arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2310.04997cs.CY
收藏

揭示眼科人工智能生命周期中的偏见并规避陷阱:一篇综述

Unmasking Biases and Navigating Pitfalls in the Ophthalmic Artificial Intelligence Lifecycle: A Review

Luis Filipe Nakayama, João Matos, Justin Quion, Frederico Novaes, William Greig Mitchell, Rogers Mwavu, Ju-Yi Ji Hung, Alvina Pauline dy Santiago, Warachaya Pha… 展开作者

Luis Filipe Nakayama, João Matos, Justin Quion, Frederico Novaes, William Greig Mitchell, Rogers Mwavu, Ju-Yi Ji Hung, Alvina Pauline dy Santiago, Warachaya Phanphruk, Jaime S. Cardoso, Leo Anthony Celi

AI总结:

本综述梳理眼科AI应用发展背景,将AI生命周期划分为七个核心步骤,逐一剖析各环节潜在伤害风险并提出缓解策略,为AI在眼科的安全广泛部署提供参考。

摘要:

在过去二十年中,数据可用性、计算能力以及新型建模技术呈指数级增长,推动了人工智能(AI)应用领域的研究兴趣、投资与成果不断扩张。得益于远程医疗筛查项目的兴起及辅助影像技术的应用,眼科学成为有望从AI中获益的众多领域之一。然而,在AI实现广泛部署前,仍需开展更多工作以规避AI生命周期内的各类陷阱。本综述文章将AI生命周期划分为七个步骤:数据收集、模型任务定义、数据预处理与标注、模型开发、模型评估与验证、部署,以及最终的部署后评估、监测与系统重新校准,并深入探讨了每个步骤中存在的伤害风险及相应的缓解策略。

英文摘要:

Over the past two decades, exponential growth in data availability, computational power, and newly available modeling techniques has led to an expansion in interest, investment, and research in Artificial Intelligence (AI) applications. Ophthalmology is one of many fields that seek to benefit from AI given the advent of telemedicine screening programs and the use of ancillary imaging. However, before AI can be widely deployed, further work must be done to avoid the pitfalls within the AI lifecycle. This review article breaks down the AI lifecycle into seven steps: data collection; defining the model task; data pre-processing and labeling; model development; model evaluation and validation; deployment; and finally, post-deployment evaluation, monitoring, and system recalibration and delves into the risks for harm at each step and strategies for mitigating them.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑