VOA慢速英语2021--NASA借助新学习算法发现301颗系外行星

Machine Learning Helps NASA Confirm 301 New Exoplanets

The American space agency NASA says it has used a new technology method to help confirm the existence of 301 new exoplanets.

Exoplanets are planets that orbit stars other than the sun. Before the latest discoveries, NASA had confirmed the existence of more than 4,569 such planets. Thousands of other "candidate" exoplanets have been identified. But these require additional study.

Exoplanets are difficult for telescopes to identify. One reason is that the bright light of the stars they orbit can hide them. The search process can involve looking for decreases in the light level of stars. Such drops could be caused by a planet passing in front of a star.

NASA has used two space telescopes to confirm thousands of exoplanets. The Kepler space telescope was launched in 2009 and operated until October 2018. At that time, NASA announced it was retiring Kepler because the spacecraft had "run out of fuel needed for further science operations."

The other space telescope is called the Transiting¹ Exoplanet Survey Satellite, or TESS. NASA launched TESS in April 2018 to build on Kepler's observations. TESS continues to operate today.

NASA's confirmations² of the 301 new exoplanets were based on data collected by the Kepler space telescope. The data was processed through a machine learning system called ExoMiner.

Machine learning systems are a form of artificial intelligence (AI). They are trained to learn a task over time by being fed huge amounts of data.

In this case, NASA said it used the machine learning method to examine existing data to identify real exoplanets from so-called "imposters."

ExoMiner is powered by data gathered from past efforts to confirm or rule out possible exoplanets. The system was designed to use the same methods that human experts use to confirm new exoplanets.

NASA said the system provides much-needed assistance to scientists who are expertly trained to confirm the existence of such planets. The agency's space telescopes collect data on thousands of stars. It is a huge effort for humans to examine so many stars. ExoMiner is designed to ease that load and improve the accuracy of identifying new exoplanets.

Jon Jenkins is an exoplanet scientist at NASA's Ames Research Center in California. He said in a statement that ExoMiner offers big improvements over other machine learning programs used to identify exoplanets in the past.

The main reason for this, Jenkins said, is that the new system permits scientists to easily confirm ExoMiner's findings.

"There is no mystery as to why it decides something is a planet or not," he said. "We can easily explain which features in the data lead ExoMiner to reject or confirm a planet."

The machine learning system was developed and tested by NASA researchers and the team's international partners. It was described by a paper published in the Astrophysical Journal.

The paper explains that ExoMiner discovered the 301 exoplanets from a list of candidates based on data from the Kepler space telescope. They had been identified and declared as possible exoplanets by scientists at the Kepler Science Operations Center. But NASA said no human researchers had been able to confirm them.

"When ExoMiner says something is a planet, you can be sure it's a planet," said Hamed Valizadegan. He is the ExoMiner project lead and oversees³ machine learning operations at the Universities Space Research Association at the Ames center.

Valizadegan added that the system is "in some ways more reliable" than both existing machine learning methods as well as human experts. He said one reason for this is that ExoMiner is free of "biases⁵" that can affect human identification operations.

The NASA team said it plans to build on ExoMiner's success by expanding the system. The goal would be to include data from TESS and future telescopes that aim to discover new exoplanets.

Words in This Story

artificial intelligence – n. an area of computer science that deals with giving machines the ability to seem like they have human intelligence

imposter – n. someone who pretends to be someone else in order to trick people

accuracy – n. correct or exact

feature – n. a typical quality or important part of something

reliable – adj. able to be trusted or believed

bias⁴ – n. prejudice in favor of or against one thing person or group compared with another

1 transiting   

通过（transit的现在分词形式）

参考例句：

• The effect of the transiting mechanic required reserve system vehicle is low. 准备金制度的传导机制的作用是很低的。
• I was busy transiting to the telescope. 我正忙着旋转望远镜。

2 confirmations   

证实( confirmation的名词复数 )； 证据； 确认； （基督教中的）坚信礼

参考例句：

• Never use transitory dialogs as error messages or confirmations. 绝不要用临时对话框作为错误信息框或确认信息框。 来自About Face 3交互设计精髓
• Dismissing confirmations thus becomes as routine as issuing them. 因此关闭确认对话框和发起确认对话框一样成为例行公事。 来自About Face 3交互设计精髓

3 oversees   

v.监督，监视( oversee的第三人称单数 )

参考例句：

• She oversees both the research and the manufacturing departments. 她既监督研究部门又监督生产部门。 来自《简明英汉词典》
• The Department of Education oversees the federal programs dealing with education. 教育部监管处理教育的联邦程序。 来自互联网

4 bias   

n.偏见，偏心，偏袒；vt.使有偏见

参考例句：

• They are accusing the teacher of political bias in his marking.他们在指控那名教师打分数有政治偏见。
• He had a bias toward the plan.他对这项计划有偏见。

5 biases   

偏见( bias的名词复数 )； 偏爱； 特殊能力； 斜纹

参考例句：

• Stereotypes represent designer or researcher biases and assumptions, rather than factual data. 它代表设计师或者研究者的偏见和假设，而不是实际的数据。 来自About Face 3交互设计精髓
• The net effect of biases on international comparisons is easily summarized. 偏差对国际比较的基本影响容易概括。