the of children under 5 years old. The early recognition high-risk districts and understanding epidemic characteristics can facilitate sectors to prevent occurrence HFMD effectively.Descriptive analysis was used summarize characteristics, spatial autocorrelation space-time scan were utilized explore distribution pattern identify hot spots with statistical significance. result presented...
Nowadays, people are usually involved in multiple heterogeneous social networks simultaneously. Discovering the anchor links between accounts owned by same users across different is crucial for many important inter-network applications, e.g., cross-network link transfer and recommendation. Many supervised models have been proposed to predict so far, but they effective only when labeled abundant. However, real scenarios, such a requirement can hardly be met most unlabeled, since manually...