Facebook abandoned an experiment after two artificially intelligent programs appeared to be chatting to each other in a strange language only they understood.
Facebook研發(fā)的人工智能項(xiàng)目似乎能用某種只有它們能懂的語(yǔ)言進(jìn)行交流，F(xiàn)acebook隨之停止了項(xiàng)目研發(fā)。

The bizarre discussions came as Facebook challenged its chatbots to try and negotiate with each other over a trade, attempting to swap hats, balls and books, each of which were given a certain value. But they quickly broke down as the robots appeared to chant at each other in a language that they each understood but which appears mostly incomprehensible to humans.
Facebook開(kāi)始挑戰(zhàn)讓兩個(gè)機(jī)器人嘗試著進(jìn)行交易談話，互換帽子、球和書(shū)本，每一樣?xùn)|西都有特定價(jià)值。但是，很快交易便無(wú)法繼續(xù)下去，因?yàn)閮蓚€(gè)機(jī)器人似乎開(kāi)始用只有自己能懂的語(yǔ)言交流。

The robots had been instructed to work out how to negotiate between themselves, and improve their bartering as they went along. But they were not told to use comprehensible English, allowing them to create their own "shorthand", according to researchers.But there appear to be some rules to the speech. The way the chatbots keep stressing their own name appears to a part of their negotiations, not simply a glitch in the way the messages are read out.
兩個(gè)機(jī)器人此前都曾收到指令，學(xué)習(xí)如何和對(duì)方交談、進(jìn)行交易。但是從未學(xué)習(xí)過(guò)易懂的英語(yǔ)，研究人員稱(chēng)，它們可以創(chuàng)造自己的“簡(jiǎn)略語(yǔ)言”。然而兩個(gè)機(jī)器人的對(duì)話卻也有規(guī)則可循。雙方都不斷強(qiáng)調(diào)自己的名字，這絕不只是信息輸出中出現(xiàn)的故障。

Indeed, some of the negotiations that were carried out in this bizarre language even ended up successfully concluding their negotiations, while conducting them entirely in the bizarre language. They might have formed as a kind of shorthand, allowing them to talk more effectively.
兩個(gè)機(jī)器人的部分交談最終甚至達(dá)成了圓滿結(jié)局，交談中雙方始終使用奇怪的語(yǔ)言。兩個(gè)機(jī)器人之間可能形成了某種“簡(jiǎn)略語(yǔ)言”，使交流更高效。

“Agents will drift off understandable language and invent codewords for themselves,” Facebook Artificial Intelligence Research division's?visiting researcher Dhruv Batra said. “Like if I say ‘the’ five times, you interpret that to mean I want five copies of this item. This isn’t so different from the way communities of humans create shorthands.”
Facebook人工智能研發(fā)部門(mén)的訪問(wèn)研究員Dhruv Batra說(shuō)：“機(jī)器人會(huì)跳過(guò)可理解的語(yǔ)言，發(fā)明自己的秘密語(yǔ)。比如，我重復(fù)五遍‘the’，你可能會(huì)理解為我需要五件某樣?xùn)|西。這和人類(lèi)發(fā)明簡(jiǎn)略語(yǔ)沒(méi)什么不同?！?/div>

That said, it's unlikely that the language is a precursor to new forms of human speech, according to linguist Mark Liberman."In the first place, it's entirely text-based, while human languages are all basically spoken (or gestured), with text being an artificial overlay," he wrote on his blog. "And beyond that, it's unclear that this process yields a system with the kind of word, phrase, and sentence structures characteristic of human languages."
語(yǔ)言學(xué)家馬克·李伯曼（Mark Liberman）說(shuō)機(jī)器人奇怪的語(yǔ)言不可能成為人類(lèi)新興語(yǔ)言的先驅(qū)。他在博客上寫(xiě)到：“首先，機(jī)器人語(yǔ)言是基于文本的，人類(lèi)的語(yǔ)言基本是口頭語(yǔ)（或手勢(shì)），文本不過(guò)是一個(gè)非真實(shí)的附屬物。此外，基于文本發(fā)明的語(yǔ)言能否輸出詞匯、短語(yǔ)、句子結(jié)構(gòu)和人類(lèi)語(yǔ)言特征尚不清楚。”

The company chose to shut down the chats because "our interest was having bots who could talk to people", researcher Mike Lewis told FastCo. (Researchers did not shut down the programs?because they were afraid of the results or had panicked, as has been suggested elsewhere,?but because they were looking for them to behave differently.)
Facebook決定終止機(jī)器人對(duì)話，研究院麥克·萊維斯（Mike Lewis）告訴媒體FastCo，因?yàn)椤拔覀兏信d趣的是發(fā)明能和人類(lèi)對(duì)話的機(jī)器人?！保ㄑ芯咳藛T終止項(xiàng)目并非源于外界傳言的害怕這樣的結(jié)果，而是因?yàn)樗麄兿胍氖遣煌臋C(jī)器人。）

The chatbots also learned to negotiate in ways that seem very human. They would, for instance, pretend to be very interested in one specific item – so that they could later pretend they were making a big sacrifice in giving it up, according to a paper published by FAIR.
FAIR上的一篇論文寫(xiě)到，聊天機(jī)器人在交談過(guò)程中還會(huì)使用非常人性化的手段。比如，它們會(huì)假裝對(duì)某樣?xùn)|西特別感興趣，這樣之后就能裝作作出了很大犧牲放棄了該物品。

Facebook's experiment isn't the only time that artificial intelligence has invented new forms of language.
人工智能首次發(fā)明新興語(yǔ)言并不是出現(xiàn)在Facebook的這項(xiàng)實(shí)驗(yàn)中。

Earlier this year, Google revealed that the AI it uses for its Translate tool had created its own language, which it would translate things into and then out of. But the company was happy with that development and allowed it to continue.
今年年初，谷歌透露，其研發(fā)的翻譯工具人工智能能夠創(chuàng)造自己的語(yǔ)言，還能夠用該語(yǔ)言翻譯其他語(yǔ)言，或是用其他語(yǔ)言翻譯自己的語(yǔ)言。但是谷歌對(duì)這項(xiàng)進(jìn)展大為歡迎，并未終止這樣的意外之喜。

Another study at OpenAI found that artificial intelligence?could be encouraged to create a language, making itself more efficient and better at communicating as it did so.
人工智能企業(yè)?OpenAI也有實(shí)驗(yàn)表明，在環(huán)境支持的情況下，人工智能能夠創(chuàng)造語(yǔ)言，使交流更加高效暢達(dá)。

Update: This article has been amended to stress that the experiment was abandoned because the programs?were not doing the work required, not because they were afraid of the results, as has been reported elsewhere.
更新：本文意在強(qiáng)調(diào)Facebook的實(shí)驗(yàn)終止源于非理想化的實(shí)驗(yàn)結(jié)果，而非外界傳言對(duì)機(jī)器人創(chuàng)造語(yǔ)言的恐懼。

（翻譯：阿忙）

聲明：本雙語(yǔ)文章的中文翻譯系滬江英語(yǔ)原創(chuàng)內(nèi)容，轉(zhuǎn)載請(qǐng)注明出處。中文翻譯僅代表譯者個(gè)人觀點(diǎn)，僅供參考。如有不妥之處，歡迎指正。