一位女博士在人工智能領域創業的那五年

{"type":"doc","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"作爲IT行業的新時代焦點產業,人工智能領域近年來催生了一大批很有前途的創業企業,湧現出衆多技術過硬、眼光獨到的優秀創業者。與此同時,越來越多的技術人與投資者開始將目光投向人工智能,希望在這一數字時代的“淘金熱”中搶佔先機,取得令人矚目的成就。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"不過,任何行業的創業之路都不可能一帆風順,人工智能創業者和企業都必須面對諸多挑戰與未知的困境。如何才能儘可能避開創業道路上的溝壑與彎路,怎樣充分利用有限的資源快速響應需求,在激烈的競爭中贏得優勢?想要回答這些問題,一位資深創業者的經歷無疑能爲我們帶來寶貴的經驗作參考。爲此,InfoQ大咖說欄目採訪了愛數智慧創始人和CEO張晴晴博士,以人工智能行業長達5年的資深創業者身份,爲大家分享了她在這一行業中積累的經驗與觀察思考成果。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"從學術研究到創業實踐:一位創業者的心路歷程"}]},{"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":"color","attrs":{"color":"#494949","name":"user"}}],"text":"2016年,在中科院擔任副研究員的張晴晴選擇離職創業,創辦了愛數智慧。張晴晴回憶,當時最重要的驅動力是學術領域在進行人工智能研究工作時非常缺乏數據資源的支持,於是張晴晴決定從做模型與算法的角色,轉變成爲全世界的AI工作者提供底層數據服務能力的角色。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"這樣的轉變對於象牙塔出身的張晴晴而言,無疑是巨大的挑戰。之前的十一年間,張晴晴一直身處研究環境,更多接觸的是個人性質的工作任務。但進入創業階段後,創業者需要建立商業邏輯、建立團隊並構建協作氛圍,還要更多地將目光轉向市場與行業趨勢等方面,而這些思維轉變都是不可或缺的。"}]},{"type":"heading","attrs":{"align":null,"level":5},"content":[{"type":"text","text":"創業者需要做到的兩件事"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"張晴晴從多年的研究者經歷中,收穫最大的就是建立了一套完整的科學思維體系。相比之下,創業之路更考驗創業者的學習能力與創業初心。作爲公司的引領者,CEO一定要明確企業的發展方向,也就是公司的發展戰略。創始人的思維高度往往決定了企業的發展高度,所以創業者需要持續思考、快速迭代,才能不斷提升企業成功的機率。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"創業者還要認識到,創業不是一個人的事情,而是一羣人的事情。創業者需要讓夥伴們瞭解大家需要做什麼事情,發現每一個人各自的能力與特長所在,並把自己的知識傳遞給他們,還要確保整個團隊向同一方向努力。在創業過程中,經費不足等客觀原因也曾導致很多意外事件,在張晴晴看來,團隊共同克服困難的過程也是創業必經之路。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"此外,作爲AI領域公司的創始人和CEO,張晴晴在觀察行業、決定公司方向的同時也要立足於技術的理解和認知層面,需要知道AI的底層原理是什麼、未來會怎樣發展、存在哪些阻礙以及如何去突破等等。與此同時,創始人也要以團隊服務者的角色,在大家遇到困惑的時候站出來幫助大家解決現在的問題。"}]},{"type":"heading","attrs":{"align":null,"level":5},"content":[{"type":"text","text":"給女性創業者的建議"}]},{"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":"color","attrs":{"color":"#494949","name":"user"}}],"text":"商業世界男性佔比很高。身爲女性創業者,張晴晴表示自己很少思考自己的性別、年齡等標籤。對於她來說,如何更好、更快完成創業目標,爲社會創造更大價值纔是關鍵所在。她也建議其他女性在創業或扮演重要角色時忘掉自己身上的一些標籤:"}]},{"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":"因爲很多事情,你沒有嘗試就告訴自己說不可以,那就真的沒有機會了。拋棄社會給你定義的東西,專注在事情本身,你會發現通過這樣的一種思維邏輯,這個世界都會爲你而開啓。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在你決定做這件事情之前,不用給自己設太多邊界。只要專注在這件事情上,邁開這一步,就是最大的成功。人的潛力無窮無盡,只要你真有足夠的毅力和定力去做,你會發現很多你過去想不到的目標可能有一天都能實現。"}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"愛數智慧成立五年來,張晴晴一直對於創業抱有敬畏之心,時刻思考每一個環節要怎樣去做才能更好達到目標。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"身處AI這樣高速進化的行業之中,創業者最重要的一點就是維持積極學習的態度。創業者不應該在任何時刻認爲自己應該停下腳步。所謂活到老,學到老,張晴晴認爲這不僅應該是創業者具備的素養,也應該是希望活出精彩人生的每一個人都要有的態度。"}]},{"type":"heading","attrs":{"align":null,"level":5},"content":[{"type":"text","text":"如何應對創業道路上的問題與挑戰"}]},{"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","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"談到創業道路上解決的問題,張晴晴首先提到了公司組織結構的轉變,又從業務角度分享了自己的觀察。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"在創業早期,團隊往往沒有細緻的分工,但當公司慢慢進入到一定規模後,管理者就要開始做職能拆解,明確分工職責。在公司很多年的老人可能對於這個過程會有些不適應,新來的小夥伴也需要逐漸融入這樣的環境,因此會有一定的挑戰。但職能拆分是持續的過程,隨着規模的擴張必然會繼續下去。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"對於公司業務的部分,愛數智慧的企業定位是爲所有人工智能企業提供底層數據服務。數據是未來的原油,但在不同的應用場景下,數據的提煉方式與規則也有所不同。過去五年來,愛數智慧逐步形成了一套提煉體系,搭建了一套數據處理系統,並在今年正式商業發佈。張晴晴希望把愛數智慧過去5年所沉澱下來認知凝結在這套系統裏面,賦能給更多需要用到數據的企業,幫助他們在業務演進過程中通過數據處理和迭代、各種標籤體系和預測分析等工具更快前進。"}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"人工智能:行業正在經歷怎樣的變化"}]},{"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":"color","attrs":{"color":"#494949","name":"user"}}],"text":"對話式AI是人工智能領域的終極問題之一。對話AI的目標是讓機器像人一樣和人類自然交流對話。在過去,人機對話交互主要侷限在一些特定場景,涉及確定的動作。但人類非常希望能夠按照自己想要的表達方式去跟機器交流,這樣的需求催生了對話式AI。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"以汽車智能座艙爲例,司機要告訴導航系統自己想去哪個地方,可能的表達方式是多種多樣的,系統沒有辦法完全預測每一個人的說話方式;同時人類還有口音和語氣的差異,這些在對話式AI裏面都是非常典型的難點,也是需要行業在未來幾年中逐步解決的問題。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"在解決這些問題的過程中,AI專家需要花費大量時間採集數據。因爲人工智能是靠數據驅動,數據清洗採集的好壞直接影響了建模性能,決定了預測效果。從海量數據裏挑選出想要的數據,就像在浩瀚海洋中找尋一個貝殼,是非常耗時耗力的事情。未來,隨着數據生產和現實生活中出現的數據越來越多,這個問題就會變得越來越突出。因此行業需要一套更高效的數據處理工具,也需要建立數據處理的標準和體系,才能不斷提高數據處理過程的效率。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"在數據的收集過程中,最重要的一點就是理解數據的使用場景。根據企業需要解決的問題來決定數據處理的重點和方法。數據處理過程結合了機器和人兩者的優勢,這種人機協作的方式也需要行業思考和深耕。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"目前,人機對話模式還處在比較機械、呆板的初級階段。張晴晴認爲這是比較系統性的問題,不只是單點問題。站在數據角度來講,目前人機交互的底層數據積累還遠遠不夠。比如說人們在用中文交流,但是每個人的用語體系都是有差異的,而機器在理解這種差異時會顯得呆板,不夠靈活。要解決這個問題,張晴晴認爲從業者需要將思考維度放的更長遠一些。比如認識一個人需要從很小的時候就開始在他的身邊,長年累月才能知道他的性格、喜好,人們很難在很短的時間對某人構成很清晰的認知。所以在未來,人機交互要進一步發展可能需要某種虛擬的陪伴機器人,跟隨主人一段時間後才能對主人有全面的認知。"}]},{"type":"heading","attrs":{"align":null,"level":5},"content":[{"type":"text","text":"標註員:新時代的“數據民工”?"}]},{"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":"color","attrs":{"color":"#494949","name":"user"}}],"text":"很多人認爲智能化標註平臺上的數據標註員像是“數據民工”,但張晴晴並不認同這樣的說法。在她看來,所有從業者嚴格來說都是工作者,而數據標註員是把他對於這個世界的認知轉化成機器可讀取的0、1等確定的數字。數據標註員的工作是連接人類所在的真實世界和計算機所在的虛擬世界的連接點。隨着需要處理的數據類型以及待解決問題越來越多樣化,數據標註員需要掌握的知識體系也越來越豐富。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"以智能醫療系統爲例,這個系統可能需要一些主任醫師和專家才能標註,他們標註出來的每一個結果都代表了他們積累了幾十年的經驗。其他領域裏也有類似的情況,比如金融股票方面的會議場景,就需要標註員具備相應的背景知識。如果標註員對這些領域不熟悉,那麼標出來的東西可能完全是錯的。從這個角度來講,張晴晴認爲,把數據標註員稱之爲數據專家會更合適。"}]},{"type":"heading","attrs":{"align":null,"level":5},"content":[{"type":"text","text":"AI如何幫助提升數據平臺效率"}]},{"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":"color","attrs":{"color":"#494949","name":"user"}}],"text":"數據處理是需要人和機器協同完成的事情。人力處理數據最大的問題就是不同的人對同一件事情的認知也會不同,而且行業需要處理的數據量越來越大,完全靠人工的方式來處理無疑也無法跟上AI落地的發展節奏。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"在這種情況下,行業應該將一些最基礎的標註能力逐步沉澱到系統層面,讓系統幫助人工更快完成標註工作。簡單來講,一些初步的標註工作讓機器去做,人類則負責審覈與校正,這些校正結果也會反哺系統,幫助系統更快迭代。但社會每天都會產生新生事物,系統也要處理新的事物和數據,這部分還是需要靠人來建立連接點,標註員連接起虛擬世界和真實世界,儘快將知識沉澱到系統裏面來做迭代。"}]},{"type":"heading","attrs":{"align":null,"level":5},"content":[{"type":"text","text":"AI行業的一些前沿動向"}]},{"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":"color","attrs":{"color":"#494949","name":"user"}}],"text":"在AI產業發展早期,從業者的終極目標就是構建大量的數據。不過,當前通過一些技術迭代,從業者在一些特定場景中也可以通過小數據獲得相對精準的結果。AI發展的另一個重點是可解釋性。缺乏可解釋性的情況下,從業者會不知道到底是哪些數據對模型提供了最根本的幫助,這會導致數據資源的浪費。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"聯邦學習也是AI領域的一大研究方向。通過聯邦學習,所有人各自訓練的模型就有機會共享,實現共同快速的迭代和優化。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"從產業賦能的角度來看,AI本質上是一種工具,可以賦能到各行各業。比如說在法律方面,有公司在做智慧法庭、焦點爭議的快速判定,或者一些摘要記錄工作。張晴晴表示,AI可以真正用於各行各業,只要在相應領域裏設定好標準、定義好目標是什麼,就能得到比較匹配的數據,獲得很好的賦能成果。"}]},{"type":"heading","attrs":{"align":null,"level":5},"content":[{"type":"text","text":"數據安全與隱私保護對AI行業的影響"}]},{"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":"color","attrs":{"color":"#494949","name":"user"}}],"text":"在AI的落地過程當中,數據的隱私保護是非常重要的關鍵環節。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"AI行業的初心是希望社會變得更美好,那麼在人們貢獻數據來加強與機器之間的理解、溝通能力的同時,一定要確保數據的安全性,確保數據的使用經過充分授權。從業者要嚴格區分數據的使用權與所有權,涉及到個人隱私的部分還需要做數據脫敏。隨着相關法律法規的健全完善,科研工作也在這一領域不斷取得進展。在未來,數據的合規性和技術迭代應該並駕齊驅。"}]},{"type":"heading","attrs":{"align":null,"level":5},"content":[{"type":"text","text":"AI行業未來需要克服的挑戰與發展趨勢"}]},{"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":"color","attrs":{"color":"#494949","name":"user"}}],"text":"在現有框架下,AI行業未來需要解決數據和算力這兩個人工智能的底層問題。在追求更多數據與算力的過程中,行業還需要克服能源消耗問題,保護地球環境不受破壞,實現可持續的發展道路。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"從行業宏觀角度來看,對話式AI一定會不斷前進,這是行業的終極目標。但在這個過程當中有很多問題要去解決,行業需要綜合考慮能源耗費、人力物力投入、數據合規性等等問題,最終才能一步一步邁向更美好的未來。"}]},{"type":"heading","attrs":{"align":null,"level":5},"content":[{"type":"text","text":"MagicHub開源社區背後的思考"}]},{"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":"color","attrs":{"color":"#494949","name":"user"}}],"text":"2021年4月,愛數智慧發起的MagicHub開源社區正式上線。之所以選擇以開源的方式跟大家共享愛數智慧的研究成果,張晴晴也有自己的思考和理念。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"首先,MagicHub是愛數智慧從數據集和數據處理工具這兩個維度打造的開源項目。縱觀全球,更多的開源項目集中在算法系統層面,而這些算法需要由底層的數據去驅動。但行業還沒有很多合適的數據能夠用於AI訓練。張晴晴表示,很多試圖嘗試進入到AI領域的工作者常會自己下載開源工具,希望能自己跑些AI算法,但發現在這個過程中沒有發現合適的數據來完成模型迭代,所以他們對AI的認知也會停留在比較淺的階段。數據的開源方式能夠讓更多從業者開始嘗試玩兒AI,產出更好的成果。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"目前,MagicHub開源社區成員有來自國內外的AI研究者、科研機構等,他們會主動到這個開源社區裏下載和分享、貢獻數據。MagicHub也是很好的學習社區。MagicHub提供了中英雙語的版本,方便國內研究者更快上手學習使用。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"另外,張晴晴建議從業者在學習瞭解各種技術時,可以去關注一些頂級會議,讀這些會議相關的組織者、發表的文章,通過這種以點帶面的方式,從業者就可以快速理解一個細分領域的發展歷程,更好地理解技術現狀背後的演進邏輯。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"結束語"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"如今,愛數智慧成立已經有5年時間,張晴晴自己也在AI領域走過了16年。張晴晴希望愛數智慧一直希望做一家專業的數據公司,做出更高效的數據處理系統。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]}]}
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