China’s National Data Administration has announced a new plan to make better use of the huge amount of data that businesses and public bodies collect every day. The effort will focus on ten fields, including manufacturing, health care, banking, transport, farming, energy, tourism, education, trade and public services.
Data has become a kind of raw material for the modern economy. Yet much of it sits in separate systems that cannot talk to each other. A hospital may hold useful records that a research team cannot reach. A factory may own machine readings that never leave its own computers. The plan aims to break down these walls.
Under the new approach, the government will push its basic data rules into local areas and key industries. It will also encourage companies, schools and public bodies to form joint teams. Together they will look for valuable uses of data, mixing public information with private and social data to create services that none of them could build alone.
Another goal is to support artificial intelligence. Modern AI systems learn from enormous sets of examples, so the quality of the data matters as much as the design of the model. The administration says it will build better training datasets and improve the way data, models and software agents work together.
The plan is still at an early stage. Real progress will depend on who owns each piece of data and who is allowed to use it. Companies worry about losing control of information that took years to gather, while citizens want to know how their personal details are protected. Clear rules on these points will decide whether the idea works.
Words to know:
- data (n. 数据，资料)
- administration (n. 管理局，行政管理机构)
- industry (n. 行业，工业)
- raw material (n. 原材料)
- dataset (n. 数据集)
- artificial intelligence (n. 人工智能)
- quality (n. 质量，品质)
- citizen (n. 公民，市民)
长难句拆解 · so 结果状语从句 + as … as 同级比较
原文：Modern AI systems learn from enormous sets of examples, so the quality of the data matters as much as the design of the model.
拆解：主干是 the quality of the data matters as much as the design of the model（数据的质量与模型的设计同等重要）。前半句 Modern AI systems learn from enormous sets of examples 是原因，so 引出结果。后半句的核心结构是 as much as，表示「和……一样多、同等重要」，比较的两端分别是 the quality of the data 和 the design of the model。matter 在这里是动词，意思是「要紧、重要」，不是名词「事情」。
直译：现代人工智能系统要从海量样本中学习，因此数据的质量与模型的设计同样重要。
同义替换 · 雅思阅读利器
雅思阅读最爱把原文的词换成另一个说法来出题，认出这些替换，定位就能快一半。
- make better use of → put … to better use（use 既可作名词也可作动词）
- focus on ten fields → centre on / concentrate on ten areas（field 与 area 同义）
- a kind of raw material → a sort of basic input（raw material ＝ 原材料）
- break down these walls → remove these barriers（wall 在此为比喻，＝ barrier）
- encourage … to form joint teams → push them to work together（form joint teams ＝ 组队合作）
- the quality of the data matters → data quality is important（matter 是动词，＝ be important）
- whether the idea works → whether it succeeds in practice（work 在此指奏效）
阅读自测 · 3 题
- T / F / NG：这项攻坚行动覆盖工业制造、医疗健康、金融服务等十大行业。（判断正误并说明依据）
- 简答：按照方案，公共数据要与哪两类数据融合应用？
- 选词填空：Real progress will depend on who ____ each piece of data.（供选：owns / owns to / owning）
参考答案与解析
- TRUE。原文列出 manufacturing, health care, banking, transport, farming, energy, tourism, education, trade and public services，正是题目所说的十大行业，只是顺序不同。做这类题时，顺序调换不算错。
- 企业数据和社会数据。原文说推动公共数据与企业数据、社会数据融合应用（mixing public information with private and social data）。注意干扰项套路：题目若只答「企业数据」就漏了一半，属于「以偏概全」。
- owns。depend on 后面跟的是宾语从句 who owns each piece of data，从句需要有自己的谓语动词，所以用第三人称单数 owns。owns to 和 owning 都不能作从句谓语。
中文译文
中国国家数据局宣布了一项新方案，要更好地利用企业和公共机构每天都在收集的海量数据。这项行动将聚焦十个领域，包括工业制造、医疗健康、金融服务、交通物流、现代农业、能源电力、文化旅游、教育、商贸流通和公共事业。
数据已经成为现代经济的一种原材料。然而，其中很大一部分滞留在彼此无法沟通的独立系统里。一家医院可能握着研究团队拿不到的有用档案；一家工厂可能拥有从未离开自己电脑的机器读数。这项方案的目标，就是推倒这些墙。
按照新的思路，政府将推动国家数据基础制度在地方和重点行业落地。它还会鼓励企业、学校和公共机构组建联合团队，一起寻找数据的高价值用途，把公共数据与企业和社会的私有数据结合起来，做出任何一方单独都做不出的服务。
另一个目标是支撑人工智能。现代人工智能系统要从海量样本中学习，因此数据的质量与模型的设计同样重要。国家数据局表示，将建设更高质量的训练数据集，并改进数据、模型与软件智能体协同工作的方式。
这项方案目前仍处于早期阶段。真正的进展将取决于每一份数据归谁所有、谁有权使用。企业担心失去历时多年才积累起来的信息控制权，而公众则想知道自己的个人信息如何得到保护。这些问题上能否有清晰的规则，将决定这个设想能否奏效。