2021,百度飛槳交出最新成績單

{"type":"doc","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"12月12日,在2021深度學習開發者大會現場,百度首席技術官、深度學習技術及應用國家工程實驗室主任王海峯公佈了百度深度學習框架飛槳的最新成績單:"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"截止目前,飛槳已經凝聚了406萬開發者,創建了47.6萬模型,服務於15.7萬企事業單位。"}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"發佈十大最新技術和生態進展"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"大會現場,飛槳發佈十大最新技術和生態進展,包括飛槳新版全景圖—產業級模型庫新增文心大模型、業界首個產業實踐範例庫、飛槳“大航海”2.0共創計劃、飛槳開源框架v2.2 體系化新增科學計算API、端到端自適應大規模分佈式訓練技術、文本任務全流程加速、多層次低成本的硬件適配方案、產業級開源模型庫模型超過400個、企業版升級自動高效的模型部署功能,以及1分鐘極速安裝完成本地高效建模的飛槳EasyDL桌面版。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https:\/\/static001.geekbang.org\/infoq\/e2\/e2cdab98f1decfe56f7157a0798c9526.png","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"具體來說,技術層面,飛槳全新發布的開源框架v2.2,具備四大特性:飛槳API更加豐富、高效、兼容,新增大量科學計算API;高效支持超大模型訓練的端到端自適應大規模分佈式訓練技術;全流程加速文本任務,解決文本領域開發在性能和訓推一體方面的痛點問題;多層次、低成本的硬件適配方案,降低框架與芯片的適配成本。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https:\/\/static001.geekbang.org\/infoq\/7b\/7b53e0911220c0095e79e06ca83b49fb.png","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"飛槳產業級模型庫新增百度最新發布的知識增強文心大模型,讓大模型進入產業應用;官方支持的產業級開源算法模型超過400個,併發布13個PP系列模型,在精度和性能上達到平衡,將推理部署工具鏈打通。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"產業落地方面,飛槳推出了業界首個產業實踐範例庫,從真實產業場景分析、完整代碼實現,到詳細過程解析,直達項目落地,覆蓋數十個高頻應用場景,推動AI落地可複製和規模化。面向產業場景提升開發效率和資源使用效能的飛槳企業版升級了自動高效的模型部署功能,同時推出1分鐘極速安裝完成本地高效建模的飛槳EasyDL桌面版。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"生態方面,飛槳全新發布“大航海”計劃2.0,在啓航、護航、領航三大航道基礎上,新增“共創”計劃,以飛槳平臺爲基座,社區開發者共創工具、模型、產業案例與實踐經驗;形成產業創新需求對接平臺,共創產學研用正循環;與生態夥伴一起建設人工智能產業賦能中心,共創區域創新生態。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"百度集團副總裁、深度學習技術及應用國家工程實驗室副主任吳甜表示,全國各城市的AI開發者規模逐年增長,應用人工智能的企業數量,也呈現了多地開花、多行業繁榮的景象。從推薦等通用場景,到客服系統智能派單等行業衍生場景,再到發電預測等行業關鍵場景……人工智能與產業場景的結合越來越深入、越來越專業。"}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"王海峯:人工智能工業大生產階段,具備融合創新和降低門檻兩大特點"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"深度學習框架下接芯片,上承應用,相當於智能時代的操作系統,對於AI技術發展和產業落地至關重要。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"“深度學習推動人工智能進入工業大生產階段。面向技術和產業發展需求的AI大生產平臺可以讓AI技術以標準化、自動化和模塊化的方式輸出給千行百業,實現規模化應用,同時以平臺爲基礎促進融合創新、共同發展。飛槳產業級深度學習開源開放平臺是典型的AI大生產平臺,賦能廣大開發者,有力支撐AI工業大生產,促進技術創新和產業智能化升級。”王海峯表示。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"王海峯認爲,目前,人工智能呈現出“融合創新”和“降低門檻”的特點:"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"一方面,AI技術及產業的融合創新越來越多。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"融合創新包括知識與深度學習融合、跨模態多技術融合、軟硬一體融合、技術與場景融合。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"知識與深度學習融合方面,百度剛剛發佈的全球首個知識增強千億大模型—鵬城-百度·文心,參數規模達到2600億,在60多項任務中取得最好效果。知識增強千億大模型的發佈,也得益於飛槳的大規模訓練技術。飛槳研製了自適應大規模分佈式訓練技術,使大模型訓練效率達到業界最好水平。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"跨模態多技術融合方面,機器需要像人類一樣理解真實世界,聽懂語音、看懂圖像視頻、理解語言。百度研製了知識增強跨模態大模型,提升跨模態理解與生成能力。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"軟硬一體融合方面,芯片和深度學習框架綜合算力、功耗、時延等因素聯合優化,才能獲得AI應用的最佳效果。飛槳已經和22家國內外硬件廠商完成了31款芯片的適配和聯合優化工作。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"技術與場景融合方面,飛槳在應用於各行各業的過程中,與實際應用場景融合創新,既切實解決了行業應用問題,又使飛槳平臺得到持續積累和提升。飛槳還專門推出系列科學計算API,支持量子計算、生命科學等應用,探索AI與前沿科技的融合創新。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"另一方面,雖然AI技術越來越複雜,但AI開發與應用的門檻卻越來越低。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"技術方面,近期成爲AI技術重要方向的大模型,具有通用性好、泛化性強、效果好等特點,可降低AI開發與應用門檻。飛槳在覈心框架上實現的動靜統一,兼顧科研開發的靈活和產業開發的高效,可滿足不同類型開發者需要。"}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"人工智能人才培養的機遇和挑戰"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"人工智能行業,無論是技術發展,還是產業升級,歸根結底要靠人才實現。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"吳甜現場分享了百度在與吉林大學攜手飛槳探索產教融合協同育人路徑,實現產學研用正循環的實例。吉大三創(創意、創新、創業)實驗室團隊針對醫藥企業藥瓶缺陷檢測場景需求,基於飛槳平臺創新研發了一套全方位、高精度的自動化檢測系統,並實現了產線落地。吉大與飛槳共建了CV及AI+質檢課程體系,將課程資源全面開放,通過飛槳學習與實訓社區AI Studio,將這一實踐經驗輸送給百萬開發者,實現了產學研用正循環。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"王海峯認爲,應該培養複合型AI人才。隨着AI與產業的深度融合,需要越來越多既懂AI,又具備產業經驗的複合型AI人才。一直以來,飛槳致力於複合型AI人才培養,聯合學術界和產業界,打通產學研用的正向循環,持續爲社會貢獻AI人才。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在大會圓桌論壇環節,百度研究院大數據實驗室和商業智能實驗室主任、美國俄勒岡大學計算機系教授竇德景,復旦大學特聘教授、原復旦大學計算機科學技術學院、軟件學院院長王曉陽,上海大學計算機工程與科學學院院長謝少榮,華東師範大學計算機科學與技術學院副院長、華東師範大學上海智能教育研究院副院長周愛民等多位在人工智能人才培養領域中有深入研究,在產教融合的道路上有深刻見解的教育專家,共同探討了現階段,AI人才培養的機遇與挑戰。"}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"AI人才培養,如何實現產教融合?"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"blockquote","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"竇德景:人工智能和產業結合日益緊密,從高校的角度來看,企業應該如何發揮技術和產業的實踐優勢,助力高校人才培養實現產教融合?"}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"周愛民:"},{"type":"text","text":" 企業和高校各自職責不同。對於高校來說,特別是對於研究生和博士階段的人才培養來說,專業碩士、非全日制碩士培養方面能夠和企業形成良好互動,企業可以建立相關的實習基地,可以邀請企業導師發揮各自的優勢。高校在人才培養上有更加全面的知識體系,覆蓋面更全,特別是人工智能爲代表的新工科的專業實踐性更強,對人才的需求非常大,因此在這一塊可以發揮高校和企業界的優勢聯合來培養人才。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"王曉陽:"},{"type":"text","text":" 業界常有人發問,爲何高校教育出來的人與產業界的應用總是脫節。這其中的一個原因是,高校教育和產業的運行節奏是不同的,需要相互配合。談到企業如何助力高校人才培養,第一企業應該要有耐心,很多時候不能說一種新技術出來了,就單純覺得學校培養的人不適合企業發展。第二,應該探索出一種共同育人的模式。從很早開始,企業希望影響學校教學,除了在專業、學位上更深入外,在通用技術方面也要多做努力。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"通用技術較容易在學校推動。很多企業想推廣產品,讓學生使用它的產品,培養潛在的用戶,其實這有點太急功近利。應該把着眼點放在如何把更通用的技術跟學校的教育結合起來。因爲整個計算機科學比較深,在哪裏區分,在某些專業推行哪一些技術,需要做研究,需要長期耕耘。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"謝少榮:"},{"type":"text","text":" 高校對本科生有系統知識的培養。特別是高年級以後的學生,有了共性的專業基礎後,要注意跟應用結合。在這時,高校學生培養希望跟企業的一些很新的、很先進的算法模型,開源平臺等結合起來。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"曾經有一個院士講,高校裏面不要用昨天的知識教今天的學生,讓他們去面對明天的需求。所以我們始終在思考這個問題,也在做各種探索,產教融合應該把最新研發的東西及時納入到專業培養裏面來,這是很有必要的一個環節。也期望像百度這樣的企業多給予高等教育支持和幫助,共同育人,不止是算法和平臺,還包括一些研發工程師甚至可以走進課堂、走進高校,一起來培養學生,培養未來的人工智能人才。"}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"現階段的高校課程設置要調整嗎?"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"blockquote","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"竇德景:產業發展新階段,國內高校各有特色,側重點不一樣,但所有的高校都得教課,咱們在制定教學大綱的時候,編程能力和AI的基礎知識,應該做相應的調整嗎?"}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"謝少榮:"},{"type":"text","text":" 上海大學設計人工智能專業培養體系的時候進行了很多調研、分析和思考,我們的培養理念是,首先讓學生打好基礎,在數學模塊、專業基礎模塊的基礎必須打牢打紮實。我們針對專業課分了六大模塊,要有基礎理論人才培養類型,也要有AI+複合型人才的培養類型,其中有兩個模塊是偏向於通用公共基礎理論方面。我自己是搞智能無人系統的,其中有一個模塊就是結合智能無人系統來進行。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"王曉陽:"},{"type":"text","text":" 計算機行業,從最底層到抽象層、應用層,分了很多層面,從橫向的切成很基礎的算法、編程語言的問題,以及最上層怎麼應用的問題。數字還要分成各種各樣的專業,像無人駕駛和自然語言分析就很不一樣,可以有縱向的劃分,在這個層面上,計算機專業本身確實跟現在整個社會對計算機科學的要求有點脫節,專業的分置和佈置都在變。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"剛纔我講到爲什麼節奏不太一樣?高校的專業一設置就是四年,學生一般進來這個專業,四年不大可能變了。這四年對於產業界來講是很長的時間,課程設置既要對目前產業發展、社會發展的需求做出響應,另一方面又要維持它的長遠性,這兩個需求永遠在互相競爭。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"目前看來,我覺得要在在橫向選擇時要有不同的側重點,如側重不同的學校以及不同特色的學校。很多企業對行業比較瞭解,對於各種各樣的行業應用需求比較瞭解,企業可以幫助學校進行專業設置等。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"周愛民:"},{"type":"text","text":" 我覺得要發揮各自學校的優勢和特色。華東師範大學最大的特色是,我們是一個師範類見長的綜合性大學,所以我們在想辦法來突出師範的特色。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"去年我們在計算機科學與技術一級學科下,自主申報了智能教育二級學科的博士點,對於這樣的學生,我們在課程體系的設置上體現了交叉的特性,首先學生需要知道計算機和人工智能的技術作爲基礎,基本計算機的基礎有相應的模塊。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"還有一個很重要的模塊就是教育,因爲是交叉的,而且不是理科和工科的交叉,而是文科和工科的交叉。對博士班的培養方案,除了基本的課程學習以外,另外一塊內容是有關智能教育這方面科學技術的研究,要深入地參與到上海智能教育研究院的課程中。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"我們不僅僅想要做簡單的科學研究,還想把做科學研究的成果最終落地到中小學課程教學過程中。所以第三個模塊就是一定要實踐,把這些科學研究的理論成果參與到相關產品的開發過程中,最終從早期的理論學習到科學的研究到產業的落地,形成一個完整的鏈條。"}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"人工智能人才培養要結合區域特色"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"blockquote","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"竇德景:中國各個地方都有自己的特色,例如長三角就是一個不論在經濟還是科技上都非常有活力的地方。結合到人工智能人才的培養,一個問題是,怎麼樣和當地的產業(企業)相結合?"}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"王曉陽:"},{"type":"text","text":" 照理說,計算機科學、計算機應用,是一個普適性學科,但到了應用層面,各個區域應用的專注點不太一樣。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"今天人工智能的應用很大程度上不光是計算機專業的問題,區域的特殊性很大程度上反映了學校的特殊性。像我們復旦大學經濟、金融相當發達,在人工智能應用方面有很大的潛力,我們現在推的這些計算機工具、人工智能工具越來越輕盈、越來越好用,好用不光是爲了計算機從業者建模更容易,更是希望對應用方來說落地更容易。如何在不同的區域推動不同專業領域使用人工智能的工具,這是一個值得思考的問題。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"復旦是一直講通識教育比較多的學校,理念是通識教育爲主,專業教育在後,我們提出,通識教育裏面應該有計算機、人工智能。專業的人士,比如學經濟的應該搞一個人工智能的小專業,這樣使得他們在使用人工智能工具的時候,在區域發展的過程當中能夠發揮比較好的作用,我覺得這塊大有發展潛力。"}]},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"複雜的國際環境下,如何培養AI人才"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"blockquote","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"竇德景:當下國際環境很複雜,在這樣的背景下,如何做好AI人才培養?"}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"王曉陽:"},{"type":"text","text":" 這個問題比較複雜,我們要在應用人才的培養上做一些工作,但在當下複雜的國際環境下,我覺得最重要的一點是在基礎研究上加大投入。像百度這樣的企業,我希望能夠在基礎演講上更加給力。如何做好基礎研究,希望高校和企業增加合作,共同探索更好的途徑。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"謝少榮:在"},{"type":"text","text":"這樣的國際形勢下,我們中國人,一直在提科技的自立自強。而實現技術自立自強,人才,特別是未來的人才非常重要。希望用我們自己的算法、模型、開源平臺,來培養我們自己未來的AI人才,這非常重要。所以我很期望像百度這樣的大企業,企業有資源和實力,我們學校也有開放的心態,期望和大家一起聯合起來,培養好未來的人,人是最重要的。"}]}]}
發表評論
所有評論
還沒有人評論,想成為第一個評論的人麼? 請在上方評論欄輸入並且點擊發布.
相關文章