府中市で高校・大学・大学院受験 英語塾 英検 TOEIC TOEFL アメリカ留学 英文読解の透視図なら とも英語塾府中 小金井 国分寺 東府中 調布からも便利です
CROWN III The Biggest EventADLINE
CROWN III The Biggest Event in Human Historyの定期テスト対策
CROWN III The Biggest Event in Human Historyk
The term “artificial intelligence” was first used by a group of scientists who gathered at Dartmouth College in 1956. They ①( idea / of / simulate / up / of / building / every / aspect / machines / that / came / with / the) human intelligence. Ever since that time, scientists have been making efforts to create such machines. According to Matsuo Yutaka, an AI expert at the University of Tokyo. AI research has developed in three stages. From the late 1950s through the 1960s, researchers ②(on / could / and / play / computer / developing / puzzles / programs / that / concentrated / solve) board games. In the 1980s, they shifted their attention to creating “expert systems,” programs designed to incorporate expert knowledge into computers. In the 2010’s, ③(systems / learn / own / to creating / their / could / researchers’ / on / shifted / which / interest), systems with “machine learning” capabilities.
Machine learning has now reached a new stage with the development of “deep learning.” Deep learning tries to replicate what is going on in human brains. In other words, computer systems are becoming more like human neural networks. Just as we learn by using neural networks, computer systems learn ④(them / amounts / make / data, / finding / applying / of / to / patterns / of / and / to / sense / vast / eventually) various tasks.
Computer systems ⑤( never / have / a / we / dreamed / of / number / accomplished / of / feats / using / learning / which / deep), say, 30 years ago. Object recognition is an essential human function―for example, our capacity to see a cat and recognize it as such. Obviously, this is very simple task for us humans, but not for a computer. Think about the fact that no two cats are exactly the same―size, weight, colors and shape of the tail vary widely. Thus, for a computer system to be able to recognize a cat, it ⑥(some / characterize / differences / while / essential / be / features / focusing / must / to / ignore / on / able / which ) a cat.
In fact, in 2012 researchers at Google carried out an experiment where the computer system received 10 million cat images from the Internet. The neural networks ⑦(autonomously / abstracted / common / shared / by patterns / which / of / the / system / certain / are / abstracted ) cat images, and then created a whole new cat.
If you are asked who painted the portraits shown below, you will most likely to answer, “Rembrandt.” But the answer is not quite correct. Actually, one was painted by Rembrandt, but the other was painted by a computer system. A Dutch team ⑧(learning / in / that / style / could / the / create / an / “new” / built / a / work / deep / program / of / entirely) Rembrandt.
First, the computer system ⑨(learning / 346 / them / using / deep / and / scanned / all / of / Rembrandt’s / analyzed / technology / paintings). Next, the computer was told what features should be incorporated in the painting: the subject should be a 30- to 40-year -old male with a beard and wearing a hat. Then, the computer processed the data and created a fully formed face and bust in the style of Rembrandt. Finally, a 3-D printer, using a deep learning program, painted the work. In the spring of 2016, this new “Rembrandt” surprised art lovers all over the world.
Another feat of deep learning took place in the field of board games. In March 2016, Alphago, software created by Demis Hassabis, surprised the world by defeating Lee Sedol, a South Korean Go champion, by 4-1. This ⑩(rely / did / so-called / it / preprogramed, / on / nor / software / was / innovative / not) “brute force” search. Hassabis explains: “Most AI programs are programed directly with the solutions, and then the program executes that solution. With Alphago, it’s very different. We use neural networks to allow Alphago to learn, first of all, by looking at professional games and learning the kinds of patterns professional human players make. And then it tries to get better than human players by practicing through self-play―⑪(mistakes / of / own / learning / from / its / against / itself / playing / older / versions / and). It makes a mistake that leads to a loss, it will slightly change its system to make the chance of that mistake slightly less in the future.”
The term “artificial intelligence” was first used by a group of scientists who gathered at Dartmouth College in 1956. They ①( idea / of / simulate / up / of / building / every / aspect / machines / that / came / with / the) human intelligence. Ever since that time, scientists have been making efforts to create such machines. According to Matsuo Yutaka, an AI expert at the University of Tokyo. AI research has developed in three stages. From the late 1950s through the 1960s, researchers ②(on / could / and / play / computer / developing / puzzles / programs / that / concentrated / solve) board games. In the 1980s, they shifted their attention to creating “expert systems,” programs designed to incorporate expert knowledge into computers. In the 2010’s, ③(systems / learn / own / to creating / their / could / researchers’ / on / shifted / which / interest), systems with “machine learning” capabilities.
Machine learning has now reached a new stage with the development of “deep learning.” Deep learning tries to replicate what is going on in human brains. In other words, computer systems are becoming more like human neural networks. Just as we learn by using neural networks, computer systems learn ④(them / amounts / make / data, / finding / applying / of / to / patterns / of / and / to / sense / vast / eventually) various tasks.
Computer systems ⑤( never / have / a / we / dreamed / of / number / accomplished / of / feats / using / learning / which / deep), say, 30 years ago. Object recognition is an essential human function―for example, our capacity to see a cat and recognize it as such. Obviously, this is very simple task for us humans, but not for a computer. Think about the fact that no two cats are exactly the same―size, weight, colors and shape of the tail vary widely. Thus, for a computer system to be able to recognize a cat, it ⑥(some / characterize / differences / while / essential / be / features / focusing / must / to / ignore / on / able / which ) a cat.
In fact, in 2012 researchers at Google carried out an experiment where the computer system received 10 million cat images from the Internet. The neural networks ⑦(autonomously / abstracted / common / shared / by patterns / which / of / the / system / certain / are / abstracted ) cat images, and then created a whole new cat.
If you are asked who painted the portraits shown below, you will most likely to answer, “Rembrandt.” But the answer is not quite correct. Actually, one was painted by Rembrandt, but the other was painted by a computer system. A Dutch team ⑧(learning / in / that / style / could / the / create / an / “new” / built / a / work / deep / program / of / entirely) Rembrandt.
First, the computer system ⑨(learning / 346 / them / using / deep / and / scanned / all / of / Rembrandt’s / analyzed / technology / paintings). Next, the computer was told what features should be incorporated in the painting: the subject should be a 30- to 40-year -old male with a beard and wearing a hat. Then, the computer processed the data and created a fully formed face and bust in the style of Rembrandt. Finally, a 3-D printer, using a deep learning program, painted the work. In the spring of 2016, this new “Rembrandt” surprised art lovers all over the world.
Another feat of deep learning took place in the field of board games. In March 2016, Alphago, software created by Demis Hassabis, surprised the world by defeating Lee Sedol, a South Korean Go champion, by 4-1. This ⑩(rely / did / so-called / it / preprogramed, / on / nor / software / was / innovative / not) “brute force” search. Hassabis explains: “Most AI programs are programed directly with the solutions, and then the program executes that solution. With Alphago, it’s very different. We use neural networks to allow Alphago to learn, first of all, by looking at professional games and learning the kinds of patterns professional human players make. And then it tries to get better than human players by practicing through self-play―⑪(mistakes / of / own / learning / from / its / against / itself / playing / older / versions / and). It makes a mistake that leads to a loss, it will slightly change its system to make the chance of that mistake slightly less in the future.”
- 下線部①から⑪を文意に沿って意味が通るように並べ替えなさい。
copyright©2005-Present 
Tomoeigojuku Fuchu all rights reserved.