双语：Facebook否认AI失控只是忘了告诉它们：“请用英文语法”！

　　Facebook has shut down a controversial chatbot experiment that saw two AIs develop their own language to communicate.

　　近日Facebook紧急关停了一个备受争议的聊天机器人实验，原因是试验中两个AI竟然发明出了自己的语言并进行沟通交流。

　　The social media firm was experimenting with teaching two chatbots, Alice and Bob, how to negotiate with one another.

　　当时Facebook正在进行一个实验，教两个机器人(爱丽丝和鲍勃)如何谈判。

　　However, researchers at the Facebook AI Research Lab (FAIR) found that they had deviated from script and were inventing new phrases without any human input.

　　然而Facebook人工智能研究实验室的研究人员发现，机器人们偏离了既定的脚本，在没有人类输入的情况下自己发明了新的语言。

　　The bots were attempting to imitate human speech when they developed their own machine language spontaneously - at which point Facebook decided to shut them down.

　　在机器人们自发地发明自己的机器语言时，它们还试图模仿人类讲话——而此时Facebook决定将其立即关停。

　　’Our interest was having bots who could talk to people,’ Mike Lewis of Facebook’s FAIR programme told Fast Co Design.

　　来自Facebook人工智能研究实验室的麦克·刘易斯向Fast Co Design表示称：“我们的本意是想研发可以和人类交谈的机器人。”

　　Facebook’s Artificial Intelligence Researchers (Fair) were teaching the chatbots, artificial intelligence programs that carry out automated one to one tasks, to make deals with one another.

　　当时Facebook的人工智能研究人员正在教授两台机器人(这些机器人是AI程序，载有自动一对一任务)如何跟对方做交易。

　　As part of the learning process they set up two bots, known as a dialog agents, to teach each other about human speech using machine learning algorithms.

　　作为学习过程的一部分，他们建造了两个机器人，被称之为对话代理，用机器学习算法互相教授人类的语言。

　　The bots were originally left alone to develop their conversational skills.

　　最开始研究人员让机器人们自己发展会话技巧。

　　When the experimenters returned, they found that the AI software had begun to deviate from normal speech.

　　而当实验人员返回时，他们发现AI软件已经开始偏离正常语言了。

　　Instead, they were using a brand new language created without any input from their human supervisors.

　　事实上，它们正在使用一种全新的语言，而这种语言是在没有任何人类输入的情况下由机器人自己发明的。

　　The new language was more efficient for communication between the bots, but was not helpful in achieving the task they had been set.

　　这种语言在机器人间交流更有效率，但是对于完成既定任务却没有任何帮助。

　　’Agents will drift off understandable language and invent codewords for themselves,’ Dhruv Batra, a visiting research scientist from Georgia Tech at Facebook AI Research (FAIR) told Fast co.

　　来自佐治亚理工学院的巴特拉的Facebook人工智能研究实验室的一名访问研究科学家，他对Fast co表示称：“机器人代理偏离了可理解的语言，发明了自己的语言。”

　　The programmers had to alter the way the machines learned language to complete their negotiation training.

　　程序员们不得不改变机器人学习语言的方式，从而完成既定的谈判训练。

　　Writing on the Fair blog, a spokesman said: ’During reinforcement learning, the agent attempts to improve its parameters from conversations with another agent.

　　在该实验室博客上发布声明时，一位发言人表示：”在强化学习过程中，机器人代理企图通过和其他代理的对话来提高自己的参数。“