在斯坦福大學(xué)的閱讀理解考試中，阿里巴巴開(kāi)發(fā)的人工智能模型成績(jī)超過(guò)了人類。

阿里巴巴上周測(cè)試了深度神經(jīng)網(wǎng)絡(luò)模型，要求AI給出超過(guò)10萬(wàn)道問(wèn)題的答案。該測(cè)試是世界上最權(quán)威的機(jī)器閱讀測(cè)驗(yàn)之一。最后阿里巴巴數(shù)據(jù)科學(xué)與技術(shù)研究院開(kāi)發(fā)的AI得到82.44分，略高于人類對(duì)手的82.304分。

阿里巴巴表示，這是機(jī)器首次在此類測(cè)試中超過(guò)人類。微軟研發(fā)的AI在該測(cè)試中成績(jī)差不多，得分為82.650，只是其成績(jī)確認(rèn)比阿里巴巴的AI晚了一天。

中國(guó)電商巨頭阿里巴巴已開(kāi)始發(fā)力AI開(kāi)發(fā)，跟騰訊、百度等對(duì)手展開(kāi)競(jìng)爭(zhēng)。AI可以豐富社交媒體信息流，實(shí)現(xiàn)廣告和服務(wù)精準(zhǔn)投放，甚至可以協(xié)助自動(dòng)駕駛。中國(guó)政府已經(jīng)在國(guó)家級(jí)規(guī)劃中公開(kāi)肯定AI技術(shù)，還提出到2030年中國(guó)要成為行業(yè)領(lǐng)跑者。

所謂自然語(yǔ)言處理技術(shù)，是指模仿人類理解詞語(yǔ)和句子的方式。斯坦福大學(xué)的測(cè)試內(nèi)容包括500多篇維基百科文章，提出的問(wèn)題旨在評(píng)估機(jī)器學(xué)習(xí)模型能否在處理大量信息后準(zhǔn)確回答問(wèn)題。

阿里研究院自然語(yǔ)言處理首席科學(xué)家司羅發(fā)表聲明稱：“也就是說(shuō)，現(xiàn)在機(jī)器可以準(zhǔn)確地回答客觀問(wèn)題，比如‘為什么會(huì)下雨？’背后的技術(shù)可以逐步用于大量應(yīng)用，比如客戶服務(wù)、博物館導(dǎo)覽和在線解答患者的咨詢，從而大大降低人力投入?！保ㄘ?cái)富中文網(wǎng)）

譯者：Charlie

審校：夏林

Alibaba has developed an artificial intelligence model that scored better than humans in a Stanford University reading and comprehension test.

Alibaba Group Holding (BABA, -0.52%) put its deep neural network model through its paces last week, asking the AI to provide exact answers to more than 100,000 questions comprising a quiz that’s considered one of the world’s most authoritative machine-reading gauges. The model developed by Alibaba’s Institute of Data Science of Technologies scored 82.44, edging past the 82.304 that rival humans achieved.

Alibaba said it’s the first time a machine has out-done a real person in such a contest. Microsoft achieved a similar feat, scoring 82.650 on the same test, but those results were finalized a day after Alibaba’s, the company said.

The Chinese e-commerce titan has joined the likes of Tencent Holdings (TCTZF, +2.68%)and Baidu (BIDU, +1.06%) in a race to develop AI that can enrich social media feeds, target ads and services or even aid in autonomous driving. Beijing has endorsed the technology in a national-level plan that calls for the country to become the industry leader 2030.

So-called natural language processing mimics human comprehension of words and sentences. Based on more than 500 Wikipedia articles, Stanford’s set of questions are designed to tease out whether machine-learning models can process large amounts of information before supplying precise answers to queries.

“That means objective questions such as ‘what causes rain’ can now be answered with high accuracy by machines,” Luo Si, chief scientist for natural language processing at the Alibaba institute, said in a statement. “The technology underneath can be gradually applied to numerous applications such as customer service, museum tutorials and online responses to medical inquiries from patients, decreasing the need for human input in an unprecedented way.”