阅读下列短文,从每题所给的A、B、C、D四个选项中选出最佳选项。If you've ever cooked a complex meal with someone, you know the level of cooperation required. American researchers recently showed that in a virtual kitchen, artificial intelligence (AI) learned how to cooperate with a human just by watching humans work together.There are some ways researchers have already trained AI to work with people. Many approaches involve reinforcement learning, where an AI interacts with an environment online-which can include other AIs or humans—and is rewarded for making decisions that lead to desired outcomes. Google Deep Mind's program Alpha Go, for example, learned the board game Go using reinforcement learning.But starting training an AI from zero to interact with people through complete trial and error can waste a lot of time and can be risky. Another choice is to train one AI to model human behavior, and then use that as a tireless human replacement for another AI to learn to interact with. Researchers have used this method in a simple game. But repeating human behavior in more complex settings can be difficult.Thus in the study, researchers used offline reinforcement learning—a method for developing strategies by analyzing recorded behavior rather than through real-time interaction. Previously, offline reinforcement learning had been used mostly to help virtual robots move or to help AIs solve mazes (迷宫), but here it was applied to the tricky problem of working with humans. Instead of learning by interacting with people, this A I learned by watching human interactions."The research work introduces away for AI to learn to team up with humans, without even practicing with us. It could help improve human-AI interactions." says Nikolaidis, director of a lab.In the future, humans will increasingly work with AI, both online and in the physical world. Sometimes we'll want an AI to silently guide our choices and strategies, like a good teammate who knows our weaknesses. "It solves an important problem how AI can learn to work with people," says Nikolaidis.
(1)
What is the function of reinforcement learning in the passage?
A . It's how AI learned to work with people
B . It's how researchers helped AI learn cooking.
C . It's how Google Deep Mind developed AlphaGo.
D . It's how AI got rewards for making correct choices.
(2)
What is the advantage of offline reinforcement learning?
A . It relies on real-time interaction.
B . It helps AI learn by interacting with people.
C . It takes less time and has less danger.
D . It is realized through mistakes and corrections.
(3)
What is AI expected to do in the future?
A . To cook a complex meal on its own.
B . To point out our weaknesses directly.
C . To offer voiceless guidance sometimes.
D . To solve difficult problems for humans.
(4)
What is the text mainly about?
A . AI learns cooperation by practising.
B . AI learns cooking from humans.
C . AI learns cooking without humans' help.
D . AI learns teamwork by watching humans.
答案: A
C
C
D