SBIR Phase I: An Artificial Intelligence-Inspired Computing Application for Detecting the Early Onset of Pneumonia (COVID-19)

这项小企业创新研究（SBIR）第一阶段项目的更广泛影响/商业潜力是开发一种人类-人工智能（AI）计算应用程序，用于检测早期发作的肺炎。它对COVID-19并发症特别有用；临床研究已经确定COVID-19与肺炎之间存在显著关联，研究观察到高达70.1%的老年COVID-19患者被诊断为肺炎。这项工作旨在通过远程健康监测收集生理数据和症状决定因素，并将其传输到我们基于人工智能的云应用程序，以检测与肺炎相关的模式。通过在医院环境外进行无障碍监测，这一拟议的应用程序可以在病情恶化的最早迹象时对患者进行护理管理，并作为一种补充诊断工具，对这种危及生命的疾病的一般检测非常有用，特别是对COVID-19患者。这个小企业创新研究第一阶段项目提出，通过开发一种预测算法策略，为有肺炎风险的门诊COVID-19患者提供最佳护理，解决当前COVID-19大流行带来的一些公共卫生挑战。拟议的应用程序使用多模态数据集（生理和用户输入）与基于云的协作AI集成。拟议的应用程序将包括一个基于云的预测分析单元，该单元接收来自远程健康监测的多模式信息，识别肺炎的早期发作，并向医疗保健提供者发出警报。提出的工作的关键创新之一是动态分析单元的动态自适应方法，该方法对低维数据进行分类，并根据需要通过包括实时患者症状来扩展维度模型。这种方法为人工智能提供了一种新的协作方法，在这种方法中，COVID-19患者在系统决策过程中积极协作。系统将自动决定应该交互式地向患者提出什么要求，以提高预测的准确性。该方法将提供增强的临床信息，允许临床医生在算法检测到与早发性肺炎相关的模式时监督快速反应。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase I project is to develop a Human-Artificial Intelligence (AI) computing application for detecting the early onset of pneumonia. It can be particularly useful for complications of COVID-19; clinical studies have identified a significant association between COVID-19 and pneumonia, with studies observing up to 70.1% of older COVID-19 patients diagnosed with pneumonia. This work aims to collect physiological data and symptomatic determinants using remote health monitoring and stream it to our AI-based cloud application to detect the pattern associated with pneumonia. Through accessible monitoring outside the hospital setting, this proposed application affords patient care management at the earliest signs of worsening and serving as a complementary diagnostic tool, useful for general detection of this life-threatening ailment - particularly for COVID-19 patients. This Small Business Innovation Research Phase I project proposes to address some of the public health challenge of the current COVID-19 pandemic by developing a predictive algorithm strategy for providing optimal care for outpatient COVID-19 patients at risk of pneumonia. The proposed application uses a multimodal dataset (physiological and user inputs) integrated with collaborative cloud-based AI. The proposed application will include a cloud-based predictive analytics unit that receives multimodal information from Remote Health Monitoring, identifies the early onset of pneumonia, and alerts healthcare providers. One of the proposed work’s key innovations is the dynamic analytics unit’s dynamically adaptive approach that performs classifications on low-dimensional data and expands the dimensionality model as needed by including real-time patient symptoms. This approach affords a novel collaborative approach to AI, where the COVID-19 patient is actively collaborating in the system decision-making process. The system will automatically decide what should be interactively requested from the patient to enhance prediction accuracy. The approach will provide enhanced clinical information, allowing for clinician oversight for rapid response when the algorithm detects a pattern associated with the early onset of pneumonia.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.