挑戰(zhàn)與希望:AI 2.0時代從大數(shù)據(jù)到知識(英文)
發(fā)布時間:2018-01-07 00:06
本文關(guān)鍵詞:挑戰(zhàn)與希望:AI 2.0時代從大數(shù)據(jù)到知識(英文) 出處:《Frontiers of Information Technology & Electronic Engineering》2017年01期 論文類型:期刊論文
更多相關(guān)文章: 深度推理 知識庫擴充 強人工智能 大數(shù)據(jù) 跨媒體
【摘要】:本文對大數(shù)據(jù)時代人工智能領(lǐng)域近期出現(xiàn)的若干理論和技術(shù)進展進行了綜述。我們認為,將數(shù)據(jù)驅(qū)動機器學習方法與人類的常識先驗與隱式直覺有效結(jié)合起來,可實現(xiàn)可解釋、更魯棒和更通用的人工智能。AI 2.0時代大數(shù)據(jù)人工智能具體表現(xiàn)為:從淺層計算到深度神經(jīng)推理;從單純依賴于數(shù)據(jù)驅(qū)動的模型到數(shù)據(jù)驅(qū)動與知識引導(dǎo)相結(jié)合學習;從領(lǐng)域任務(wù)驅(qū)動智能到更為通用條件下的強人工智能(從經(jīng)驗中學習)。下一代人工智能(AI 2.0)將改變計算本身,將大數(shù)據(jù)轉(zhuǎn)變?yōu)橹R以支持人類社會作出更好決策。
[Abstract]:This paper reviews the progress of theory and technology in artificial intelligence field, the era of big data recently appeared. We believe that the data driven prior knowledge of machine learning and implicit method of intuition and human can be effectively combined, can realize the explanation, more robust and more general AI.AI 2 big data era of artificial intelligence specific performance as from the shallow depth of the neural computing to data from simple reasoning; to rely on data driven and model driven combination of learning and knowledge leads to more intelligent; strong artificial intelligence general condition from the field of task driven (learning from experience). The next generation of artificial intelligence (AI 2) will change the calculation itself and the big data into knowledge to support the human society to make better decisions.
【作者單位】: College
【分類號】:TP18
【正文快照】: 1 Introduction The idea of inanimate objects coming to life as intelligent beings has been around for a long time.Modern artificial intelligence(AI)was formally founded in 1956 at a workshop at Dartmouth College,where the term‘a(chǎn)rtificial intelligence’
【相似文獻】
相關(guān)會議論文 前1條
1 王佳;;對中文屋思想實驗四個主要版本的考察[A];第三屆全國科技哲學暨交叉學科研究生論壇文集[C];2010年
相關(guān)重要報紙文章 前1條
1 毛向輝;Google的邊界[N];21世紀經(jīng)濟報道;2006年
相關(guān)碩士學位論文 前1條
1 陳成海;評塞爾的中文屋論證[D];浙江大學;2010年
,本文編號:1390108
本文鏈接:http://sikaile.net/kejilunwen/zidonghuakongzhilunwen/1390108.html
最近更新
教材專著