未来谁是职场抢手人才
美国学者格斯特林表示，5年内所有软件应用都将内置智能，这样一来，谁会成为市场上炙手可热的关键人才？
测试中可能遇到的词汇和知识：
mint 铸造，铸币[mɪnt]
inbuilt 内置的；内藏的；嵌入的['ɪnbɪlt]
ubiquity 普遍存在；到处存在[juː'bɪkwətɪ]
analogue 类似情况；对等的人['ænəlɒg]
wholesale 批发的；大规模的['həʊlseɪl]
truism 众所周知；真实性['truːɪz(ə)m]
阅读马上开始，建议您计算一下阅读整篇文章所用的时间，对照下方的参考值就可以评估出您的英文阅读水平。
如果您读完全文用时为： 那么，您的阅读速度相当于 每分钟阅读的英文单词数
4分8秒 母语为英语者的朗读速度 140
2分33秒 母语为英语的中学生的阅读速度 250
1分15秒 母语为英语的大学生的阅读速度 350
0分8秒 母语为英语的速读高手 1000
Data scientists at forefront of changes in technology businesses （684words）
By Richard Waters
-----------------------------------------------------
For a field supposedly starved of talent, data science seems to have been minting a lot of new experts in a hurry.
The depth of interest was on display this week in San Francisco, where 1,600 people turned up for a data science summit organised by Turi, a company run by University of Washington machine learning professor Carlos Guestrin.
Mr Guestrin argues that all software applications will need inbuilt intelligence within five years, making data scientists — people trained to analyse large bodies of information — key workers in this emerging “cognitive” technology economy. Whether or not he is right about the coming ubiquity, there is already a core of critical applications that depend on machine learning, led by recommendation programmes, fraud detection systems, forecasting tools and applications for predicting customer behaviour.
The adaptation of what was until recently the preserve of research scientists into production-grade business applications could point to a profound change in corporate competitiveness. The companies showing off their skills in data science and machine learning at the Turi event — including Uber, Pinterest and Quora — were all born in the digital era.
Some companies that grew up in the analogue world, such as Walmart, are also investing massively in this field, says Anthony Goldbloom, chief executive of Kaggle, a company that runs online data science competitions. But he predicts that they are unlikely ever to catch up with Amazon and its ilk, which have a head start and are moving fast. Repeated across different sectors, that could point to wholesale change in industry leadership as intelligent systems take a more central role.
One factor weighing on many traditional companies will be the high cost of mounting a serious machine-learning operation. Netflix is estimated to spend $150m a year on a single application — its movie recommendation system — and the total bill is probably four times that once all its uses of the technology are taken into account, says a person familiar with its applications.
Many companies that were born digital — particularly internet outfits that have a deluge of real-time customer interactions to mine — are all-in when it comes to data science. Pinterest, for instance, maintains more than 100 machine learning models that could be applied to different classes of problems, and it constantly fields requests from managers eager to use this resource to tackle their business problems, says Jure Leskovec, its chief scientist.
Another problem for many non-technology companies is talent. Despite the surging ranks of data scientists, some skills are in very short supply, particularly in deep learning — the most advanced form of machine learning. Of the freelance computer science experts who use Kaggle, only about 1,000 have deep learning skills, compared to 100,000 who can apply other machine learning techniques, says Mr Goldbloom. He adds that big companies are often reluctant to bend their pay scales to hire the top talent in this field, even if the algorithms developed by a single high-paid expert can have a disproportionately large effect on their business.
The biggest barrier to adapting to the coming era of “smart” applications, however, is likely to be cultural. Some companies, such as General Electric, have been building their own Silicon Valley presence to attract and develop the digital skills they will need. But they will have to push their new data scientists and machine-learning experts out into operating divisions and bring them closer to line managers to reap the full benefits.
This confluence, between the science and the business practice is critical. It has become a truism to say that all managers will need to let their decision-making be led by the data from now on. But that requires a complete change in mindset that is easier said than done. The challenge is made even harder, says Mr Goldbloom, by the fact that managers are required to redesign their work processes around the new “smart” applications, in ways that effectively design themselves partly out of a job.
Despite the obstacles, some may master this difficult transition. But companies that were built, from the beginning, with data science and machine learning at their core, are likely to represent serious competition.
请根据你所读到的文章内容，完成以下自测题目：
1. What will be needed in the software applications within five years as mentioned?
A. inbuilt intelligence
B. big data
C. linguistics
D. virtual reality
2. What are most recommendation programmes and fraud detection systems depended on now?
A. linguistics
B. manual labour
C. machine learning
D. positioning system
3. Which one is obstacle for many traditional companies to popularize learning operation?
A. technological problem
B. difference of opinion
C. high cost
D. talent crisis
4. What is the biggest barrier to adapting to the coming era of “smart” applications?
A. money
B. cultural
C. time
D. talent
[1] 答案 A. inbuilt intelligence
解释：文章提到所有软件应用在5年内都将需要内置的智能，从而使数据科学家——经过培训，能够对海量数据进行分析的人员——成为这一新兴“认知”技术经济中的关键工作者。
[2] 答案 D. positioning system
解释：目前已有一些核心的关键应用依赖机器学习，最主要的是推荐程序、欺诈探测系统、预报工具和旨在预测顾客行为的应用。
[3] 答案 C. high cost
解释：拖累许多传统公司的一个因素，是开展真正的机器学习运作的高成本。
[4] 答案 B. cultural
解释：适应即将到来的“智能”应用时代的最大障碍，可能是文化上的。
