Andrew Ng is hunched over his smartphone, in apantomime of key-pecking, squinting, typo-riddendiscomfort. “This is how we do it today, he says.

吳恩達(dá)(Andrew Ng)駝著背低著頭，略帶夸張地在他的智能手機(jī)上比劃著不停點(diǎn)擊屏幕、瞇著眼卻仍然錯(cuò)字連篇的那種不自在的樣子。“我們?nèi)缃袷沁@樣做的，他稱。

“And this is how we should be doing it, says the chief scientist for Baidu, China’s largest searchengine. He sits back in his chair, speaking to no one in particular with his phone placed on thetable. The one-finger typing agony of millions of smartphone users should one day become athing of the past, he says. All it would take is the creation of a reasonably accurate, pocket-sized electronic version of a human brain.

“而我們應(yīng)該這樣做，這位百度(Baidu)的首席科學(xué)家稱。他靠在座位上，沒有特定對(duì)象地說著話，手機(jī)放在桌子上。他說，數(shù)百萬智能手機(jī)用戶用一個(gè)手指敲字的痛苦有一天應(yīng)該成為過去。而這只需要?jiǎng)?chuàng)造一種達(dá)到合理精確度、與口袋大小相當(dāng)?shù)碾娮影嫒祟惔竽X。百度是中國最大的搜索引擎。

Mr Ng is an expert in deep learning, a branch of artificial intelligence that focus on teachingcomputers how to talk, listen, read, and think like us. The area is fast becoming a priority forthe world’s biggest technology companies, including Baidu as it tackles the era of the mobileinternet.

吳恩達(dá)是深度學(xué)習(xí)(deep learning)領(lǐng)域的專家，該領(lǐng)域是人工智能的一個(gè)分支，專注于讓計(jì)算機(jī)學(xué)習(xí)如何像我們一樣聽、說、讀、思。由于該領(lǐng)域與移動(dòng)互聯(lián)網(wǎng)時(shí)代緊密相連，它正迅速成為包括百度在內(nèi)的全球最大科技公司的優(yōu)先發(fā)展領(lǐng)域。

“The whole world is switching to mobile devices but no one has created a usable interface toinput into the devices, he says. With the development of artificial intelligence, “soon you’ll beable to order food and just say ‘Can I have some food delivered to my house before I gethome?’ out loud.

“整個(gè)世界都在轉(zhuǎn)向移動(dòng)設(shè)備，但是還沒人創(chuàng)造出向移動(dòng)設(shè)備輸入指令的有用接口，他稱。隨著人工智能的發(fā)展，“很快你將可以在訂購食物時(shí)只需要大聲說一句‘能在我回家前送些食物到我家中嗎?’

“It won’t even feel like technology, it will just be in the background.

“感覺上甚至都不像是科技，而就在后臺(tái)里。

In addition to better voice recognition, AI is being talked about for any number of uses frompredicting advertising clicks to recognising faces.

除了更好的語音識(shí)別，從預(yù)測廣告點(diǎn)擊量到人臉識(shí)別技術(shù)的很多領(lǐng)域都在討論使用人工智能。

Since joining Baidu last year, Mr Ng has been steadily working to implement this vision. A UKnative with Chinese roots, he founded in 2011 Google Brain, the US technology company’s deeplearning project, and led it until he joined the Chinese company last year. Poaching him wasregarded as a coup in the technology world.

自從去年加入百度以來，吳恩達(dá)一直在為實(shí)現(xiàn)這個(gè)愿景而穩(wěn)扎穩(wěn)打。作為一名出生在英國的華人，他在2011年創(chuàng)建了“谷歌大腦(Google Brain)——谷歌的深度學(xué)習(xí)項(xiàng)目，并且在去年加入百度前一直領(lǐng)導(dǎo)著該項(xiàng)目。百度撬走吳恩達(dá)被認(rèn)為是科技界的一次政變。

He describes the advanced computers at Baidu’s Sunnyvale, California, lab as “rocket engineswhose software can be taught to mimic the functioning of the human mind. Their “fuel is data,which he gets from Baidu’s trove of online video and audio output as he works to teach theelectronic brain to listen and speak.

他把百度位于加州森尼韋爾(Sunnyvale)實(shí)驗(yàn)室中的先進(jìn)計(jì)算機(jī)比作“火箭引擎，計(jì)算機(jī)中的軟件可以學(xué)習(xí)模擬人類思想的功能。在吳恩達(dá)教電子大腦聽和說時(shí)，它們的“燃料就是他從百度在線視頻和音頻輸出資料庫中得到的數(shù)據(jù)。

The company has an advantage in deep-learning algorithms for speech recognition in thatmost video and audio in China is accompanied by text — nearly all news clips, television showsand films are close-captioned and almost all are available to Baidu and Iqiyi, its video affiliate.

百度在語音識(shí)別深度學(xué)習(xí)算法方面具有優(yōu)勢(shì)，因?yàn)橹袊蠖鄶?shù)視頻和音頻都伴有文本——幾乎所有資訊剪輯、電視節(jié)目及電影都有詳細(xì)的字幕，而百度及其視頻子公司愛奇藝(Iqiyi)可以獲得幾乎所有此類內(nèi)容。

While a typical academic project uses 2,000 hours of audio data to train voice recognition,says Mr Ng, the troves of data available to China’s version of Google mean he is able to use100,000 hours.

吳恩達(dá)說，一個(gè)典型的學(xué)術(shù)項(xiàng)目會(huì)利用2000小時(shí)的音頻數(shù)據(jù)來訓(xùn)練語音識(shí)別，但百度——中國版谷歌——擁有的龐大數(shù)據(jù)庫意味著他可以利用10萬小時(shí)。

He declines to specify just how much the extra 98,000 hours improves the accuracy of hisproject, but insists it is vital.

他拒絕詳細(xì)說明額外9.8萬小時(shí)在多大程度上提升了其項(xiàng)目的精確度，但堅(jiān)稱這至關(guān)重要。

“A lot of people underestimate the difference between 95 per cent and 99 per cent accuracy.It’s not an ‘incremental’ improvement of 4 per cent; it’s the difference between using itoccasionally versus using it all the time, he says.

“許多人低估了95%精確度與99%精確度之間的區(qū)別。這不是4%的“增量提升;這是偶爾使用與始終使用之間的區(qū)別，他說。

Thanks to the strides made in Chinese language voice recognition — a particular challengebecause of the number of homonyms and the importance of context — Baidu will soon roll outDeepspeech, a voice recognition software similar to Apple’s Siri.

由于在漢語語音識(shí)別方面取得了巨大進(jìn)步(漢語中的大量同音異義詞和語境的重要性使之極具挑戰(zhàn))，百度即將推出Deepspeech——一款類似于蘋果(Apple)的Siri的語音識(shí)別軟件。

Other Chinese companies including Alibaba and Tencent are also making advances in AI, butthanks largely to Mr Ng’s reputation Baidu is now judged by industry experts to be ahead ofits domestic peers, ranking up alongside US rivals Facebook, Google, and IBM.

包括阿里巴巴(Alibaba)、騰訊(Tencent)在內(nèi)的其他中國企業(yè)在人工智能方面也取得了進(jìn)步，但主要得益于吳恩達(dá)的聲望，行業(yè)專家如今認(rèn)為百度要領(lǐng)先于國內(nèi)同行，可與美國競爭對(duì)手Facebook、谷歌和IBM比肩。

“Artificial intelligence is an oligopoly, says Yang Jing, founder of AI Era, an association for theartificial intelligence industry in China. “It’s a game for the titans.

“人工智能是寡頭壟斷行業(yè)，中國人工智能行業(yè)協(xié)會(huì)新智元(AI Era)創(chuàng)始人楊靜說，“這是一個(gè)巨頭間的游戲。

Baidu already saves Rmb17m ($2.7m) per day at its data centres by using deep-learningalgorithms to predict hard drive malfunctions, and it is also using AI to optimise the use ofadvertisements and photos to improve clickthrough rates. It would not reveal how much it isspending on AI development overall.

百度通過在數(shù)據(jù)中心利用深度學(xué)習(xí)算法預(yù)測硬盤故障已經(jīng)可以每天節(jié)省1700萬元人民幣(合270萬美元)，而且還利用人工智能優(yōu)化廣告和相片的使用來提升點(diǎn)擊率。該公司并未透露在人工智能開發(fā)上共計(jì)投入多少資金。

But in spite of lofty long-term ambitions, translating deep learning into money-making projectsis still largely on the horizon.

盡管雄心勃勃，但要將深度學(xué)習(xí)轉(zhuǎn)變成賺錢的項(xiàng)目仍有很長一段路要走。

Mr Ng is undaunted. “There’s no question that [AI] is creating huge economic value; there’s noquestion that this will continue to create huge advances, he says. “There is still a huge gapbetween the way machines learn and the way humans learn.

吳恩達(dá)毫無畏懼。“毫無疑問，(人工智能)正在創(chuàng)造巨大的經(jīng)濟(jì)價(jià)值;毫無疑問，這將繼續(xù)創(chuàng)造巨大的進(jìn)步，他說，“機(jī)器的學(xué)習(xí)方式與人類的學(xué)習(xí)方式之間仍存在巨大差距。

Andrew Ng is hunched over his smartphone, in apantomime of key-pecking, squinting, typo-riddendiscomfort. “This is how we do it today, he says.

吳恩達(dá)(Andrew Ng)駝著背低著頭，略帶夸張地在他的智能手機(jī)上比劃著不停點(diǎn)擊屏幕、瞇著眼卻仍然錯(cuò)字連篇的那種不自在的樣子。“我們?nèi)缃袷沁@樣做的，他稱。

“And this is how we should be doing it, says the chief scientist for Baidu, China’s largest searchengine. He sits back in his chair, speaking to no one in particular with his phone placed on thetable. The one-finger typing agony of millions of smartphone users should one day become athing of the past, he says. All it would take is the creation of a reasonably accurate, pocket-sized electronic version of a human brain.

“而我們應(yīng)該這樣做，這位百度(Baidu)的首席科學(xué)家稱。他靠在座位上，沒有特定對(duì)象地說著話，手機(jī)放在桌子上。他說，數(shù)百萬智能手機(jī)用戶用一個(gè)手指敲字的痛苦有一天應(yīng)該成為過去。而這只需要?jiǎng)?chuàng)造一種達(dá)到合理精確度、與口袋大小相當(dāng)?shù)碾娮影嫒祟惔竽X。百度是中國最大的搜索引擎。

Mr Ng is an expert in deep learning, a branch of artificial intelligence that focus on teachingcomputers how to talk, listen, read, and think like us. The area is fast becoming a priority forthe world’s biggest technology companies, including Baidu as it tackles the era of the mobileinternet.

吳恩達(dá)是深度學(xué)習(xí)(deep learning)領(lǐng)域的專家，該領(lǐng)域是人工智能的一個(gè)分支，專注于讓計(jì)算機(jī)學(xué)習(xí)如何像我們一樣聽、說、讀、思。由于該領(lǐng)域與移動(dòng)互聯(lián)網(wǎng)時(shí)代緊密相連，它正迅速成為包括百度在內(nèi)的全球最大科技公司的優(yōu)先發(fā)展領(lǐng)域。

“The whole world is switching to mobile devices but no one has created a usable interface toinput into the devices, he says. With the development of artificial intelligence, “soon you’ll beable to order food and just say ‘Can I have some food delivered to my house before I gethome?’ out loud.

“整個(gè)世界都在轉(zhuǎn)向移動(dòng)設(shè)備，但是還沒人創(chuàng)造出向移動(dòng)設(shè)備輸入指令的有用接口，他稱。隨著人工智能的發(fā)展，“很快你將可以在訂購食物時(shí)只需要大聲說一句‘能在我回家前送些食物到我家中嗎?’

“It won’t even feel like technology, it will just be in the background.

“感覺上甚至都不像是科技，而就在后臺(tái)里。

In addition to better voice recognition, AI is being talked about for any number of uses frompredicting advertising clicks to recognising faces.

除了更好的語音識(shí)別，從預(yù)測廣告點(diǎn)擊量到人臉識(shí)別技術(shù)的很多領(lǐng)域都在討論使用人工智能。

Since joining Baidu last year, Mr Ng has been steadily working to implement this vision. A UKnative with Chinese roots, he founded in 2011 Google Brain, the US technology company’s deeplearning project, and led it until he joined the Chinese company last year. Poaching him wasregarded as a coup in the technology world.

自從去年加入百度以來，吳恩達(dá)一直在為實(shí)現(xiàn)這個(gè)愿景而穩(wěn)扎穩(wěn)打。作為一名出生在英國的華人，他在2011年創(chuàng)建了“谷歌大腦(Google Brain)——谷歌的深度學(xué)習(xí)項(xiàng)目，并且在去年加入百度前一直領(lǐng)導(dǎo)著該項(xiàng)目。百度撬走吳恩達(dá)被認(rèn)為是科技界的一次政變。

He describes the advanced computers at Baidu’s Sunnyvale, California, lab as “rocket engineswhose software can be taught to mimic the functioning of the human mind. Their “fuel is data,which he gets from Baidu’s trove of online video and audio output as he works to teach theelectronic brain to listen and speak.

他把百度位于加州森尼韋爾(Sunnyvale)實(shí)驗(yàn)室中的先進(jìn)計(jì)算機(jī)比作“火箭引擎，計(jì)算機(jī)中的軟件可以學(xué)習(xí)模擬人類思想的功能。在吳恩達(dá)教電子大腦聽和說時(shí)，它們的“燃料就是他從百度在線視頻和音頻輸出資料庫中得到的數(shù)據(jù)。

The company has an advantage in deep-learning algorithms for speech recognition in thatmost video and audio in China is accompanied by text — nearly all news clips, television showsand films are close-captioned and almost all are available to Baidu and Iqiyi, its video affiliate.

百度在語音識(shí)別深度學(xué)習(xí)算法方面具有優(yōu)勢(shì)，因?yàn)橹袊蠖鄶?shù)視頻和音頻都伴有文本——幾乎所有資訊剪輯、電視節(jié)目及電影都有詳細(xì)的字幕，而百度及其視頻子公司愛奇藝(Iqiyi)可以獲得幾乎所有此類內(nèi)容。

While a typical academic project uses 2,000 hours of audio data to train voice recognition,says Mr Ng, the troves of data available to China’s version of Google mean he is able to use100,000 hours.

吳恩達(dá)說，一個(gè)典型的學(xué)術(shù)項(xiàng)目會(huì)利用2000小時(shí)的音頻數(shù)據(jù)來訓(xùn)練語音識(shí)別，但百度——中國版谷歌——擁有的龐大數(shù)據(jù)庫意味著他可以利用10萬小時(shí)。

He declines to specify just how much the extra 98,000 hours improves the accuracy of hisproject, but insists it is vital.

他拒絕詳細(xì)說明額外9.8萬小時(shí)在多大程度上提升了其項(xiàng)目的精確度，但堅(jiān)稱這至關(guān)重要。

“A lot of people underestimate the difference between 95 per cent and 99 per cent accuracy.It’s not an ‘incremental’ improvement of 4 per cent; it’s the difference between using itoccasionally versus using it all the time, he says.

“許多人低估了95%精確度與99%精確度之間的區(qū)別。這不是4%的“增量提升;這是偶爾使用與始終使用之間的區(qū)別，他說。

Thanks to the strides made in Chinese language voice recognition — a particular challengebecause of the number of homonyms and the importance of context — Baidu will soon roll outDeepspeech, a voice recognition software similar to Apple’s Siri.

由于在漢語語音識(shí)別方面取得了巨大進(jìn)步(漢語中的大量同音異義詞和語境的重要性使之極具挑戰(zhàn))，百度即將推出Deepspeech——一款類似于蘋果(Apple)的Siri的語音識(shí)別軟件。

Other Chinese companies including Alibaba and Tencent are also making advances in AI, butthanks largely to Mr Ng’s reputation Baidu is now judged by industry experts to be ahead ofits domestic peers, ranking up alongside US rivals Facebook, Google, and IBM.

包括阿里巴巴(Alibaba)、騰訊(Tencent)在內(nèi)的其他中國企業(yè)在人工智能方面也取得了進(jìn)步，但主要得益于吳恩達(dá)的聲望，行業(yè)專家如今認(rèn)為百度要領(lǐng)先于國內(nèi)同行，可與美國競爭對(duì)手Facebook、谷歌和IBM比肩。

“Artificial intelligence is an oligopoly, says Yang Jing, founder of AI Era, an association for theartificial intelligence industry in China. “It’s a game for the titans.

“人工智能是寡頭壟斷行業(yè)，中國人工智能行業(yè)協(xié)會(huì)新智元(AI Era)創(chuàng)始人楊靜說，“這是一個(gè)巨頭間的游戲。

Baidu already saves Rmb17m ($2.7m) per day at its data centres by using deep-learningalgorithms to predict hard drive malfunctions, and it is also using AI to optimise the use ofadvertisements and photos to improve clickthrough rates. It would not reveal how much it isspending on AI development overall.

百度通過在數(shù)據(jù)中心利用深度學(xué)習(xí)算法預(yù)測硬盤故障已經(jīng)可以每天節(jié)省1700萬元人民幣(合270萬美元)，而且還利用人工智能優(yōu)化廣告和相片的使用來提升點(diǎn)擊率。該公司并未透露在人工智能開發(fā)上共計(jì)投入多少資金。

But in spite of lofty long-term ambitions, translating deep learning into money-making projectsis still largely on the horizon.

盡管雄心勃勃，但要將深度學(xué)習(xí)轉(zhuǎn)變成賺錢的項(xiàng)目仍有很長一段路要走。

Mr Ng is undaunted. “There’s no question that [AI] is creating huge economic value; there’s noquestion that this will continue to create huge advances, he says. “There is still a huge gapbetween the way machines learn and the way humans learn.

吳恩達(dá)毫無畏懼。“毫無疑問，(人工智能)正在創(chuàng)造巨大的經(jīng)濟(jì)價(jià)值;毫無疑問，這將繼續(xù)創(chuàng)造巨大的進(jìn)步，他說，“機(jī)器的學(xué)習(xí)方式與人類的學(xué)習(xí)方式之間仍存在巨大差距。