您当前的位置:首页 >> 人物
【科技】深度学习能做什么? | Day 753




In a paper1 published on 7 January in Nature Medicine, researchers describe thetechnology behind the diagnostic aid, a smartphone app called Face2Gene. Itrelies on machine-learning algorithms and brain-like neural networks toclassify distinctive facial features in photos of people with congenital andneurodevelopmental disorders.  

Nature Medicine 

自然医学 (杂志)(Nature Medicine, ISSN 1078-8956)是一个学术期刊, 发表生物医学领域,包括基础研究和早期临床研究的研究文章, 综述, 研究新闻及评论。文章议题包括癌症,心血管疾病,基因治疗,免疫学,免疫疫苗和神经科学。该杂志旨在发表“展示新颖疾病过程的机理, 有直接证据的生理相关的研究结果”的原创论文

machine-learning algorithms


brain-like neural networks




Using the patterns that it (the app) infers from the pictures, the model homes inon possible diagnoses and provides a list of likely options. Doctors have beenusing the technology as an aid, even though it's not intended to provide definitivediagnoses. 

1. model 算法;模型 

2. home in on 匹配;定位

Researchers at FDNA, a digital-health company in Boston, Massachusetts, first trained the artificial intelligence (AI) system to distinguish Cornelia de Lange syndrome and Angelman syndrome — two conditions with distinct facial features — from other similar conditions. They also taught the model to classify different genetic forms of a third disorder known as Noonan syndrome. 

FDNA 是马萨诸塞州波士顿的一家数字医疗公司。公司的研究者们首先训练人工智能系统来区分德朗热综合征和天使人综合征,这两种疾病患者都有区别于其他疾病的明显面部特征。同时,研究者们还训练该模型对另一种疾病——努南综合征的不同基因形式进行分类。


Then the researchers, led by FDNA chief technology officer Yaron Gurovich, fedthe algorithm more than 17,000 images of diagnosed cases spanning 216 distinctsyndromes. When presented with new images of people’s faces, the app’s bestdiagnostic guess was correct in about 65% of cases. And when considering multiple predictions, Face2Gene's top-ten list contained the right diagnosis about90% of the time. 

接下来,在公司首席技术官 Yaron Gurovich 的带领下,研究者们给算法输入涵盖 216 种不同综合征的 17000 多张确诊病例的图像。在用新面孔进行测试时,该 APP 的最佳诊断猜测准确率达到了 65%。如果考虑多个预测结果,则 Face2Gene 的 top-10 准确率可以达到约 90%。


The program’s accuracy has improved slightly as more healthcare professionalsupload patient photos to the app, says Gurovich. There are now some 150,000images in its database.

Gurovich 表示,随着更多医疗专家将病人的照片上传到该 APP,该项目的准确率也得到略微提高。现在该项目的数据库中大约有 15 万张照片。

But the algorithm is only as good as its training data set. 

— Silos and Bias 

It does raise a number of ethical and legal concerns, say researchers. Theseinclude ethnic bias in training data sets and the commercial fragmentation ofdatabases, both of which could limit the reach of the diagnostic tool.The algorithm is only as good as its training data set — and there’s a risk,especially where rare disorders that affect only small numbers of people worldwideare concerned, that companies and researchers will begin to silo and commodifytheir data sets.And ethnic bias in training data sets that contain mostly Caucasian faces remains aconcern. A 2017 study of children with an intellectual disability found that whereasFace2Gene’s recognition rate for Down syndrome was 80% among white Belgianchildren, it was just 37% for black Congolese children. With a more-diversetraining data set, however, the algorithm’s accuracy for African faces improved,showing that more-equitable representation of diverse populations is achievable.

1. training data set 训练数据集 

2. commercial fragmentation 商业竞争


3. Silo Effect 谷仓效应 

指企业内部因缺少 沟通,部门间各自为政,只有垂直的指挥系统,没有水平的协同机制,就象一个个的谷仓,各自具有独立的进出系统,但缺少了谷仓与谷仓之间的沟通和互动。这类情况下各部门之间未能建立共鸣而没法和谐运作。

AI 不止于此




在 “分子厨房 ”里创造新物质 

德国明斯特大学的研究生Marwin Segler和小伙伴们正尝试用AI简化分子合成的过程——让AI学会从数百个积木般的零件库中挑选原子等,并依据数千个合成规则进行连接它们。Segler团队用合成40个不同分子对程序进行测试,并与传统的分子设计程序进行比较,结果证明AI比传统方法快了95%。Segler希望用这种方法改进药品的生产流程。











请小编们喝咖啡 ►