Finding combinatorial drug repositioning therapy for Alzheimer's disease and related dementias

寻找治疗阿尔茨海默病和相关痴呆症的组合药物重新定位疗法

基本信息

批准号：
10598207
负责人：
Xiaoqian Jiang
金额：
$ 30.27万
依托单位：
UNIVERSITY OF TEXAS HLTH SCI CTR HOUSTON
依托单位国家：
美国
项目类别：
财政年份：
2020
资助国家：
美国
起止时间：
2020-04-01 至 2025-03-31
项目状态：
未结题

来源：
https://reporter.nih.gov/project-details/10598207
关键词：
Accountability Address Algorithms Alzheimer&apos s Disease Alzheimer&apos s disease related dementia Alzheimer&apos s disease therapy Area Artificial Intelligence Black Populations Climacteric Clinical Cost Savings Custom Data Data Analyses Data Collection Data Protection Data Set Decision Making Development Diabetes Mellitus Diagnostic Drug Combinations Economic Factors Economics Ensure Equilibrium Ethics Goals Health Health Services Accessibility Healthcare Healthcare Industry Heterogeneity Hypertension Incidence Knowledge Lead Learning Legal Machine Learning Measurement Medicine Methodology Methods Minority Groups Modeling Outcome Parents Patients Performance Pharmaceutical Preparations Phase Play Population Procedures Process Protective Agents ROC Curve Recommendation Research Risk Sampling Sensitivity and Specificity Source Structure Subgroup System Techniques Technology Training Trust Work algorithmic bias base burden of illness causal model clinical application combinatorial design drug repurposing health disparity human error improved innovation large datasets machine learning model novel opportunity cost outcome prediction parity patient population patient subsets precision medicine sample collection social social health determinants social implication software development tool treatment planning trustworthiness underserved community user-friendly

项目摘要

Summary Artificial intelligence and machine learning (ML) models are becoming increasingly popular in clinical applications. If we allow these “autonomous” ML models to make recommendations for clinical decisions, it is important to ensure that they do introduce algorithmic unfairness (e.g., differences in the burden of disease or opportunities of treatment for different populations). We propose novel technological solutions to mitigate algorithmic unfairness. We will address two major types of data biases (subgroup and representation) to reduce their negative impact on ML models. Based on contextual information and novel causal inference techniques, we will identify potential outliers and task-irrelevant confounders and address them with customized mitigation strategies (e.g., down-sampling and factor reduction) to avoid learning erroneous information that might lead to health disparities. In addition, we will propose FairAUC (a new optimization mechanism) to maximize prediction accuracy while considering fairness by design. As opposed to post-hoc fairness rectification approaches, our method will automatically consider both objectives in the training phase to strike the optimal balance between accuracy and fairness.

概括人工智能和机器学习（ML）模型在临床应用。如果我们允许这些“自主” ML模型提出建议临床决定，重要的是要确保它们确实引入算法不公平（例如，疾病伯恩的差异或不同人群的治疗机会）。建议减轻算法不公平的新型技术解决方案。我们将解决两个主要类型的数据偏见（亚组和表示），以减少其对ML的负面影响型号。基于上下文信息和新颖的因果推理技术，我们将确定潜在的异常值和任务 - 涉及混杂因素，并通过自定义缓解措施来解决它们策略（例如，下采样和减少因素），以避免学习错误的信息这可能会导致健康差异。此外，我们将提出Fairauc（一种新的优化机制），以最大程度地提高预测准确性，同时考虑设计公平。而不是反对为了事后公平纠正方法，我们的方法将自动考虑同时考虑训练阶段的目标，以实现准确性和公平之间的最佳平衡。