人工智能如何帮医生拯救生命

人工智能已经展现出了帮助医生预测病人是否易患阿尔茨海默病（Alzheimer’s）等慢性疾病的前景。

不过，在上周三圣迭戈举办的《财富》头脑风暴健康大会（Fortune’s Brainstorm Health conference）上，辉瑞制药（Pfizer）的副总裁和基因组科技部门主管莫滕·瑟高表示，尽管医疗业发展迅速，但在推出人工智能疗法和药品上还仍然处于初期阶段。

瑟高表示，辉瑞多年来都在利用机器学习等人工智能技术筛选数据，帮助研究新药配方（尤其是多种药物的组合），并确定临床试验的最佳参与者。他在会上谈到人工智能时说：“在一些情况下，这没什么新鲜的。”

然而，他解释道，新鲜之处在于越来越大的数据洪流，如医疗设备获取的基因组数据和传感器数据。这种数据流让我们更难理清隐藏在其中的关键联系，从而帮助研究人员发现新疗法。

目前，辉瑞使用了深度学习来挖掘电子健康记录和实验室数据。这种方法曾因谷歌（Google）用以训练计算机识别照片中的猫而流行起来。瑟高表示，辉瑞可以借此更好地了解自体免疫和脂肪肝等疾病的进展。

瑟高称，这些深度学习技术已经展现了在大规模人群中寻找疾病模式的前景，不过最终的目标是帮助个体病人。

辉瑞还与IBM合作，利用其沃森（Watson）的数据处理技术进行药学研究。公司同时也在与谷歌、亚马逊（Amazon）和其他云计算供应商合作，采用他们的对应技术。

瑟高相信，许多云计算供应商在未来都会拥有可供医药公司研究和开发使用的人工智能技术。不过他也表示：“当然，这不会在短期内发生。”

该领域的迅速创新总有一天能够帮助人类利用基因组、蛋白质和医学成像数据预测和诊断疾病，而联邦法规目前还没有跟上创新的步伐。不过瑟高对此充满希望。根据辉瑞与监管者的会面，他认为联邦药品管理局（Federal Drug Administration）在人工智能辅助医学治疗方面“很开明”。（财富中文网）

译者：严匡正

Artificial intelligence has shown promise in helping doctors predict which patients may be susceptible to chronic diseases like Alzheimer’s.

But despite the rapid advances, the healthcare industry is still in the early days of rolling out AI-powered treatments and drugs, Morten Sogaard, Pfizer’s vice president and head of genome sciences and technologies, said at Fortune’s Brainstorm Health conference in San Diego on Wednesday.

Pfizer (pfe, -0.42%) has been using AI techniques like machine learning for years to sift through data, help research new drug compounds (essentially the combination of multiple drugs), and determine the best participants for clinical trials, he said. “In some cases, it is nothing new,” Sogaard said about AI in healthcare.

What is new, however, is the rising flood of information like genomic data and sensor data from medical devices, he explained. This influx has made it more difficult to understand key connections that could help researchers discover new treatments.

Currently, Pfizer is using deep learning, which Google (goog, +1.04%) helped to popularize as a way to train computers to recognize cats in photos, to mine electronic health records and lab data. By doing so, Pfizer can better understand how ailments like autoimmune and fatty liver diseases progress, he explained.

Sogaard said that these deep learning techniques have shown promise in finding disease patterns across large groups of people, but the ultimate goal is to eventually help individual patients.

Pfizer has also partnered with IBM (ibm, -0.30%) to use its Watson data-crunching technology in pharmaceutical research. But the company is also open to partnering with Google, Amazon (amzn, -0.62%), and other cloud-computing providers to incorporate their own respective technologies.

Sogaard believes a handful of cloud computing providers will have AI technologies that drug companies could eventually use for research and development. However, “it will not all happen the day after tomorrow, of course,” he said.

Federal regulations have not yet caught up to the rapid pace of innovation that could one day help predict and diagnose diseases using a combination of genomic, protein, and medical imaging data. But Sogaard is hopeful, and based on Pfizer’s meetings with regulators, he believes the Federal Drug Administration is “open-minded” to AI-assisted medical treatment.