经济学人精读 The Economist [56]
选自 |January 20 2018 | Science and Technology | 科技板块
更多英语阅读听力口语内容，请关注
微信公众号MyEnglishTrip
我是Eva
一个认真学习英语的美少女
#Eva导读#
在美国，计算机已经被用来正确的预测一个人是否应该接受协助保释和判决的决定很多年了。支持者这一做法的人认为，经过大量数据训练的算法，也对囚犯是否会再次犯罪作出判断。研究人员就此问题做了实验。实验结果显示，计算机算法的判断与人的判断的准确率是相同的。既然计算机算法的准确率并没有超过人的判断，那么计算机算法的价值就很难评判。
#以上，个人总结和理解，欢迎批评指正，欢迎留言讨论
#有输出才有进步
Computers and criminal justice[计算机与刑事审判]
Algorithm’s dilemma[算法的困境]
Are programs better than people at predicting recidivism?[程序会比人更擅长预测累犯吗]
IN AMERICA, computers have been correctly predicted whether someone used to assist bail[保释]and sentencing decisions for many years[在美国，计算机已经正确的预测一个人是否应该接受协助保释和判决的决定很多年了]. Their proponents argue that the rigorous[缜密的]logic of an algorithm, trained with a vast amount of data, can make judgments about whether a convict[囚犯]will reoffend[再次犯罪]that are unclouded by human bias[支持者认为，经过大量数据训练的算法，逻辑缜密，可以在不受人类偏见的影响下，对囚犯是否会再次犯罪作出判断]. Two researchers have now put one such program, COMPAS, to the test[现在，有两位研究人员将试验这样的一个叫做COMPAS的系统]. According to their study, published in ScienceAdvances, COMPAS did neither better nor worse than people with no specialexpertise[根据他们发表在《科学进展》杂志中的研究，COMPAS比那些没有特殊专业知识的人相比，既没有更好也没有更差].
Julia Dressel and Hany Farid of Dartmouth College in New Hampshire selected 1,000 defendants at random from a database of 7,214people arrested in Broward County, Florida between 2013 and 2014, who had been subject to COMPAS analysis[来自新罕布什尔州达特茅斯学院的JD和HF，在佛罗里达州B郡2013-2014年间被逮捕的7214人的数据库中，随机选取了1000名被告人]. They split their sample into 20 groups of 50[他们将样本分成20组，每组50人]. For each defendant they created a short description that included sex, age and prior convictions, as well as the criminal charge faced[对每一位被告人，他们都添加了一个简短的描述，包括性别、年龄、前科和面临的刑事控告].
They then turned to[开始使用]Amazon Mechanical Turk, a website which recruits volunteers to carry out small tasks in exchange for cash[之后，他们开始使用亚马逊土耳其机器人，一个可以有偿雇佣志愿者完成小任务的网站]. They asked 400 such volunteers to predict, on the basis of the descriptions, whether a particular defendant would be arrested for another crime within two years of his arraignment (excluding any jail time he might have served)—a fact now known because of the passage of time[他们会要求400名这样的志愿者，基于描述，预测某个被告人是否会在他传讯的两年之内因再次犯罪被捕（不包括他可能服刑的监狱时间）——因为时间已经过去，所以现在已经知道的事实]. Each volunteer saw only one group of 50people, and each group was seen by 20 volunteers[每一位志愿者只会看到一组50人，每一组人会被20位志愿者看到]. When Ms Dressel and Dr Farid crunched the numbers[（快速大量的）处理数字], they found that the volunteers correctly predicted whether someone had been rearrested 62.1% of the time[当D和F处理这些数字时，他们发现志愿者有62.1%的概率正确预测一个人是否会再次被捕]. When the judgments of the 20 who examined a particular defendant’s case were pooled, this rose to 67%[当20组检测某一个被告人案例的判断集合在一起时，正确率提升到了67%]. COMPAS had scored 65.2%—essentially the same as the human volunteers[COMPAS的得分是65.2%——实际上与人类志愿者分数相同].
To see whether mention of a person’s race (a thorny[棘手的]issue in the American criminal-justice system)would affect such judgments, Ms Dressel and Dr Farid recruited 400 more volunteers and repeated their experiment, this time adding each defendant’s race to the description[为了检测提到一个人的种族（在美国刑事审判系统中一个棘手的问题）是否会影响这样的判断，D和F重新雇佣了400多名志愿者，并重复了他们的试验，这一次将被告人的种族加入到描述中]. It made no difference[结果没有区别]. Participants identified those rearrested with66.5% accuracy[参与者以66.5%的准确率鉴别出了那些再次被捕的人].
All this suggests that COMPAS, though not perfect, is indeed as good as human common sense at parsing pertinent[有关的]facts to predict who will and will not come to the law’s attention again[所有这些表明，COMPAS，尽管不完美，但在预测谁将会或者不会遇到法律的再次警告时，与人类分析有关事实的常识性判断一样好]. That is encouraging[这个结果是鼓舞人心的]. Whether it is good value, though, is a different question, for Ms Dressel and Dr Farid have devised[发明]an algorithm of their own that was as accurate as COMPAS in predicting rearrest when fed the Broward County data, but which involves only two inputs—the defendant’s age and number of prior convictions[然而，这是否是一个有利的价值是一个难题，D和F已经发明了自己的一套算法，当用B郡的数据预测再次犯罪时，与COMPAS一样准确，但是仅需要两个输入——被告的年龄和前科的次数].
As Tim Brennan, chief scientist at Equivant, which makes COMPAS, points out, the researchers’algorithm, having been trained and tested on data from one and the same place, might prove less accurate if faced with records from elsewhere[尽管创造COMPAS的公司Equivant的首席科学家TB指出，研究人员的算法是通过单一且同一地点的数据训练和检测，在遇到其他地方的记录时，可能检测出的准确率更低]. But so long as the algorithm behind COMPASitself remains proprietary, a detailed comparison of the virtues of the two is not possible[但是，只要支持COMPAS自身的算法仍然是专有的，那么详细比较两者的优点是不可能的].
Jan 22 | 519 words
更多英语阅读听力口语内容，请关注
微信公众号MyEnglishTrip
我是Eva
一个认真学习英语的美少女