10-2110-2110-2110-2110-2110-2110-2110-2110-2110-21最新范文01-0101-0101-0101-0101-0101-0101-0101-0101-0101-0101-0101-0101-0101-0101-01英语翻译_名师谈能力考：最后一个月，N2单项如何提高？_沪江英语
分类学习站点
名师谈能力考：最后一个月，N2单项如何提高？
编辑点评：“距2015年7月日语能力考还有很少的时间！”现在不知有多少考生把这句话挂在书桌案头。的确，能力考越来越贵，不一次通过真心伤不起……于是，本期沪江专访请来了青岛华兴日语学校的沈虎林老师，为大家讲解N2最后备考要点！ [1] 登场人物：沈虎林 [2]
&距2015年7月日语能力考还有很少的时间！&现在不知有多少考生把这句话挂在书桌案头。的确，能力考涨价报名费越来越贵，考个N2的居然要450RMB！如果不能一次性通过那还真是伤不起&&但话说回来，想通过N2也并不是那么容易，一个月时间努力一把就通过了，不采取任何措施可能就挂了。很多童鞋都在寻求最后的备考方式&定心丸&。
于是本期【沪江专访】连线青岛华兴日语学校中高级日语部部长沈虎林老师，沈老师仔细讲了通过最后一个月复习，N2词汇、语法、听力、阅读四个单项分别如何提高，一起来听听吧~
&&&&&&&&&& & & &
[本期登场人物]
嘉宾：沈虎林
多年从事日语翻译和日语教学工作，现任青岛华兴日语学校中高级日语部部长，从事中高级日语教学以及能力考试辅导。对能力考试出题基准有着深入细致的研究。也是青岛华兴日语学校一二级全日制班主讲老师。在新能力考试中联创佳绩。
【N2四大单项如何在短期内提高？】
①词汇：最后一个月，复习有所侧重
如果想在考试中有所侧重，多拿分的话，沈老师认为还是以双汉字、动词、接头、接尾词为主要复习点，其它为辅。因为在这几项在历年考试中出题率占80%左右。能做到有的放矢，事半功倍。很多同学单词记两三遍还是个模糊记忆，考试时也不会拿来用，最终成了&死单词&，关于这个问题，沈老师给出以下建议：
（1）记单词需要反复性和灵活性。反复性指的就是通过反复的记忆去熟练单词的含义。灵活性指的是不要死记单词，要去组词或造句来理解后记。特别是惯用句。比如&相談に乗る、話が弾む、時間を潰す、お金を下ろす&等。如果单独地去记各自的词义，有时候还是不能正确地选择，所以以惯用词的形式去记忆，又省时间有能够牢固地掌握，而且还能再实际会话中运用。
（2）具体抽象掌握法：具体抽象掌握法指的是先掌握单词具体的含义，然后在去记其它抽象的含义。这类词以动词为最多，其次是副词，名词。比如说 &遮る&这个动词，原意是遮挡，遮蔽的意思。&月は雲に遮られた&月亮被云彩挡住了。那么我们通过遮挡这个词义就能联想到它还有阻挡，打断的意思。 &話を遮る&就表示打断别人的说话。像这样先去掌握一个词的基本用法，在这个基础上在去掌握其他的用法，这样记起来比较容易，又能够清晰地记住。
（3）最后就是强化训练法：通过反复的做模拟练习题来熟练单词的用法和熟悉考试的题型，来提高对单词的使用语感和掌握每个题型的做题方法。
②语法：冲刺后最易&抓&分，记语法也有诀窍！
沈老师认为，词汇、语法、读解、听力四项当中，语法相对来说比较容易掌握，因为大多数语法都是惯用句，只要充分的理解这个语法点无论什么样的句子用法都可以举一反三。但是很多同学反映语法点太多，易混还记不住。这里沈老师也给出一些建议：
（1）首先，语法不能死记硬背。N2中的所谓的语法其实说的是惯用句型。很多人认为惯用句就是习惯性用法。只能硬记，这种观点本身就是错误的，其实99%的惯用句都是由单词和语法组成的。比如说：&～にしたがって（に+したがう+て）、といっても（と+いう+ても）、ために（ため+に）&等，都是由单词和语法组成的，所以在记惯用句的时候首先要去分解句型，先确定这个句型是由什么组合而成，然后根据单词和语法的含义来理解这个句型，这样才能真正的掌握这个句型的用法。其中有一部分文语的残留，其实也一样。比如：&～ざるをえない（ざる+を+える+ない）&这个句型就是典型的文语用法。这里的　&ざる&其实是&ない&的文语否定助动词&ぬ&的连体型。在文语语法中助动词后面是可以接宾格助词&を&这样就可以理解否定的否定，其实是肯定句。像这样的句型能力考试中也有很多，我们都把它分解掌握的话就不会忘记也不会混淆。
（2）其次，不要用中文单词或句型来记日语惯用句。因为汉语是独立语，本身一个字就是一个词，所以每个字都有很多种词的搭配和用法。但日语中一个词就是一个词，如果用汉语的形式去掌握日语，一旦遇见汉语的用法一样，但日语的用法不一样的句型时，就无法分辨，最终导致使用错误。比如说：&～うちに、～際に、～ところに、～最中に&这4个句型日语的用法其实完全不一样，但用汉语表达的时候都可以翻译为&在.......的时候&。这样如果我们只记住了它汉语的意思，那么在做日语题的时候就会发现这些句型用汉语都可以用，最终导致没有能够正确地掌握这些句型，而变得混淆。
③听力：了解听力题型，考试百战百胜
沈老师在专访中谈到，练习听力最有效的方法就是首先要摸清听力题型。如果连考试会出现什么题型都不清楚一味盲目去听的话效果不会很明显。因此，要有针对性地去练习，现在N2考试听力一共分为五个部分：
（1）其中第一、二题是有选项题。那么做这样的题首先在播放内容之前要把4个选项快速看一遍并记住内容，然后听内容，一边听一边逐一排除答案。像这样的题提问方式基本上是围绕着&1、これから何をしますか；2、～はいつですか；3、～はどうしてですか；4、～は何をしなければなりませんか&等。平时练习的时候学会在原文中找答案通常以什么形式出现，通常在文章的哪个部分出现，然后有意识的做上标示，进行归纳总结，最后就会发现大多数的类似题答案基本上都会在同样的位置上体现。这样我们在考试的时候即使前段没听懂，但到了快出现答案的时候仔细听的话有可能就会找出答案。
（2）题型三应该是听力考试中的难点之一。之所以难是因为没有选项，而且还是先听内容然后设问，因此很多考生原本就对内容没完全听懂，而且又不知道到底要问什么，所以导致没有方向感，不知道怎么选好。其实像这样的题基本都是围绕话题的，所以听的时候第一句话一般都是要说的话题，最后几句一般都是对这个话题的概要和观点的讲述。所以建议考生多去注意一些说话人总结性和概括性的句子。比如：&～のではないでしょうか。～と思います。～と考えられます。&等的相关句型表现，这些句型前面的内容通常是作者想要说的观点。
（3）题型四（即答题）：这类的题题量多，而且口语用法比较多，所以对平时不怎么说日语的考生来说有一定的难度。首先，要去掌握一些日语口语的缩略型的用法，多总结一些常用日语缩略语。其次是通过练习要去掌握会话的语感和日本人说话的习惯，这一点很重要。
（4）题型五：这类的题型基本上就是四选一或者是四选二的题。做这类题的时候建议最好是记笔记。把内容里出现的选项逐一的记下来，然后用排除法或对应法选择即可。
N2各专项冲刺班：&&&&&&
④阅读：生词满篇，做阅读也有技巧
很多考生确实对读解感到棘手，因为读解是考核考生的一个综合能力题，尤其是如何在考场那么短的时间里看完这么多生词满篇的文章是个难题。沈老师分析称，N2的读解共分5个部分：
（1）题型一是短篇（大约200-300之间，共5道题）做题时间：3-4分钟。首先为了节省时间还是要先看问题，短篇问题的提问方式大体分为主观题和客观题。主观题主要说的是作者的观点和整篇文章的中心思想。提问方式大体是&この文章で筆者が言いたいことはなにか；～とあるが、筆者はどう思っていますか。&等。 看完4个选项后通常不符合道理的，或者是违背恒常规律的一般都不会是答案，还有就是谓语使用命令型的或者是肯定某个观点的选项通常也不会作为答案出现，原因是主张或观点本身就是以个人的立场出发阐述的，所以一般不会用命令或很肯定的方式来表达自己的观点或主张，所以类似这样的选项一般都不会作为答案。（有些个别的问题有可能会以这样的形式出现）另外主观题一般也不会把文章里的某一部分内容原班不动拿出来作为答案，通常都是替换说法。也就是说文章里的内容和选项的内容表达的意思一样，但写法不一样，或单词、句型的使用不一样。那么我们就在选项中找和文章内容意思一致的句子，这样的方式会提高选择的准确率。在一个是客观题，问题方式大多是&文章の内容とあっているものはどれですか。&这类题也是主要找替换说法的句子就可以。注意点：千万不要把自己的想法或观点加进去。这是很多考生选错的主要原因之一。通常选项里会出现一些很符合真理，道理的句子，但记住文章里没有出现就不是答案。所以答案一定会再文章中体现，再抽象的题也如此。
（2）题型2是中篇（700百字左右，共3篇，每篇3个问题）做题时间：7-8分钟；题型4是长篇(1000字左右，共1篇，3个问题）做题时间：15分钟。做中长篇文章时关键要找出文章的话题，就是说这篇文章主要写的内容是围绕着什么写的，找出主题后在文章中找答案相对正确率会很高。因为时间有限，所以先看看文章的出处有时候也能知道主题，阅读文章的第一部分和最后部分，先确定这篇文章的主题，最后按照问题找答案。注意时间的控制。
（3）第三题综合理解题和第五题情报检索题相对比较容易一些。因为这类的题不需要阅读理解，只要在文章当中寻找答案即可。注意点就是不要漏掉任何一个条件，仔细的寻找答案。
接下来沈老师还会谈他对能力考改革的看法，
相关推荐：
想要和小伙伴一起备考吗？可以加入沪江官方备考QQ群：，期待你的加入。翻译翻译有关翻译的一切都在这里关注专栏更多置顶文章最新文章{&debug&:false,&apiRoot&:&&,&paySDK&:&https:\u002F\\u002Fapi\u002Fjs&,&wechatConfigAPI&:&\u002Fapi\u002Fwechat\u002Fjssdkconfig&,&name&:&production&,&instance&:&column&,&tokens&:{&X-XSRF-TOKEN&:null,&X-UDID&:null,&Authorization&:&oauth c3cef7c66aa9e6a1e3160e20&}}{&database&:{&Post&:{&&:{&title&:&经济学人官译：公平竞争&,&author&:&iqlake&,&content&:&\u003Cp\u003EGlobal economy\u003C\u002Fp\u003E\u003Cp\u003E\u003Ca href=\&https:\u002F\\u002Fnews\u002Fleaders\u002F-does-it-play-fair-how-china-battling-ever-more-intensely-world-markets\&\u003EDoes China Play Fair? 【戳】\u003C\u002Fa\u003E\u003C\u002Fp\u003E\u003Cp\u003ECompetition from China will only get fiercer. That calls for cool heads and wise policies\u003C\u002Fp\u003E\u003Cp\u003E全球经济\u003C\u002Fp\u003E\u003Cp\u003E\u003Ca href=\&https:\u002F\u002Fgbr.businessreview.global\u002Farticles\u002Fview\u002F59c52ac\u002Fzh_CN\u002Fen_GB\&\u003E中国在公平竞争么？【CLICK】\u003C\u002Fa\u003E\u003C\u002Fp\u003E\u003Cp\u003E来自中国的竞争只会更加激烈。对此需要冷静的头脑和明智的政策\u003C\u002Fp\u003E\u003Cp\u003E\u003Cbr\u003E\u003C\u002Fp\u003E\u003Cp\u003EIF DONALD TRUMP had \u003Cb\u003Eslapped punitive tariffs on\u003C\u002Fb\u003E all Chinese exports to America, as he promised, he would have started a trade war. Fortunately, the president hesitated, partly because he wants China’s help in \u003Cb\u003Ethwarting\u003C\u002Fb\u003E North Korea’s nuclear ambitions. But that is not the end of the story. Tensions over China’s industrial might now threaten the architecture of the global economy. America’s trade representative this week called China an “unprecedented” threat that cannot be \u003Cb\u003Etamed\u003C\u002Fb\u003E by existing trade rules. The European Union, worried by \u003Cb\u003Ea spate of\u003C\u002Fb\u003E Chinese acquisitions, is drafting stricter rules on foreign investment. And, all the while, China’s strategy for modernising its economy is \u003Cb\u003Eadding further strain\u003C\u002Fb\u003E. 如果特朗普像他承诺过的那样，对中国出口到美国的所有商品\u003Cb\u003E征收惩罚性关税\u003C\u002Fb\u003E，那么他就会挑起一场贸易战。所幸这位总统犹豫了，部分原因是他需要中国的帮助来\u003Cb\u003E挫败\u003C\u002Fb\u003E朝鲜的核野心。但故事并未到此结束。中国的产业实力引发的紧张局势如今威胁到了全球经济的结构。本周，美国贸易代表称中国是“前所未有的”威胁，无法由现有的贸易规则\u003Cb\u003E制约\u003C\u002Fb\u003E。中国\u003Cb\u003E一连串\u003C\u002Fb\u003E的海外收购引发了欧盟的担忧，它正在起草针对外国投资的更严格规定。与此同时，中国经济现代化的战略也在\u003Cb\u003E加剧这种紧张局势\u003C\u002Fb\u003E。\u003C\u002Fp\u003E\u003Cp\u003EAt the heart of these tensions is one simple, overwhelming fact: firms around the world face ever more intense competition from their Chinese rivals. China is not the first country to industrialise, but none has ever made the \u003Cb\u003Eleap\u003C\u002Fb\u003E so rapidly and on such a \u003Cb\u003Emonumental\u003C\u002Fb\u003E scale. Little more than a decade ago Chinese boom towns \u003Cb\u003Echurned out\u003C\u002Fb\u003E zips, socks and cigarette lighters. Today the country is at the global frontier of new technology in everything from mobile payments to \u003Cb\u003Edriverless cars\u003C\u002Fb\u003E. 这些紧张局势的核心是一个简单而又无法辩驳的事实：世界各地的公司都面临着来自中国对手的日益激烈的竞争。中国不是第一个工业化的国家，但从没有哪一个国家曾如此迅速地实现了\u003Cb\u003E如此大\u003C\u002Fb\u003E规模的\u003Cb\u003E飞跃\u003C\u002Fb\u003E。仅仅十多年前，中国新兴的城镇还在\u003Cb\u003E大量生产\u003C\u002Fb\u003E拉链、袜子和打火机。如今，从移动支付到\u003Cb\u003E无人驾驶汽车\u003C\u002Fb\u003E，中国在方方面面都处于全球新技术的前沿。\u003C\u002Fp\u003E\u003Cp\u003EEven as China’s achievements \u003Cb\u003Einspire awe\u003C\u002Fb\u003E, there is growing concern that the world will be dominated by an economy that does not play fair. Businesses feel threatened. Governments that have seen Brexit and the election of Mr Trump, worry about the effects of job losses and shrinking technological leadership. Yet if the outcome is to be good, they must all think clearly about the real nature of China’s challenge. 中国的成就\u003Cb\u003E令人敬畏\u003C\u002Fb\u003E，但同时人们也越来越担心世界将被一个不公平行事的经济体所主宰。各行各业感受到了威胁。已目睹了英国退欧和特朗普当选的各国政府则担心职位流失和技术领先优势缩减的影响。不过，如果想要得到好的结果，它们都必须清楚地思考由中国带来的挑战的真正本质。\u003C\u002Fp\u003E\u003Cp\u003E\u003Cb\u003EGo, in three dimensions\u003C\u002Fb\u003E\u003C\u002Fp\u003E\u003Cp\u003E\u003Cb\u003E从三方面入手\u003C\u002Fb\u003E\u003C\u002Fp\u003E\u003Cp\u003EUndoubtedly, China has form. It \u003Cb\u003Ekept its currency cheap\u003C\u002Fb\u003E for years, it finances its state-owned giants with \u003Cb\u003E\u003C\u002Fb\u003E and its cyber-spies steal secrets. Yet \u003Cb\u003Edepictions\u003C\u002Fb\u003E of corporate China as just an undemocratic, state-run monster, thieving and cheating to get ahead, are crude and out of date. Home-grown innovation is flourishing. The innovators are mainly private, not the many heads of a single creature called China Inc. To separate hype from reality, think of Chinese competition as having three dimensions: illegal, intense and unfair. \u003Cb\u003EEach needs a different response.\u003C\u002Fb\u003E 毫无疑问，中国在不当竞争方面是有前科的。多年来它一直\u003Cb\u003E压低汇率\u003C\u002Fb\u003E以提振出口；它为国有巨头提供\u003Cb\u003E低息贷款\u003C\u002Fb\u003E；它的网络间谍窃取机密。但如果还是把中国企业\u003Cb\u003E描述\u003C\u002Fb\u003E成不民主的国营怪物，靠剽窃和欺骗冒头，就简单粗暴且过时了。中国本土的创新正在蓬勃发展。创新的主体是私营企业，而不是一个叫“中国有限公司”的生物的诸多分身。为区分炒作和现实，可以从三个方面看待中国的竞争：非法、激烈、不公平。\u003Cb\u003E每一面需要不同的应对方式。\u003C\u002Fb\u003E\u003C\u002Fp\u003E\u003Cp\u003EFirst, consider illegality. The best example is the \u003Cb\u003Eblatant theft\u003C\u002Fb\u003E of intellectual property that makes for the most sensational headlines, such as the charges laid in 2014 against five Chinese military officers for hacking into American nuclear, solar and metals firms. The good news is such crimes are declining. An agreement with America in 2015 seemingly led to a \u003Cb\u003Emarked drop\u003C\u002Fb\u003E in Chinese hacks of foreign companies and, as Chinese firms produce more of value, they are themselves demanding better intellectual-property protection at home. 首先来看非法。最好的例证就是中国\u003Cb\u003E明目张胆地窃取\u003C\u002Fb\u003E知识产权，一次次成为轰动的头条新闻。比如2014年，五名中国军官被指控入侵美国核电厂、太阳能企业和金属公司的电脑网络。好消息是此类犯罪正在减少。2015年与美国达成的一项协议似乎令中国对外国公司的黑客行为\u003Cb\u003E大幅减少\u003C\u002Fb\u003E，而且随着中国企业创造的价值越来越高，它们自己也要求在国内获得更好的知识产权保护。\u003Cbr\u003E\u003C\u002Fp\u003E\u003Cp\u003EThe second dimension—intense but legal competition—is far more important. Chinese firms have proven that they can \u003Cb\u003Emake good products for less\u003C\u002Fb\u003E. Consumer prices for televisions, adjusted for quality, fell by more than 90% in the 15 years after China joined the World Trade Organisation (WTO). China’s share of global exports has risen to 14%, the highest any country has reached since America in 1968. That may fall as China \u003Cb\u003Eloses its grip on\u003C\u002Fb\u003E low-value industries such as textiles. But it is gaining a new reputation in high tech. If data are the new oil, China’s tech industry has vast reserves in the information generated by the hundreds of millions of its people online—unprotected by privacy rules. Whether you make cars in Germany, semiconductors in America or robots in Japan, the chances are that in future some of your \u003Cb\u003Efiercest rivals\u003C\u002Fb\u003E will be Chinese. 第二个方面——激烈但合法的竞争——要重要得多。中国的公司已经证明，它们能\u003Cb\u003E以更少的成本做出好产品\u003C\u002Fb\u003E。中国加入世界贸易组织（WTO）的15年里，按质量调整后，电视机的零售价格下降了90%以上。中国在全球出口中所占的份额已经上升到14%，这是自1968年的美国以来的最高水平。随着中国\u003Cb\u003E失去对\u003C\u002Fb\u003E纺织等低附加值产业\u003Cb\u003E的控制\u003C\u002Fb\u003E，这一份额可能会下降。但它正在高科技领域赢得新的声誉。如果说数据是新的石油，中国的科技行业就有巨大的“储量”，这些信息由数亿网民产生，且不受隐私条款的保护。无论你是在德国生产汽车，在美国生产半导体，还是在日本生产机器人，未来某些\u003Cb\u003E最强劲的竞争对手\u003C\u002Fb\u003E很可能会是中国人。\u003C\u002Fp\u003E\u003Cp\u003ELast, and hardest to deal with, is unfair competition: sharp practice that breaks no global rules. The government demands that firms give away technology as the cost of admission to China’s vast market (see article). Foreign firms have been targeted in the biggest of China’s anti-monopoly cases. The government restricts access to lucrative sectors, while financing assaults on those same industries abroad. Such behaviour is dangerous precisely because today’s rules offer no redress.\u003C\u002Fp\u003E\u003Cp\u003E\u003Cb\u003EDon’t get angry. Get even\u003C\u002Fb\u003E\u003C\u002Fp\u003E\u003Cp\u003ESorting Chinese competition into these categories helps calibrate the response. Blatant illegality is the most straightforward. Governments must prosecute and seek redress, whether through the courts or the WTO. Firms can better protect themselves against cyber-thieves—from China and elsewhere.\u003C\u002Fp\u003E\u003Cp\u003EThough it is politically hard, the best response to intense competition is to welcome it. Consumers will gain from lower costs and faster innovation. Misguided attempts to hold back the tide would not only lose those potential gains but might also blow up the world trading system, with catastrophic results. Rather than try to stop the loss of jobs, governments should provide retraining and a decent safety net. Both companies and governments need to spend more on education and research. Six years ago Barack Obama said America faced a new “Sputnik moment” in China’s rise. Since then not much extra has been devoted to research, training and infrastructure.\u003C\u002Fp\u003E\u003Cp\u003EThe hardest category is competition that is unfair, but not illegal. One approach is to coax China into behaving better by acting collectively. America, Europe and big Asian countries could jointly publish information about economic harm from China’s policies—as they did by sharing details about overcapacity in the steel industry, nudging China into cutting its excesses. They should demand reciprocity, requiring China to give foreign companies the same access that its own firms enjoy in their markets. Governments need to review their policies for screening investments from China so that they can block genuine threats to national security (though only those). And they should also require that investors with state backing report this in full, and punish those hiding their true identity.\u003C\u002Fp\u003E\u003Cp\u003EMuch of the responsibility for putting this right falls on China. It may ask why it should hold itself back. After all, 19th-century Germany and America grew rich behind subsid Britain and Japan were bullies. Yet, having done so well out of the global commercial ecosystem, China should recognise that it has become one of its custodians. Abuse it—illegally or by overburdening it—and it will break.要纠正这些问题，主要责任还是在中国身上。它可能会问，为什么它应该保持克制。毕竟，19世纪的德国和美国就是靠补贴和关税壁垒富裕起来，英国和日本从前也恃强凌弱。不过，既然中国从全球商业生态系统中获益良多，它应该认识到自己已经成为这个系统的守护者之一。滥用这个系统——无论是违法或是让它超负荷——都会让它崩溃。\u003C\u002Fp\u003E&,&updated&:new Date(&T00:40:32.000Z&),&canComment&:false,&commentPermission&:&review&,&commentCount&:0,&likeCount&:29,&state&:&published&,&isLiked&:false,&slug&:&&,&isTitleImageFullScreen&:false,&rating&:&none&,&sourceUrl&:&&,&publishedTime&:&T08:40:32+08:00&,&links&:{&comments&:&\u002Fapi\u002Fposts\u002F2Fcomments&},&url&:&\u002Fp\u002F&,&titleImage&:&https:\u002F\\u002Fv2-a7bfd97bc98ef7bed567218b_r.jpg&,&summary&:&&,&href&:&\u002Fapi\u002Fposts\u002F&,&meta&:{&previous&:null,&next&:null},&snapshotUrl&:&&,&commentsCount&:0,&likesCount&:29},&&:{&title&:&经济学人官译：中欧铁路&,&author&:&iqlake&,&content&:&\u003Cp\u003ELogistics\u003C\u002Fp\u003E\u003Cp\u003EFreight gain\u003C\u002Fp\u003E\u003Cp\u003ENew rail routes between China and Europe will change trade patterns\u003C\u002Fp\u003E\u003Cp\u003E物流\u003C\u002Fp\u003E\u003Cp\u003E货运发展\u003C\u002Fp\u003E\u003Cp\u003E中欧之间的新铁路线将改变贸易模式\u003C\u002Fp\u003E\u003Cp\u003E\u003Cbr\u003E\u003C\u002Fp\u003E\u003Cp\u003EASTANA in Kazakhstan is one of the world’s most remote capitals, surrounded by thousands of kilometres of empty steppe. This summer Astana attempted to launch itself onto the global stage by hosting the World Expo, which closed on September 10th and underwhelmed many attendees. But there are other ways to have an impact. On the city’s north side, away from the Expo’s exhibits, a series of diesel trains, each pulling dozens of containers, roll through the old railway station. Most are heading from China to Europe. Last year over 500,000 tonnes of freight went by train between the two, up from next to nothing before 2013. Airlines and shipping firms are watching things closely. 哈萨克斯坦的阿斯塔纳（Astana）是世界上最偏僻的首都之一，周围是绵延数千公里的空旷大草原。今年夏天，阿斯塔纳主办了世博会，试图借此登上国际舞台。世博会于9月10日闭幕，并未给许多参会者留下深刻印象。不过，要想产生影响还有别的办法。在远离世博会会址的阿斯塔纳北部，一列列装运着几十个集装箱的柴油机车驶过老火车站，大多是从中国开往欧洲的。去年，中欧之间有超过50万吨货物经铁路往来，而2013年之前还寥寥无几。航空公司和船运公司正在密切关注这一发展趋势。\u003C\u002Fp\u003E\u003Cp\u003E\u003Cbr\u003E\u003C\u002Fp\u003E\u003Cp\u003EThe trains rumbling through Astana result from a Chinese initiative, in tandem with countries like Kazakhstan, to build a “New Silk Road” through Central Asia. The earlier overland routes were once the conduits for most trade between Europe and China and I they faded into irrelevance when European ships started circumnavigating the Cape of Good Hope.中国提出与哈萨克斯坦等国家一起打造一条贯穿中亚的“新丝绸之路”，这些隆隆驶过阿斯塔纳的货运列车正是这一倡议的产物。古丝绸之路曾是欧洲与中国及印度之间大多数贸易往来的通道。当欧洲船只开始绕行好望角之后，这一陆上贸易通道就渐渐衰落了。\u003C\u002Fp\u003E\u003Cp\u003E\u003Cbr\u003E\u003C\u002Fp\u003E\u003Cp\u003EChina has long wanted to develop its inland regions and push industry to “go west”, in order to spread economic growth more evenly. Manufacturers have been loth to shift, in part because of the higher cost of moving goods to ports for export. Developing a rail-freight network to Europe—an important part of China’s “One Belt One Road” policy—opens up a new route to market for its poorest areas. The land route through Central Asia is relatively short. A container ship too large for the Suez canal must make a 24,000km journey to reach Europe. Trains travel no more than 11,000km to reach the same destination.长期以来中国一直希望发展内陆地区，推动产业“走向西部”，促进更均衡的经济发展。制造商一直不愿转移，部分原因是这样会拉高将货物运往出口港口的成本。发展一个通往欧洲的铁路货运网络是中国“一带一路”政策的重要组成部分，为中国最贫困地区开辟了一条通往市场的新路线。通过中亚到达欧洲的陆路通道相对较短。无法通过苏伊士运河的大型集装箱船必须航行2.4万公里才能到达欧洲，走铁路的话全程不超过1.1万公里。\u003C\u002Fp\u003E\u003Cp\u003E\u003Cbr\u003E\u003C\u002Fp\u003E\u003Cp\u003EKazakhstan has spent over 1.1trn tenge ($3.2bn) on upgrading its railway lines and rolling stock since 2011. That includes $250m on the Khorgos Gateway, a dry port at the border with China that lifts containers from Chinese trains onto Kazakh ones to overcome a change in track width (a problem that has stymied previous efforts to build railway routes between Europe and China).自2011年以来，哈萨克斯坦已经为升级铁路线路和铁路机车投入超过1.1万亿坚戈（32亿美元），其中包括投向中哈边界的无水港霍尔果斯口岸的2.5亿美元。由于中哈铁路轨距不同（这一差异阻碍了过去在中欧之间修建铁路线的努力），中国列车上的集装箱要在此换装到哈萨克斯坦的列车上。\u003C\u002Fp\u003E\u003Cp\u003E\u003Cbr\u003E\u003C\u002Fp\u003E\u003Cp\u003EVolumes of freight travelling between China and Europe by rail are rising quickly. Between 2013 and 2016 cargo traffic quintupled in weight. In the first half of this year the value of goods travelling by train rose by 144% compared with the same period in 2016. Western firms have been keen to embrace rail freight because it helps them to lower costs, says Ronald Kleijwegt, an expert on the industry. In the case of high-tech electronics, for example, which consumers like to receive quickly, making them on China’s coast and air-freighting them to Europe is extremely pricey. 中欧之间的铁路货运量正在快速上升。2013年至2016年间，货运重量增长了四倍。今年上半年，经铁路运输的商品价值相比2016年同期增长了144％。铁路货运行业专家罗纳德·克莱伊韦格特（Ronald Kleijwegt）说，西方企业十分欢迎中欧铁路货运的发展，因为这有助它们降低成本。以高科技电子行业为例，高科技电子产品的消费者希望能快速收到货物，而在中国沿海地区制造这些产品再空运至欧洲的成本极高。\u003C\u002Fp\u003E\u003Cp\u003E\u003Cbr\u003E\u003C\u002Fp\u003E\u003Cp\u003EHow worried should shipping firms and airlines be? Kazakhstan’s national rail company, KTZ, says it will have capacity for 1.7m containers to pass through the country between Europe and China each year by 2020; that is a tenth of the volume currently carried by sea and air between the two. In the longer term, a full modernisation of the existing main three rail routes from China to Europe could produce 3m containers a year in capacity. 航运和航空公司对此该多担心呢？哈萨克斯坦国家铁路公司（KTZ）表示，到2020年，它在经由哈萨克斯坦往返中欧的线路上将拥有每年170万个集装箱的运力，是目前两地间海运和空运年运输量的十分之一。而在更长久的时间里，中欧之间现有三条主要铁路线的全面现代化将带来每年300万个集装箱的运力。\u003C\u002Fp\u003E\u003Cp\u003E\u003Cbr\u003E\u003C\u002Fp\u003E\u003Cp\u003EBut there are reasons to doubt that will happen. For one thing, China plans to stop handing out government subsidies for additional rail-freight capacity from 2020, which will slow the network’s expansion. Sea freight has little to fear in the near term, says Soren Skou, chief executive of Maersk, the world’s biggest container-shipping line. Trains may take away some future growth from ships, he concedes, but not their existing business.但有理由对此前景表示怀疑。一方面，中国计划从2020年起停止向新增铁路运力提供政府补贴，这将让铁路运输网络扩张的速度放慢。全球最大的集装箱航运公司马士基（Maersk）的首席执行官施索仁（Soren Skou）说，短期内海运业没什么可担心的。他承认，火车可能会抢走船舶的一部分未来增量业务，但其现有业务不会被抢夺。\u003C\u002Fp\u003E\u003Cp\u003E\u003Cbr\u003E\u003C\u002Fp\u003E\u003Cp\u003EAir cargo is more vulnerable. Last year, 180,000 tonnes of cargo travelled on trains to western Europe from China (the remainder was destined for Russia and eastern Europe). That is a small fraction of the 52m tonnes that came by sea, but a big chunk of the 700,000 tonnes that came by air. Much of that air cargo could switch to rail in future, says Mr Kleijwegt, with one important proviso—that Russia would need to lift the retaliatory sanctions it placed in 2014 on imports of Western food, which stop most foodstuffs from travelling by land between Europe and China. That is unlikely for the time being. But it was only a decade ago that people thought the idea of freight trains between Europe and China was a joke, says Mr Kleijwegt—and no one laughs at that any more.空运更易受影响。去年，18万吨货物从中国经铁路运往西欧（其余运往俄罗斯和东欧），这在经海路运抵的5200万吨中只是个零头，但在经空运抵达的70万吨中却是很大一块。克莱伊韦格特说，未来大部分空运货物都可能转而走铁路，不过一个重要的前提条件是，俄罗斯需解除从2014年起对西方进口食品施行的报复性制裁措施，这一措施阻碍了中欧之间大多数食品的陆路运输。目前看来制裁不大可能取消。不过克莱伊韦格特说，仅仅十年前，中欧之间开通货运列车还被看作一个笑话，而如今再也没人取笑了。 \u003C\u002Fp\u003E&,&updated&:new Date(&T00:39:53.000Z&),&canComment&:false,&commentPermission&:&review&,&commentCount&:0,&likeCount&:14,&state&:&published&,&isLiked&:false,&slug&:&&,&isTitleImageFullScreen&:false,&rating&:&none&,&sourceUrl&:&&,&publishedTime&:&T08:39:53+08:00&,&links&:{&comments&:&\u002Fapi\u002Fposts\u002F2Fcomments&},&url&:&\u002Fp\u002F&,&titleImage&:&https:\u002F\\u002Fv2-dd5e58e8de_r.jpg&,&summary&:&&,&href&:&\u002Fapi\u002Fposts\u002F&,&meta&:{&previous&:null,&next&:null},&snapshotUrl&:&&,&commentsCount&:0,&likesCount&:14},&&:{&title&:&经济学人官译：免费不是免费&,&author&:&iqlake&,&content&:&\u003Ch2\u003EFree exchange\u003C\u002Fh2\u003E\u003Ch2\u003EPriceless\u003C\u002Fh2\u003E\u003Ch2\u003EFree stuff on the internet comes at a cost\u003C\u002Fh2\u003E\u003Ch2\u003E自由交流\u003C\u002Fh2\u003E\u003Ch2\u003E道是无价却有价\u003C\u002Fh2\u003E\u003Ch2\u003E互联网免费午餐有代价\u003C\u002Fh2\u003E\u003Cp\u003E\u003Cbr\u003E\u003C\u002Fp\u003E\u003Cp\u003EFACEBOOK, whose users visit for an average of 50 minutes a day, promises members: “It’s free and always will be.” It certainly sounds like a steal. But it is only one of the bargains that apparently litter the internet: YouTube watchers devour 1bn hours of videos every day, for instance. These free lunch the problem is calculating how much it is. Because consumers do not pay for many digital services in cash, beyond the cost of an internet connection, economists cannot treat these exchanges like normal transactions. The economics of free are different.\u003C\u002Fp\u003E\u003Cp\u003EFacebook的用户平均每天花50分钟使用该网站。它“免费，且永远免费”的承诺对用户来说无疑像捡了个大便宜。表面上看，互联网上净是这种“大便宜”，Facebook只是其中之一——YouTube的观众每天观看视频的时间总计多达10亿小时。这些免费午餐确实是有代价的，问题在于如何计算这些代价。除了连接互联网的费用，消费者没有为很多数字服务付费，经济学家因而不能将这类交换视为常规交易。免费服务的经营模式有所不同。\u003C\u002Fp\u003E\u003Cp\u003EUnlike conventional merchants, companies like Facebook and Google have their users themselves produce value. Information and pictures uploaded to social networks draw others to the site. Online searches, selections and “likes” teach algorithms what people want. (Now you’ve bought “The Communist Manifesto”, how about a copy of “Das Kapital”?) \u003C\u002Fp\u003E\u003Cp\u003E有别于传统商家，Facebook 和谷歌这样的公司让用户自己来为它们创造价值。上传到社交网络的信息和图片会将其他人吸引过来。在线搜索、选择和“点赞”让算法了解人们的需求。比如，既然你买了本《共产党宣言》，那是否考虑再买一本《资本论》呢？\u003C\u002Fp\u003E\u003Cp\u003EThe prevalence of free services is partly a result of history. In the early years of the internet, consumers became used to getting stuff for nothing. They have little idea of how much
since digital companies have access to billions of people, the value of one person’s data is tiny anyway. More fundamentally, scarcity is not a constraint in the digital world as it is in the physical one. Data are both inexhaustible and super-cheap to transport. In 1993 MCI Mail was charging people 50 cents for the first 500 characters of a digital message, increasing by ten cents for each extra 500. The internet slashed that price to zero. Charging would have been impractical, so small is the marginal cost. \u003C\u002Fp\u003E\u003Cp\u003E免费服务盛行在一定程度上是历史的产物。在互联网时代早期，消费者就已习惯了免费午餐。他们不清楚自己的数据值多少钱。而数字公司接触到的人数以亿计，因此单个人的数据不管怎么说价值都很小。更重要的是，不同于实体世界，数字世界里没有“稀缺”一说。数据取之不竭，传输又超级便宜。1993年，美国通讯公司MCI的电邮服务对前500个字符收费50美分，之后每增加500个字符续收10美分。而互联网则将这笔费用削减至零。边际成本微乎其微，还收费的话就不切实际了。\u003C\u002Fp\u003E\u003Cp\u003EUsers may pay nothing, but companies like Google and Facebook have fixed costs to cover: engineers, data centres, etc. To make money, they squeeze their users indirectly, by charging companies to put appropriate advertisements in front of captive eyeballs. In the second quarter of 2017, Facebook eked an average of $4.65 out of each of its users by peppering screens with ads and promoted posts. (By comparison, just eight cents came from payments and other fees, mainly from people paying for stuff within virtual games.)\u003C\u002Fp\u003E\u003Cp\u003E用户也许是不用支付任何费用，但谷歌和Facebook这样的公司需要承担花在工程师、数据中心等上面的固定成本。为了赢利，它们间接榨取用户，将精准定位的广告推送到用户眼前，让他们不由自主地浏览，为此向公司收取广告费。通过充斥屏幕的广告和推广帖，2017年第二季度Facebook平均从每个用户那里赚取了4.65美元。相比而言，只有8美分来自用户付费和其他费用，主要是购买虚拟游戏中的装备。\u003C\u002Fp\u003E\u003Cp\u003EIn the absence of prices, economists struggle to work out what people are getting back when they barter their data and attention for digital services. Some evidence suggests that they are doing rather well. A recent study by Erik Brynjolfsson, Felix Eggers and Avinash Gannameneni of the Massachusetts Institute of Technology offered people different cash amounts in exchange for giving up Facebook for a month. Based on the responses, they then estimated its average annual value to the consumer at around $750. A simpler survey in the same study (without real cash offers) suggested that on average people value free search engines at $16,600 per year, maps at $2,800 and video at $900.\u003C\u002Fp\u003E\u003Cp\u003E由于不存在价格，经济学家很难计算出人们在用数据和注意力换取数字服务时实际得到了多少。一些证据显示人们得到免费服务挺值的。在最近的一项研究中，麻省理工学院的埃里克·布莱恩约弗森（Erik Brynjolfsson）、菲利克斯·艾格斯（Felix Eggers）以及阿维纳什·甘纳门耐尼（Avinash Gannameneni）向受试者提供数目不等的现金，作为交换，受试者放弃使用Facebook一个月。他们根据受试者的反应估算出，对消费者来说Facebook的平均年价值在750美元左右。同一项研究中一个更简单的调查（没有真正提供现金）显示，平均下来，人们对免费搜索引擎的估价为每年16,600美元，免费地图2800美元，免费视频900美元。\u003C\u002Fp\u003E\u003Cp\u003EThis sounds like a wonderful deal for the consumer, but it generates problems elsewhere. Take taxes. Professionals are not allowed to evade tax by selling their services for benefits in kind, so why should consumers not be taxed if they are paid for their data in the form of services? Statisticians also struggle in a post-price world. GDP is mostly measured by transactions at market prices. A recent study by Leonard Nakamura of the Federal Reserve Bank of Philadelphia and Jon Samuels and Rachel Soloveichik of the Bureau of Economic Analysis used the amount spent on advertising to estimate uncounted output, and calculated that in 2013 American GDP should have been $19bn higher than reported.\u003C\u002Fp\u003E\u003Cp\u003E听起来，这对于消费者是个很不错的交易，但它却在其他方面引发了问题，比如税收。既然不允许专业人士以出售服务换取实物福利的方式逃税，那如果消费者提供了数据而以服务的形式获得报酬，他们怎么就可以不交税？统计人员同样因为无价格领域而为难。GDP大多由以市场价成交的交易来衡量。在最近一项研究中，费城联邦储备银行的莱纳德·纳卡穆拉（Leonard Nakamura）以及美国商务部经济分析局（BEA）的乔恩·塞缪尔斯（Jon Samuels）、雷切尔·索洛维切克（Rachel Soloveichik）通过广告费来估算未计入的产值，估测出2013年美国GDP应该比实际报告的多190亿美元。\u003C\u002Fp\u003E\u003Cp\u003EPrivacy activists also worry. Consumers tend to respond much more strongly to “free” offers than to prices that are only fractionally higher than zero. When Amazon first offered free shipping in European countries, orders surged—but not in France, where by mistake it charged around ten cents. The activists’ concern is that the “free” label fosters poor decisions, making people, for example, reveal more about themselves than they would in a more formal exchange. Researchers talk of the “privacy paradox”: when asked, people say that they care much more about their privacy than their actions would suggest.\u003C\u002Fp\u003E\u003Cp\u003E隐私权倡议人士也感到担忧。比起那些几近为零的价格，消费者对免费品还是情有独钟得多。亚马逊在欧洲首次推出免运费服务时，订单激增，但法国除外——由于操作失误，亚马逊在法国收取大约10美分的运费。隐私权倡议人士担心，“免费”的旗号会助长轻率的决定，例如，比起较为正规的交易，人们在获取免费服务时会泄露更多的隐私。研究人员谈到了“隐私权悖论”的现象：当被问及隐私权时，人们表示他们很关心自己的隐私，而他们的行为却远非如此。\u003C\u002Fp\u003E\u003Cp\u003EThe free economy also troubles competition authorities. Excessive market power can be defined as the ability to raise prices above what would be charged in a competitive market. With no prices to compare, and other options only a click away, companies such as Google seem to operate in an environment of cut-throat competition. It is naive to think so. Consumers are more captive than the low cost of switching might imply. Google, for example, commands a market share for internet search of over 90% in most countries in the European Union, where antitrust authorities in June fined it EUR2.4bn ($2.7bn) for promoting its own comparison shopping services above its competitors’. Its services may have been free, but the trustbusters judged that its market power was curbing consumers’ choices. In the absence of prices, lack of competition will show up in other ways: demanding more information from users than they want to give, or irritating them by stuffing their service chock-full of adverts.\u003C\u002Fp\u003E\u003Cp\u003E免费经济同样困扰着竞争监管机构。过高的市场支配力可以定义为，企业能够将价格提升到高于竞争性市场中要价的水平。没有明码标价可以比较，且用户只需点一下鼠标就可轻易转向其他选择——这样看来，谷歌这样的公司好像是在一个竞争惨烈的环境中运营。如果这样想就未免天真了。虽然转换服务商的成本极低，消费者并没有看上去那么自由。例如，谷歌在欧盟大多数国家控制着超过90%的互联网搜索的市场份额，今年6月，因为在搜索结果中将自己的比价购物服务置于其他竞争对手之上，欧盟反垄断机构向其罚款24亿欧元（27亿美元）。它的服务或许是免费的，但是反垄断人士判定，它的市场支配力限制了消费者的选择。在没有标价的情况下，竞争不足会以其他方式呈现，比如要求消费者提供超出其意愿的信息，或者在服务中填满广告而让消费者不堪其扰。\u003C\u002Fp\u003E\u003Ch2\u003ENo such thing as a free exchange\u003C\u002Fh2\u003E\u003Cp\u003EOpinion is divided on whether the free economy needs fixing, and if so, how. In his book “Who Owns the Future?”, Jaron Lanier suggests that tiny payments for digital contributions might correct yet another problem, a misallocation of labour. If companies paid people for useful data, rather than mopping up what they leave behind as they use online services, then prices could nudge people towards more productive online activity. Others advocate tougher regulation, mandating that consumers have the option of paying for a version of their social-media platforms free of advertisements and digital profiles. Neither seems imminent, and each comes with its own problems. But both would at least force people to start counting the cost of that priceless lunch.\u003C\u002Fp\u003E\u003Cp\u003E天下没有免费的交易\u003C\u002Fp\u003E\u003Cp\u003E免费经济是否需要修正？如果需要，该怎么修正？对此人们意见不一。杰伦·拉尼尔（Jaron Lanier）在《谁拥有未来》（Who Owns the Future?）一书中表示，如果用户在网络上的贡献能得到哪怕很少的报酬，可能还会纠正另一个问题——劳动力分配不当。假设公司为有用的数据向人们付费，而不是将他们使用网络服务时留下的信息“一网打尽”，那么价格可能会将人们推向更富效益的网络活动。另一些人倡导更严厉的监管——规定消费者可以付费来选用一个没有广告也无需提供个人资料的社交媒体平台。以上两种措施似乎都不会很快到来，而且也都有各自的问题。但是至少两者都会迫使人们开始计算——这种免费午餐究竟代价几何？\u003C\u002Fp\u003E&,&updated&:new Date(&T01:09:42.000Z&),&canComment&:false,&commentPermission&:&review&,&commentCount&:0,&likeCount&:16,&state&:&published&,&isLiked&:false,&slug&:&&,&isTitleImageFullScreen&:false,&rating&:&none&,&sourceUrl&:&&,&publishedTime&:&T09:09:42+08:00&,&links&:{&comments&:&\u002Fapi\u002Fposts\u002F2Fcomments&},&url&:&\u002Fp\u002F&,&titleImage&:&https:\u002F\\u002Fv2-06cacccd5a39a99_r.jpg&,&summary&:&&,&href&:&\u002Fapi\u002Fposts\u002F&,&meta&:{&previous&:null,&next&:null},&snapshotUrl&:&&,&commentsCount&:0,&likesCount&:16},&&:{&title&:&经济学人官译：人脸识别技术&,&author&:&iqlake&,&content&:&\u003Ch2\u003EFacial technology (1)\u003C\u002Fh2\u003E\u003Ch2\u003EKeeping a straight face\u003C\u002Fh2\u003E\u003Ch2\u003EIn the first of two stories about faces and technology, artificial intelligence is used to spot signs of sexuality\u003C\u002Fh2\u003E\u003Ch2\u003E人脸识别技术（1）\u003C\u002Fh2\u003E\u003Ch2\u003E不露声色\u003C\u002Fh2\u003E\u003Ch2\u003E人脸与科技专题之一——用人工智能识别性取向\u003C\u002Fh2\u003E\u003Cp\u003E\u003Cbr\u003E\u003C\u002Fp\u003E\u003Cp\u003EMODERN artificial intelligence is much feted. But its talents boil down to a superhuman ability to spot patterns in large volumes of data. Facebook has used this ability to produce maps of poor regions in unprecedented detail, with an AI system that has learned what human settlements look like from satellite pictures. Medical researchers have trained AI in smartphones to dete a Google system can make precise guesses about the year a photograph was taken, simply because it has seen more photos than a human could ever inspect, and has spotted patterns that no human could.\u003C\u002Fp\u003E\u003Cp\u003E现代人工智能（以下简称AI）深受推崇。归根到底，就是因为它那从大量数据中发掘模式的超人能力。Facebook已利用这种能力绘制出空前详尽的贫困地区地图，它使用的AI系统学会了如何在卫星图片中辨认人类居住区域；医疗研究人员训练智能手机中的AI程序来检测癌症病变；谷歌的一套系统可以准确猜出照片拍摄的年份，只因为该系统看过的照片数量之多非人力所能及，因而能发现人类无法察觉的模式。\u003Cbr\u003E\u003C\u002Fp\u003E\u003Cp\u003EAI’s power to pick out patterns is now turning to more intimate matters. Research at Stanford University by Michal Kosinski and Yilun Wang has shown that machine vision can infer sexual orientation by analysing people’s faces. The researchers suggest the software does this by picking up on subtle differences in facial structure. With the right data sets, Dr Kosinski says, similar AI systems might be trained to spot other intimate traits, such as IQ or political views. Just because humans are unable to see the signs in faces does not mean that machines cannot do so.\u003C\u002Fp\u003E\u003Cp\u003EAI发掘模式的能力正被用于更为私密的事情。斯坦福大学的米哈尔·科辛斯基（Michal Kosinski）和王轶伦的研究显示，机器视觉可通过分析人脸来推断性取向。他们认为，软件是通过检测面部结构上的细微差异而做出推断的。科辛斯基表示，只要使用恰当的数据集，也许还可以训练类似的AI系统来识别其他私密特征，如智商或政见。人类无法发现人脸发出的这些信号，并不意味着机器也识别不了。\u003Cbr\u003E\u003C\u002Fp\u003E\u003Cp\u003EThe researchers’ program, details of which are soon to be published in the \u003Ci\u003EJournal of Personality and Social Psychology,\u003C\u002Fi\u003E relied on 130,741 images of 36,630 men and 170,360 images of 38,593 women downloaded from a popular American dating website, which makes its profiles public. Basic facial-detection technology was used to select all images which showed a single face of sufficient size and clarity to subject to analysis. This left 35,326 pictures of 14,776 people, with gay and straight, male and female, all represented evenly.\u003C\u002Fp\u003E\u003Cp\u003E这项研究的详情即将发表于《个性与社会心理学期刊》（Journal of Personality and Social Psychology）。研究使用的数据下载自美国一家公开会员资料的热门约会网站，包含36,630名男性的13,0741张照片，以及38,593名女性的170,360张照片。研究人员首先利用基本的人脸检测技术，在所有照片中挑选出单张人脸的大小和清晰度可满足分析需要的照片，最终选出14,776人的35,326张照片，其中包括同性恋和异性恋、男性和女性，比例均等。\u003Cbr\u003E\u003C\u002Fp\u003E\u003Ch2\u003EOut of the numbers\u003C\u002Fh2\u003E\u003Cp\u003EThe images were then fed into a different piece of software called VGG-Face, which spits out a long string of numbers to r their “faceprint”. The next step was to use a simple predictive model, known as logistic regression, to find correlations between the features of those faceprints and their owners’ sexuality (as declared on the dating website). When the resulting model was run on data which it had not seen before, it far outperformed humans at distinguishing between gay and straight faces.\u003C\u002Fp\u003E\u003Ch2\u003E数字乾坤\u003C\u002Fh2\u003E\u003Cp\u003E接着，这些图片被输入到另一个名为VGG-Face的软件中，它给每个人都分配了一长串数字，作为其“脸纹”。下一步是采用名为“逻辑回归”的简单预测模型，探寻“脸纹”特征与其主人性取向（按约会网站上公布的情况）之间的关联。当所得出的模型分析它之前没见过的人脸数据时，区分同性恋和异性恋的能力远高于人类。\u003C\u002Fp\u003E\u003Cp\u003EWhen shown one photo each of a gay and straight man, both chosen at random, the model distinguished between them correctly 81% of the time. When shown five photos of each man, it attributed sexuality correctly 91% of the time. The model performed worse with women, telling gay and straight apart with 71% accuracy after looking at one photo, and 83% accuracy after five. In both cases the level of performance far outstrips human ability to make this distinction. Using the same images, people could tell gay from straight 61% of the time for men, and 54% of the time for women. This aligns with research which suggests humans can determine sexuality from faces at only just better than chance.\u003C\u002Fp\u003E\u003Cp\u003E面对随机显示的一张男同照片和一张直男照片，该模型的辨别准确率为81%。当显示同一人的五张照片时，模型的准确率达91%。在女性数据上，模型的表现较为逊色，凭一张照片辨别性取向的准确率为71%，五张照片时的准确率为83%。两种情况下，其表现均远超人类的区分能力。面对同样的照片，人们对男性性取向的识别率为61%，对女性性取向的识别率为54%。有研究显示人类从他人面部特征辨别性取向的能力仅略胜于瞎猜，上述数字印证了这一点。\u003C\u002Fp\u003E\u003Cp\u003EDr Kosinski and Mr Wang offer a possible explanation for their model’s performance. As fetuses develop in the womb, they are exposed to various levels of hormones, in particular testosterone. These are known to play a role in developing facial structures, and may similarly be involved in determining sexuality. The researchers suggest their system can pick up subtle signals of the latter from the former. Using other techniques, the program was found to pay most attention to the nose, eyes, eyebrows, cheeks, hairline and chin for determ the nose, mouth corners, hair and neckline were more important for women.\u003C\u002Fp\u003E\u003Cp\u003E科辛斯基和王轶伦为该模型的表现提供了一种可能的解释。胎儿在子宫中发育时，会暴露于各种不同水平的激素（尤其是睾丸酮）环境中。目前已知激素水平会影响到面部结构的发育，而它们对性取向可能也有类似的作用。研究人员表示，系统可以从面部结构获取有关性取向的细微信号。通过使用其他技术，研究人员发现该程序在识别男子性取向时最留意观察鼻子、眼睛、眉毛、面颊、发际线和下巴；对于女性，程序则更关注鼻子、嘴角、头发和领口。\u003C\u002Fp\u003E\u003Cp\u003EThe study has limitations. Firstly, images from a dating site are likely to be particularly revealing of sexual orientation. The 91% accuracy rate only applies when one of the two men whose images are shown is known to be gay. Outside the lab the accuracy rate would be much lower. To demonstrate this weakness, the researchers selected 1,000 men at random with at least five photographs, but in a ratio of gay to straight that more accurately ref approximately seven in every 100. When asked to select the 100 males most likely to be gay, only 47 of those chosen by the system actually were, meaning that the system ranked some straight men as more likely to be gay than men who actually are.\u003C\u002Fp\u003E\u003Cp\u003E这项研究有其局限性。首先，约会网站上的照片很可能会特别彰显个人的性取向。91%的准确率是有前提的——已知显示照片的两名男子中有一人是同性恋。在实验室之外，准确率会低得多。为展示这一弱点，研究人员随机选取1000名男子，每人至少有五张照片，但其中男同和直男的比例更接近于现实世界的情况——每100人中约有7人为同性恋。在被要求挑出100名最有可能是同性恋的男子时，系统选择的人中只有47人是男同。这表明，在系统看来，某些直男比真正的男同更像同性恋。\u003C\u002Fp\u003E\u003Cp\u003EHowever, when asked to pick out the ten faces it was most confident about, nine of the chosen were in fact gay. If the goal is to pick a small number of people who are very likely to be gay out of a large group, the system appears able to do so. The point is not that Dr Kosinski and Mr Wang have created software which can reliably determine gay from straight. That was not their goal. Rather, they have demonstrated that such software is possible.\u003C\u002Fp\u003E\u003Cp\u003E然而，当要求系统挑选出它最有把握是同性恋的十张脸时，它选出的十人中有九人的确是同性恋。假如目标是从一大群人中挑出少数很可能是同性恋的人，系统似乎是可以胜任的。重点并不是科辛斯基和王轶伦创造出了能可靠识别性取向的软件——这不是他们的目标，而是他们证明了这样的软件已能够实现。\u003C\u002Fp\u003E\u003Ch2\u003ETo calculate the selves of others\u003C\u002Fh2\u003E\u003Cp\u003EDr Kosinski is no stranger to controversial research. He invented psychometric profiling using Facebook data, which relies upon information in a person’s profile to model their personality. The Trump campaign used similar models during last year’s presidential campaign to target voters, an approach which has generated criticism.\u003C\u002Fp\u003E\u003Ch2\u003E计算他人隐私\u003C\u002Fh2\u003E\u003Cp\u003E科辛斯基并非第一次涉足争议性研究。他之前发明过基于Facebook数据的心理测量分析方法，利用人们个人资料中的信息估测其个性。去年美国总统大选时，特朗普的竞选团队就采用过类似的模型来定位目标选民，这种方法引来了批评。\u003C\u002Fp\u003E\u003Cp\u003EDr Kosinski says he conducted the research as a demonstration, and to warn policymakers of the power of machine vision. It makes further erosion of privacy “inevitable”; the dangers must be understood, he adds. Spouses might seek to know what sexuality-inferring software says about their partner (the word “gay” is 10% more likely to complete searches that begin “Is my husband…” than the word “cheating”). In parts of the world where being gay is socially unacceptable, or illegal, such software could pose a serious threat to safety. Dr Kosinski is at pains to make clear that he has invented no new technology, merely bolted together software and data that are readily available to anyone with an internet connection. He has asked \u003Ci\u003EThe Economist\u003C\u002Fi\u003E not to reveal the identity of the dating website he used, in order to discourage copycats.\u003C\u002Fp\u003E\u003Cp\u003E科辛斯基表示，他做这项研究是要演示机器视觉的威力，并使政策制定者对此有所警惕。这种威力会令个人隐私“不可避免地”被进一步侵蚀，而人们必须要了解这种危险。配偶也许会想知道性取向识别软件会如何判断自己的伴侣（搜索关键词“我丈夫是不是……”时以“同性恋”结尾的几率比“出轨了”高10%）。在同性恋不为社会接受或被视为非法的地区，这样的软件可能会造成严重的安全威胁。科辛斯基竭力强调自己并没有发明什么新技术，只是把互联网上任何人都可以取得的软件和数据结合起来而已。他请《经济学人》对其使用的约会网站保密，以阻止他人效仿。\u003Cbr\u003E\u003C\u002Fp\u003E\u003Cp\u003EIt is true that anyone wishing to replicate Dr Kosinski’s work to determine intimate traits from faces will face significant challenges in applying laboratory science to the outside world. But they will be helped by ever-growing volumes of data and improving algorithms. “The latter, over time, inevitably win,” says Alessandro Acquisti of Carnegie Mellon University, who has shown that an individual’s social security number can be discovered using face recognition and online information. For those with secrets to keep, all this is bad news.\u003C\u002Fp\u003E\u003Cp\u003E诚然，任何人若想复制科辛斯基的研究、通过人脸特征推断私密特质，都会在把实验室技术应用于外部世界时面临重大挑战。但不断增加的数据量和不断改进的算法会帮助他们。“假以时日，他们定会胜出。” 卡内基梅隆大学的亚历山德罗·阿奎斯蒂（Alessandro Acquisti）说道。他已证明，运用人脸识别技术和网上信息就能查出一个人的社保号。对那些有秘密要隐藏的人来说，这一切都是坏消息。\u003C\u002Fp\u003E&,&updated&:new Date(&T02:18:15.000Z&),&canComment&:false,&commentPermission&:&review&,&commentCount&:0,&likeCount&:4,&state&:&published&,&isLiked&:false,&slug&:&&,&isTitleImageFullScreen&:false,&rating&:&none&,&sourceUrl&:&&,&publishedTime&:&T10:18:15+08:00&,&links&:{&comments&:&\u002Fapi\u002Fposts\u002F2Fcomments&},&url&:&\u002Fp\u002F&,&titleImage&:&https:\u002F\\u002Fv2-e0e9bc2063_r.jpg&,&summary&:&&,&href&:&\u002Fapi\u002Fposts\u002F&,&meta&:{&previous&:null,&next&:null},&snapshotUrl&:&&,&commentsCount&:0,&likesCount&:4},&&:{&title&:&经济学人官译：癌细胞上&,&author&:&iqlake&,&content&:&\u003Ch2\u003EUnruly origins\u003C\u002Fh2\u003E\u003Ch2\u003EBeginnings, and how to spot them\u003C\u002Fh2\u003E\u003Ch2\u003EIt all starts with a single cell\u003C\u002Fh2\u003E\u003Ch2\u003E放肆的开端\u003C\u002Fh2\u003E\u003Ch2\u003E开端，以及如何发现它们\u003C\u002Fh2\u003E\u003Ch2\u003E一切都开始于一个细胞\u003C\u002Fh2\u003E\u003Cp\u003E\u003Cbr\u003E\u003C\u002Fp\u003E\u003Cp\u003EDIANE MILLEY, a teacher, remembers getting the small, dry cough just before \u003Cb\u003Eschool broke up（词汇）for the summer\u003C\u002Fb\u003E in 2013. She wasn’t worried: she considered herself generally healthy—she ran three times a week and went to the gym. Her doctor in Bradford, Massachusetts, put her on a course of antibiotics. When they didn’t work she had an X-ray. It showed \u003Cb\u003Enodules（词汇）\u003C\u002Fb\u003Eacross her lungs. A \u003Cb\u003Ebronchoscopy（词汇）\u003C\u002Fb\u003Ewas ordered to retrieve a \u003Cb\u003Etissue sample（词汇）\u003C\u002Fb\u003E from her lung. As she came round from the anaesthetic she remembers overhearing two medical staff talking. One said “It’s \u003Cb\u003Emalignant\u003C\u002Fb\u003E.” She had late-stage lung cancer.戴安·米莉（Diane Milley）是一名教师。她记得2013年\u003Cb\u003E夏天学校临放假前\u003C\u002Fb\u003E，她开始轻微的干咳。她并没有放在心上——她觉得自己总体上很健康，每周跑步三次还去健身房。马萨诸塞州布拉德福德的医生对她用了一个疗程的抗生素。治疗无效后她拍了一张X光片，显示她的肺部存在\u003Cb\u003E结节\u003C\u002Fb\u003E。随后医生让她做\u003Cb\u003E支气管镜检查\u003C\u002Fb\u003E，从她的肺部提取\u003Cb\u003E组织样本\u003C\u002Fb\u003E。当她从麻醉中醒来时，听到两名医务人员在说话。其中一人说：“是\u003Cb\u003E恶性\u003C\u002Fb\u003E的。”她患的是晚期肺癌。\u003C\u002Fp\u003E\u003Cp\u003EMs Milley’s body, like all human bodies, contained tens of trillions of copies of her genome. In theory, all those copies should be more or less the same. In practice, over the years, they all get \u003Cb\u003Eknocked around（句型）\u003C\u002Fb\u003E in different ways. The oxygen that \u003Cb\u003Epowers\u003C\u002Fb\u003E cell \u003Cb\u003Emetabolism（词汇）\u003C\u002Fb\u003Edamages the DNA on which the genes are stored a so do background radiation and exposure to the many low-level \u003Cb\u003Ecarcinogens（词汇）\u003C\u002Fb\u003E; so do sunlight and in so do choices about diet and \u003Cb\u003Erecreational drugs（词汇）\u003C\u002Fb\u003E, notably alcohol and tobacco (from which Ms Milley abstained).和所有人的身体一样，米莉的身体也含有自身基因的数十万亿个副本。理论上说，所有这些副本都应该是基本相同的。然而实际上，经年累月，它们以不同的方式遭到了\u003Cb\u003E破坏\u003C\u002Fb\u003E。为细胞\u003Cb\u003E代谢\u003C\u002Fb\u003E提\u003Cb\u003E供能\u003C\u002Fb\u003E量的氧会自然而然地破坏存储基因的DNA，背景辐射和接触许多低级\u003Cb\u003E致癌物\u003C\u002Fb\u003E也会如此，日晒和病毒感染也会如此，饮食选择和\u003Cb\u003E娱乐性药物\u003C\u002Fb\u003E，特别是酒精和烟草（米莉均不沾），也会如此。\u003C\u002Fp\u003E\u003Cp\u003E\u003Cb\u003EThe vast bulk of（句型）\u003C\u002Fb\u003Ethis damage is quickly fixed by DNA-repair \u003Cb\u003Eenzymes\u003C\u002Fb\u003E; fewer than one mutation in a thousand persists. But \u003Cb\u003Ewear and tear\u003C\u002Fb\u003E builds up. Many such changes make little or no difference. A few will \u003Cb\u003Ebe of consequence to（句型）\u003C\u002Fb\u003Ethe cell concerned, reducing or eliminating its capacity to do its job. But the loss of a single cell’s contribution matters not a jot.\u003Cb\u003E绝大多数\u003C\u002Fb\u003E此类损害都很快被DNA修复\u003Cb\u003E酶\u003C\u002Fb\u003E修复了；一千个突变中只有不到一个会留存下来。但这种\u003Cb\u003E损耗\u003C\u002Fb\u003E会累积。很多此类变化都无足轻重，少数则会\u003Cb\u003E影响\u003C\u002Fb\u003E相关细胞，降低或完全消除其履行职责的能力。但损失一个细胞的贡献造成的影响微不足道。\u003C\u002Fp\u003E\u003Cp\u003EThere are some genes, though, where uncorrected damage can matter a lot. Foremost are the genes which control cell growth, such as HER2, which tells the cell how to make a protein called human \u003Cb\u003Eepidermal-growth-factor\u003C\u002Fb\u003E receptor type 2. This is a protein that, when it sees a particular hormone, tells the cell it is in to divide. Mutations in the HER2 gene can make cells proliferate when there is no need. When they do so their daughter cells, which will share that HER2 \u003Cb\u003Emutation\u003C\u002Fb\u003E, will go on to do the same.然而，对于某些基因而言，未经修复的损害可能就大有关系了。首先是那些控制细胞生长的基因，如HER2，它告诉细胞如何生成一种名为人类\u003Cb\u003E表皮生长因子（词汇）\u003C\u002Fb\u003E受体2的蛋白质。这种蛋白质在遇到某种特定的激素时会指示它所在的细胞进行分裂。HER2基因的突变可能导致细胞在毫无必要的时候增殖 。如果发生这种情况，其拥有同样HER2\u003Cb\u003E突变（词汇）\u003C\u002Fb\u003E的子细胞也会继续做同样的事情。\u003C\u002Fp\u003E\u003Cp\u003EAmong some 20,000 genes in the genome there are dozens which, like HER2, can cause unwanted cell division when they go wrong. To \u003Cb\u003Eforestall\u003C\u002Fb\u003E such problems there are various tumour-suppressor genes whose job is to make sure that cells damaged in this way shut themselves down. The best known is the gene for a protein called p53, which stops cells from reproducing if their DNA is damaged. But these tumour-suppressor genes, too, are \u003Cb\u003Esubject to\u003C\u002Fb\u003E mutation.在基因组中的大约20,000个基因中，有几十个基因会像HER2一样，在出错时可能引起有害的细胞分裂。为了\u003Cb\u003E预防\u003C\u002Fb\u003E这类问题，身体里还有各种各样的肿瘤抑制基因，确保遭到此类损伤的细胞会自行关闭。最著名的是一种名为p53的蛋白质的基因，它可以阻止DNA遭损坏的细胞自我复制。但是，这些肿瘤抑制基因\u003Cb\u003E也可能（句型）\u003C\u002Fb\u003E发生突变。\u003C\u002Fp\u003E\u003Ch2\u003EThe numbers game\u003C\u002Fh2\u003E\u003Cp\u003EThus over time, as genetic damage accumulates, the likelihood rises that somewhere in the body’s trillions of cells there is one that has, through five or six mutations in key genes, developed the ability to \u003Cb\u003Egrow without check\u003C\u002Fb\u003E. This likelihood is not the same for everyone. Some people start off with quirks in their genome which make them more susceptible. Take the genes BRCA1 and BRCA2, which describe proteins that repair DNA; people who inherit a damaged version of one or the other face a higher risk of cancer (in particular, breast and ovarian cancer) because, with one crucial function already \u003Cb\u003Ecompromised\u003C\u002Fb\u003E, it takes fewer mutations for a tumour to get going.\u003C\u002Fp\u003E\u003Ch2\u003E数字游戏\u003C\u002Fh2\u003E\u003Cp\u003E因此，随着时间的推移和基因损伤的累积，在身体里的数万亿个细胞中，有一个细胞因为关键基因发生了五六次突变而能够\u003Cb\u003E无拘无束生长（句型，贬义）\u003C\u002Fb\u003E的可能性就变大了。这种可能性并不是人人平等的。有些人的基因组一开始就有些怪异，这让他们更容易受到影响。拿基因BRCA1和BRCA2来说吧，它们表达的是用于修复DNA的蛋白质。如果一个人继承的其中一个基因受损，就会面临更高的患癌风险 （特别是乳腺癌和卵巢癌）。这是因为，既然一个关键功能已经\u003Cb\u003E受损（词汇，手表翻译no compromise永不磨损）\u003C\u002Fb\u003E，那么只需要更少的突变就可发生肿瘤。\u003C\u002Fp\u003E\u003Cp\u003EOnce a cancer has begun its \u003Cb\u003Eunruly growth\u003C\u002Fb\u003E it will pick up more and more mutations: the cancer genome project at the Sanger Institute, outside Cambridge in England, has found that cancers can have as few as ten mutations or as many as a few hundred. Though all the cells in the cancer are descended from one parent cell, they become increasingly diverse over time. Some cells come loose and start new tumours of their own elsewhere. The body’s immune system will often \u003Cb\u003Erecognise\u003C\u002Fb\u003E that something is \u003Cb\u003Eamiss\u003C\u002Fb\u003E and try to fight the cancer and slow its spread. Sometimes it wins, stopping the cancer or killing it. Sometimes it doesn’t.一旦肿瘤开始\u003Cb\u003E肆意生长（句型）\u003C\u002Fb\u003E，它将会汇聚越来越多的突变。位于英国剑桥镇外的桑格研究所（Sanger Institute）的癌症基因组计划发现，癌症的突变数量少至十个，多达几百个。虽然肿瘤的所有细胞都是一个亲代细胞的子孙，但随着时间的推移 ，它们变得越来越五花八门。一些细胞会松脱，到其他地方再自己开始生长新肿瘤。身体的免疫系统通常会\u003Cb\u003E发现\u003C\u002Fb\u003E有些东西\u003Cb\u003E不对\u003C\u002Fb\u003E了，并且试图与癌症斗争并减缓其蔓延。有时它会赢，阻止甚至杀死癌症。有时它不会。\u003C\u002Fp\u003E\u003Cp\u003EWhen Ms Milley’s cancer was diagnosed all the things that could
the tumour was well developed and had spread through the lung and beyond. It would have been far better for her if it had been diagnosed earlier (see chart). But with lung cancer, as with many other forms of the disease, there are often few symptoms until the disease is already \u003Cb\u003Eat an advanced stage\u003C\u002Fb\u003E. If cancer could be reliably detected earlier, many lives might be saved.当米莉确诊患癌时，一切可能发生的糟糕情况都已发生：肿瘤发育完全，并扩散到整个肺部和身体其他部分。如果能更早诊断出来，她的情况会好得多（见图）。但是在肺癌以及许多其他类型的癌症中，在发展到\u003Cb\u003E晚期（词汇）\u003C\u002Fb\u003E之前通常不会有什么症状。如果我们能够更早地、可靠地检测到癌症，也许能挽救许多生命。\u003C\u002Fp\u003E\u003Cp\u003EIn some wealthy countries, some cancers—for example, those of the breast, \u003Cb\u003Eprostate\u003C\u002Fb\u003E and \u003Cb\u003Ecervix\u003C\u002Fb\u003E—are regularly \u003Cb\u003Esought out\u003C\u002Fb\u003E before they start to cause symptoms. Now researchers are trying to improve diagnostic tools even further, so that more types of cancer can be found early on (and with greater reliability). For some it is a terribly personal hunt. Billy Boyle, the president of a small biotech company, Owlstone Medical, based in Cambridge, in England, is one of them. He lost his wife Kate, mother to their two young boys, on Christmas morning in 2014. She died of colon cancer that had been picked up too late. Mr Boyle says that if colorectal cancer is detected early, 95% of sufferers survive. Only 6% survive if the cancer reaches stage four. For many cancers, early detection is “\u003Cb\u003Eour greatest opportunity\u003C\u002Fb\u003E to improve survival,” says Mr Boyle.在一些富裕国家，人们会在出现症状之前定期\u003Cb\u003E筛查（句型）\u003C\u002Fb\u003E某些癌症（例如乳腺癌、\u003Cb\u003E前列腺（词汇）\u003C\u002Fb\u003E癌和\u003Cb\u003E宫颈（词汇）\u003C\u002Fb\u003E癌）。如今研究人员正在努力进一步改进诊断工具，以便早日（并且以更高的可靠性）发现更多类型的癌症。对于某些人来说，这样的探索是源于极其个人化的经历。位于英国剑桥的小型生物技术公司Owlstone Medical的总裁比利·博伊尔（Billy Boyle）就是其中之一。2014年圣诞节的早晨 ，他失去了他的妻子、两个小男孩的母亲凯特。她死于结肠癌，发现时已经太晚了。博伊尔说，如果结肠直肠癌能够在早期发现，95%的患者能活下来。如果癌症达到第四阶段，生存率只剩6%。博伊尔说，对许多癌症而言，早期监测是“提高生存率的\u003Cb\u003E最大机会\u003C\u002Fb\u003E”。\u003C\u002Fp\u003E\u003Cp\u003EMr Boyle wants to detect cancer on the breath using an ion-mobility spectrometer—a gadget that weighs chemicals by passing them through an oscillating electric field. The breath contains a wide range of organic molecules that reflect what is going on in the body’s metabolism. Cancers, which affect the metabolism, should in so doing change the pattern of molecules on the breath. Although Owlstone’s system is very small—it fits on a chip the size of a coin—it is sensitive, identifying molecules at a level of \u003Cb\u003Ea few parts per billion\u003C\u002Fb\u003E. The firm hopes that when it has identified \u003Cb\u003Emolecular\u003C\u002Fb\u003E “fingerprints” associated with particular cancers it will be able to detect the disease earlier than other tests do.博伊尔想要使用离子迁移光谱仪（一种让化学物质通过振荡电场来称重的小机器）来通过呼吸检测癌症。呼出的气体中含有许多有机分子，反映了身体发生的新陈代谢。癌症会影响新陈代谢，所以应该也会改变呼出气体中分子的模式。虽然Owlstone公司的系统非常小——可以装在硬币大小的芯片上——它却十分敏感，可以在\u003Cb\u003E十亿分之几（词汇）\u003C\u002Fb\u003E的水平上识别分子。该公司希望，在确定了与特定癌症相关的\u003Cb\u003E分子\u003C\u002Fb\u003E“指纹”后，它将能够比其他测试更早地检测到癌症。\u003Cbr\u003E\u003C\u002Fp\u003E\u003Cp\u003EImproved diagnostics can do more than pick up cancers sooner. They can also reveal the cancers’ weaknesses. Because cancer drugs work in different ways, some will do well against a tumour with one set of mutations but leave unscathed one that has become cancerous by some other pathway. Troy Cox, head of Foundation Medicine, a diagnostics company based in Cambridge, Massachusett}