《知產(chǎn)人物 IP PEOPLE》對話商湯絕影王曉剛：科技領(lǐng)航，AI普惠生活

#本文僅代表作者觀點，不代表IPRdaily立場#

“《知產(chǎn)人物 IP PEOPLE》英文國際期刊就人工智能這個熱門話題與商湯科技聯(lián)合創(chuàng)始人、首席科學(xué)家、絕影智能汽車事業(yè)群總裁王曉剛進行了深入交流，探討了商湯的科研新成果與未來發(fā)展方向，近距離感受到了國內(nèi)人工智能的成長?！?/strong>

來源：IPRdaily中文網(wǎng)（iprdaily.cn）

《知產(chǎn)人物 IP PEOPLE》英文國際期刊

A Talk with Wang Xiaogang at SenseTime: Leverage leading AI for good

隨著AI時代的到來，人類文明再次迎來新一輪革新，人工智能也逐漸成為了資本競逐的賽道。先進、便捷、智能、高效、科技——這是一個充滿現(xiàn)代氣息與未來憧憬的行業(yè)，我們可以看到，AI正在改變著人類的生活方式。

The advent of AI starts a new chapter of human civilization. Gradually AI has become a new arena where businesses compete for dominance. Characterized by sophistication, convenience, intelligence, efficiency, and technology, the modern and promising AI industry is changing people’s lives.

2016年，阿爾法圍棋以4：1的總比分戰(zhàn)勝世界圍棋冠軍、職業(yè)九段棋手李世石后，人工智能這項產(chǎn)業(yè)如雨后春筍般掘地而起。在千巖競秀的新賽道里，商湯科技憑借著頂尖的科學(xué)家團隊和強大的科技實力突出重圍，在不到十年的時間里，成為了亞洲智能軟件企業(yè)中的翹楚。

After a remarkable 4-1 victory in 2016 when AlphaGo, a computer Go program, defeated the 9-dan professional and world champion Lee Sedol, AI startups have sprung up around the world. Backed by its team of top scientists and strong technological strength, SenseTime has survived the cut-throat competition in the AI industry and turned into a leading intelligent software vendor in Asia within less than a decade.

AI機器視覺作為人工智能領(lǐng)域的核心技術(shù)之一，為我國實施智能制造戰(zhàn)略提供了重要的技術(shù)支撐。商湯創(chuàng)始團隊專注于人工智能計算機視覺領(lǐng)域已超過20年，持續(xù)的投入讓商湯在這領(lǐng)域積累了眾多原創(chuàng)技術(shù)能力，并在國內(nèi)人工智能加速商業(yè)落地的洪潮中抓住了資本市場機遇。2021年12月30日，商湯成功在港交所上市，成為中國AI的第一股，硬科技背后的價值持續(xù)得到資本方的認可與期待。

One of the core technologies of AI, machine vision provides tremendous technological support for the implementation of China’s smart manufacturing strategy. SenseTime’s founding team has specialized in computer vision for over 20 years. Their continuing input in this field has translated into an extensive portfolio of original technologies and enabled the company to grab the opportunities for fundraising amid the wave of AI commercialization in China. On December 30, 2021, SenseTime was successfully listed on Hong Kong Stock Exchange and became the first AI company to go public in China. This demonstrates investors’ confidence in and expectations of the value of hard & core technology.

智慧商業(yè)、智慧城市、智慧生活、智能汽車是商湯四大核心業(yè)務(wù)板塊，從商湯發(fā)布的2021年年報來看，四大板塊過去一年的營收均實現(xiàn)了數(shù)字增長。當(dāng)今，智能駕駛的時代已經(jīng)到來，智能汽車作為商湯四大核心業(yè)務(wù)板塊之一，無人駕駛也一直是長期研究的重點。2021年，商湯推出了SenseAuto絕影智能汽車平臺，在智能駕駛、智能座艙、車路協(xié)同、L4級無人駕駛、無人駕駛小巴推進全線產(chǎn)品化商用。商湯接下來也將利用絕影平臺使城市交通管理、物流運輸、出行服務(wù)等變得便捷，將AI更好地賦能于行業(yè)和人們智能出行的需求。

According to SenseTime’s 2021 annual report, its four business domains, namely Smart Business, Smart City, Smart Life, and Smart Auto, all achieved revenue growth over the past year. Autonomous driving technology has been a long-term focus of SenseTime’s research effort to reach the technology frontier in the driverless era. In 2021, SenseTime launched SenseAuto—an intelligent automotive application platform—to commercialize its full line of products, including smart pilot system, smart cabin system, vehicle-to-everything (V2X) system, L4 autonomous driving system, and autonomous minibus system. SenseTime will leverage the SenseAuto platform to harness the power of AI to meet the demands of people and industries by providing convenient solutions for urban traffic management, logistics and transport, and traveling.

近期，《知產(chǎn)人物 IP PEOPLE》英文國際期刊就人工智能這個熱門話題與商湯科技聯(lián)合創(chuàng)始人、首席科學(xué)家、絕影智能汽車事業(yè)群總裁王曉剛進行了深入交流，探討了商湯的科研新成果與未來發(fā)展方向，近距離感受到了國內(nèi)人工智能的成長。

Recently, IP PEOPLE had an in-depth conversation on AI with Wang Xiaogang, Co-founder and Chief Scientist at SenseTime, and President of SenseAuto Business Unit. During the interview, Dr. Wang shared the latest R&D outcomes and future direction of SenseTime, and the AI development in China.

《IP PEOPLE》 Issue 1

Editor：Wei Lyu

訪談?wù)?Interview excerpts

《IP PEOPLE》：商湯科技目前已經(jīng)是行業(yè)領(lǐng)先的人工智能軟件企業(yè)，在智慧商業(yè)、智慧城市、智慧生活和智能汽車四大核心業(yè)務(wù)板塊均已處于業(yè)內(nèi)領(lǐng)先地位，您認為商湯得以快速成長保持高質(zhì)量發(fā)展的主要原因是什么？

IP PEOPLE: SenseTime has become an industry-leading AI software provider, and the four core business domains—Smart Business, Smart City, Smart Life, and Smart Auto—are all among the best in the industry. What is the main reason behind SenseTime’s rapid and high-quality development?

王曉剛：得益于中國大力促進科技創(chuàng)新以及數(shù)字基礎(chǔ)設(shè)施的建設(shè)，AI 技術(shù)與實體經(jīng)濟正不斷加速融合，給了包括商湯在內(nèi)的人工智能企業(yè)廣闊的應(yīng)用空間、場景。商湯在創(chuàng)立之初完成了最重要的原始人才積累，如今亦完善了人才造血機制，創(chuàng)始團隊很早就持續(xù)專注于人工智能計算機視覺領(lǐng)域，如今已超過20年。過去4年，商湯累計研發(fā)投入超80億人民幣。持續(xù)的投入，讓商湯在這一領(lǐng)域積累了眾多原創(chuàng)技術(shù)能力，并能夠更好的預(yù)判并抓住市場機遇。

Wang Xiaogang: Thanks to China’s big efforts to stimulate technological innovation and develop digital infrastructure, AI technology is integrating with the real economy at an increasingly fast pace. This has provided AI companies, including us, with enormous opportunities and scenarios for technology application. We formed teams essential to our business when we founded SenseTime and have gradually improved the talent cultivation system. Our founding team has been focusing on computer vision for more than 20 years so far. In the past four years, we invested RMB 8 billion in R&D. The ongoing input allows us to accumulate original technological capabilities, and better forecast and grab the market opportunity.

《IP PEOPLE》：商湯一直以“堅持原創(chuàng)，讓AI引領(lǐng)人類進步”為使命，其中的“原創(chuàng)”應(yīng)該怎樣理解呢？

IP PEOPLE: “To create a better AI-empowered future through innovation” has been SenseTime’s mission. What does “innovation” mean to SenseTime?

王曉剛：創(chuàng)業(yè)之初，我們就下定決心要做真正意義上的原創(chuàng)，推動行業(yè)邊界。人工智能在諸多行業(yè)突破“工業(yè)紅線”，技術(shù)演進速度非?？?。當(dāng)你創(chuàng)新的速度比別人“模仿”的速度還要快時，就能形成足夠的優(yōu)勢，這個優(yōu)勢在于對人才、運算工具的聚集和場景的覆蓋。做到這點，商湯就可以享受由技術(shù)突破帶來的時間窗口進而布局商業(yè)。原創(chuàng)帶來增量價值，商湯今天的成果，也證實了堅持做原創(chuàng)技術(shù)的正確性。

Wang Xiaogang: We decided to pursue innovation to push the boundaries of the industry in the early days of our company. AI has scaled up efficiency in many industries and is evolving at a very fast speed. We would gain a competitive edge over others if we innovate faster than they copy us, and innovation is achieved with talent, tools, and coverage of user scenarios. If we can accomplish that, we would be able to utilize the time that our rivals need to catch up with us to plan for our business. Innovation brings incremental value. Our achievement has validated our choice of technology innovation.

《IP PEOPLE》：商湯的研發(fā)投入逐年增加，研發(fā)成果的取得十分不易，在保護科技創(chuàng)新成果方面，商湯采取了哪些措施？

IP PEOPLE: SenseTime’s investment in R&D has increased year by year. What has SenseTime done to protect the hard-won technological innovations?

王曉剛：作為一家人工智能高科技企業(yè)，公司利用專利、軟著、商標(biāo)商號、技術(shù)秘密等不同知識產(chǎn)權(quán)類型來合理保護科技創(chuàng)新成果。持續(xù)利用知識產(chǎn)權(quán)來維護市場秩序，打擊惡意侵權(quán)行為。隨著逐年增加對研發(fā)和知識產(chǎn)權(quán)的投入，商湯集團的專利年申請量和授權(quán)量連續(xù)5年以100%以上比例增長，在海外市場競爭中有力的保護了公司創(chuàng)新成果。

Wang Xiaogang: As an AI company, we protect our innovations through different intellectual property (IP) rights, such as patent, software copyright, trademark, trade name, and know-how. We believe in the power of IPR to maintain market order and crack down on malicious infringement. With increasing investment in R&D and IP year by year, the number of our annual patent applications and that of patent licenses we grant have increased by more than 100% for five consecutive years, which offered effective protection of our innovations in overseas markets.

為了創(chuàng)造激勵創(chuàng)新保護創(chuàng)新的氛圍，公司每年投入上千萬元獎勵創(chuàng)新成果的發(fā)明人或設(shè)計人，同時設(shè)立了線索舉報機制來收集市場侵權(quán)證據(jù)，從而確保公司知識產(chǎn)權(quán)保護的有效性。

To create an atmosphere that encourages and protects innovation, we give rewards worth over RMB 10 million each year to inventors or designers of innovations. In addition, we have established a reporting system to collect infringement evidence and ensure the effectiveness of our IP protection.

《IP PEOPLE》：截至2021年底，商湯的全球?qū)＠Y產(chǎn)累計11,494件，較2020年底增長了96%，請問是怎樣的契機致使專利得到了如此大幅的增長呢？

IP PEOPLE: As of the end of 2021, SenseTime had 11,494 global patents, a 96% increase compared to 2020. What triggered such a significant increase in patents?

王曉剛：商湯集團全球?qū)＠Y產(chǎn)的迅猛增長源自公司對科技研發(fā)和知識產(chǎn)權(quán)的高度重視。2021年，公司進一步增加了對研發(fā)的投入，擴大了研發(fā)團隊規(guī)模，隨之產(chǎn)生了更多創(chuàng)新成果和知識產(chǎn)權(quán)保護。另外，隨著知識產(chǎn)權(quán)保護體系和創(chuàng)新激勵機制的進一步成熟，公司知識產(chǎn)權(quán)團隊對創(chuàng)新成果進行了更深入有效的發(fā)掘和布局，創(chuàng)新效率得到了有效提升。2021年商湯集團海外市場拓展取得了明顯突破，在海外專利布局方面明顯加大了投入，這也是專利數(shù)量繼續(xù)大幅增長的重要原因之一。

Wang Xiaogang: It is a result of SenseTime’s tremendous effort in technological R&D and IP protection. In 2021, we ramped up investments in R&D and expanded our R&D team, yielding more innovations and intellectual properties. With further improvement of the IP protection system and innovation incentive mechanism, our IP team dug deeper into our innovations and made effective plans for their development. As a result, the innovation efficiency has significantly improved. In 2021, we made great breakthroughs in overseas market expansion and substantially increased investment in patent portfolio planning overseas, which also contributed to the surge in the number of patents.

但是，公司從沒有將專利資產(chǎn)數(shù)字作為目標(biāo)，在數(shù)量增加的同時公司更加重視知識產(chǎn)權(quán)的質(zhì)量和有效保護，在創(chuàng)新挖掘過程中進行嚴(yán)格評估和篩選，增加創(chuàng)新成果的創(chuàng)新性和侵權(quán)易檢測性，同時加強對授權(quán)專利的梳理和價值評估，打造高質(zhì)量的商湯專利資產(chǎn)庫。

The number looks good, but it is never our goal. In fact, we attach more importance to the quality and effective protection of IP than quantity. We are committed to building a high-quality patent portfolio by rigorously evaluating and screening the innovations to make them more patentable and make it easier to detect infringements, and by strengthening the organization and value assessment of granted patents.

《IP PEOPLE》：發(fā)明專利占比高達78%，這在商湯的商業(yè)發(fā)展中有起到怎樣的作用嗎？

IP PEOPLE: With 78% of the patents being inventions, what role has this played in SenseTime’s business development?

王曉剛：商湯很大比例的產(chǎn)品和服務(wù)體現(xiàn)為軟件和算法，更加適合用發(fā)明專利來加以保護，這是公司專利資產(chǎn)中發(fā)明專利占比較高的主要原因。事實上，幾年前公司專利資產(chǎn)中發(fā)明專利占比更高，一度接近90%。

Wang Xiaogang: Primarily, a high proportion of SenseTime’s products and services are software and algorithms, which render them to be patented as inventions. A few years ago, the percentage of invention patents was even higher, hitting nearly 90%.

我們并不會刻意追求發(fā)明專利高占比，而是以打造有效的知識產(chǎn)權(quán)武器為目標(biāo)合理地布局公司的專利資產(chǎn)組合。隨著公司產(chǎn)品的多樣化以及對知識產(chǎn)權(quán)保護實踐的不斷探索，必要時公司也會充分布局實用新型專利和外觀設(shè)計專利，來對公司各種形態(tài)的產(chǎn)品加以全面保護，發(fā)揮它們?nèi)菀讬z測侵權(quán)的優(yōu)勢來保護公司創(chuàng)新成果，這說明了公司產(chǎn)品序和知識產(chǎn)權(quán)保護理念的成熟。

However, this is not what we are after. We aim at using IP as an effective business weapon through reasonable planning for the company’s patent portfolio. Continuing to expand our product range and explore IP protection practices, we will make strategic planning for the utility model and design patent portfolios when necessary, in order to ensure adequate protection of various types of products. It is easy to identify infringing utility model and design patents, and we will make full use of this to defend our IP. This demonstrates the maturity of our product series and IP protection policy.

《IP PEOPLE》：您一直領(lǐng)導(dǎo)人工智能產(chǎn)業(yè)方面的落地，在創(chuàng)新成果轉(zhuǎn)變?yōu)閷嶋H應(yīng)用的過程中，商湯有遇到哪些問題？是否有值得參考借鑒的經(jīng)驗分享呢？

IP PEOPLE: You have been leading SenseTime’s AI development. What challenges have you encountered during the transformation of innovations into practical applications? Could you share your experience?

王曉剛：人工智能從原創(chuàng)研究到產(chǎn)業(yè)的成功落地，本身就非常依賴于產(chǎn)學(xué)研的相互作用與密切配合。商湯對于“AI的長期發(fā)展離不開前沿技術(shù)的持續(xù)研究”有著深入的認識。我們與 50 多所高校建立了合作，同時建立了15個校企聯(lián)合實驗室，來打造“產(chǎn)學(xué)研”協(xié)同模式，將實驗室的學(xué)術(shù)研究課題與產(chǎn)業(yè)應(yīng)用中的實際問題相結(jié)合，不僅打通了技術(shù)創(chuàng)新到產(chǎn)業(yè)應(yīng)用之間的橋梁，更實現(xiàn)人才從學(xué)術(shù)研究到應(yīng)用研究的一體化培養(yǎng)。

Wang Xiaogang: The development of AI, from original research to product creation, depends heavily on industry-university-research cooperation. Highly aware of the fact that “the sustainable development of AI would be impossible without the continuous research on cutting-edge technology”, SenseTime has formed partnerships with more than 50 universities and established 15 joint laboratories for industry-university-research collaboration. By associating the research topics in laboratories with practical problems in industrial applications, we have built a bridge between technological innovation and industrial application, and enabled talent training that combines academic research and applied research.

在商湯，我們基于原創(chuàng)技術(shù)和市場洞察，探索出了獨具特色的商湯模式，即“1（基礎(chǔ)研究）+1（產(chǎn)品及解決方案）+X（行業(yè)）”模式。具體而言，我們不斷把突破“工業(yè)紅線”的學(xué)術(shù)成果落地在有需求的行業(yè)中，而不斷升級的行業(yè)需求又會促使研究者進一步探究，并最終形成一個良性循環(huán)，為行業(yè)智能化轉(zhuǎn)型注入新動能。

Based on our original technologies and insight into the market, we have created a unique model, namely “1 (basic research) + 1 (products and solutions) + X (industry)”. Specifically, we keep bringing research breakthroughs to industries in need, and industries’ growing demands for more advanced technology will prompt researchers to keep expanding the frontier of knowledge. As a result, a virtuous circle is formed to inject vitality into the intelligent transformation of industries.

《IP PEOPLE》：商湯科技在2021年登榜全球智能駕駛專利Top10，也推出了商湯絕影智能汽車平臺，那么在智能汽車領(lǐng)域，您認為商湯的主要競爭優(yōu)勢在哪里？

IP PEOPLE: In 2021, SenseTime was ranked among the top 10 global intelligent driving patent holders and launched SenseAuto, an intelligent automotive application platform. What is the main competitive advantage of SenseTime?

王曉剛：智能汽車正在從傳統(tǒng)的“制造”到“制造+服務(wù)”模式轉(zhuǎn)變。汽車要變得更加“聰明”，需要產(chǎn)業(yè)鏈各方緊密協(xié)同，找到一條讓技術(shù)以低成本且高效的方式去豐富汽車智能應(yīng)用場景的通道。背靠商湯深厚的學(xué)術(shù)積累與行業(yè)領(lǐng)先的全棧式人工智能能力，絕影智能汽車平臺以SenseCore商湯AI大裝置為基石，打通算法、算力和平臺，大幅降低人工智能生產(chǎn)要素價格，實現(xiàn)高效率、低成本、規(guī)模化的AI創(chuàng)新和落地。

Wang Xiaogang: The smart auto industry is undergoing a transformation from “manufacturing” to the “manufacturing + service” model. To make automobiles more intelligent, all parties in the industry chain need to work closely to figure out how to use technologies to enable intelligent application scenarios for automobiles in a cost-effective manner. Backed by its research experience and industry-leading full-stack AI capabilities, SenseAuto leverages the SenseCore AI infrastructure to integrate algorithms, computing power, and platforms, drastically reducing the cost of factors of production in AI and enabling efficient, low-cost AI innovation and implementation at scale.

目前，商湯絕影已經(jīng)實現(xiàn)AI生產(chǎn)效率行業(yè)第一，并形成了量產(chǎn)智能駕駛、智能車艙、自動駕駛功能車、車路協(xié)同、賦能引擎的全套產(chǎn)品矩陣，具備了相對完備的測試和數(shù)據(jù)積累體系。此外，絕影還與商湯豐富的行業(yè)板塊實現(xiàn)相互協(xié)同。以智慧生活板塊與絕影的聯(lián)動來舉例，2021年世界人工智能大會，絕影也首次結(jié)合相關(guān)虛實融合技術(shù)，在2021世界人工智能大會上推出了世界上第一款自動駕駛AR小巴。

SenseAuto now leads the industry in AI productivity, with a full range of products developed, namely SenseAuto Pilot, SenseAuto Cabin, SenseAuto RoboX, SenseAuto Connect and SenseAuto Empower. A sound testing and data accumulation system has taken shape. Furthermore, SenseAuto can be integrated with other business domains of SenseTime. For example, SenseAuto debuted the world’s first AR-powered self-driving minibus at the 2021 World Artificial Intelligence Conference.

此外，通過與諸多國內(nèi)外頂尖車企客戶的磨合，商湯逐漸熟悉了主機廠的整套技術(shù)流程，質(zhì)量標(biāo)準(zhǔn)體系，為技術(shù)深入落地打下堅實基礎(chǔ)。截至2021年，商湯絕影已簽約合作的車企客戶超30家，合作的前裝定點數(shù)量也達到2300萬臺。

Through cooperation with many leading auto companies at home and abroad, SenseTime has gradually mastered the entire technical process of auto makers and quality standard system, laying a solid foundation for further technology implementation. By 2021, we signed contracts with more than 30 automobile companies, and we have been selected as the supplier to pre-install our SenseAuto products for more than 23 million automobiles.

在智能汽車領(lǐng)域，健全的質(zhì)量體系意味著行車安全、交互流暢。絕影自2016年開始探索智能汽車領(lǐng)域以來，不斷執(zhí)行嚴(yán)苛的質(zhì)量管理，2016年，商湯率先通過ISO9001認證；2019、2020年先后通過ASPICE CL1和CL2等級認證，是亞洲最早拿到ASPICE認證的AI平臺公司。

In the smart auto industry, a robust quality system can ensure safe driving and enable seamless interactions. SenseAuto has maintained strict quality management since it ventured into the smart auto industry in 2016. SenseTime got the ISO 9001 certification in 2016, and was certified with ASPICE CL1 and CL2 in 2019 and 2020 respectively, making it the first AI company in Asia to obtain the ASPICE certification.

今年一季度，商湯完成組織架構(gòu)升級，成立智能汽車事業(yè)群，更大力度發(fā)展智能汽車業(yè)務(wù)，致力于將絕影打造為汽車行業(yè)內(nèi)最具影響力的AI賦能平臺，引領(lǐng)汽車行業(yè)的智能化變革。

In the first quarter of 2022, we upgraded our organizational structure and established the SenseAuto Business Unit. We will step up our efforts to develop the smart auto business and stay committed to developing SenseAuto into the most influential AI-powered platform and leading the intelligent transformation of the automobile industry.

《IP PEOPLE》：在AI領(lǐng)域，人才是各國各大企業(yè)都持續(xù)爭奪的重要資源。據(jù)了解，商湯科技的研發(fā)人才隊伍占員工總數(shù)的70%，在科技企業(yè)中已經(jīng)是相當(dāng)高的占比數(shù)，那么對于研發(fā)人才，商湯科技是如何建設(shè)人才體系，激發(fā)創(chuàng)新活力的呢？

IP PEOPLE: Talent has always been an important resource that AI companies around the world scramble for. We have learned that SenseTime’s R&D team accounts for 70% of the total number of employees, which is a rather high percentage among tech companies. So, how does SenseTime develop R&D teams to stimulate innovation?

王曉剛：商湯從成立之日起，就堅持以人才發(fā)展為根本，積極投入大量資源，搭建符合員工職業(yè)發(fā)展路徑及全生命周期的學(xué)習(xí)體系。為夯實員工的職業(yè)發(fā)展體系，職業(yè)發(fā)展中心為員工提供不同階段、不同崗位的專業(yè)力培訓(xùn)。針對校招新員工，安排了AI領(lǐng)航員和AI先鋒等學(xué)習(xí)項目，集中賦能與在崗培訓(xùn)實踐聯(lián)動，助力員工提升崗位勝任力；面向社招員工，商湯建設(shè)了豐富的線上學(xué)習(xí)內(nèi)容以及線下集中培訓(xùn)。

Wang Xiaogang: Talent development has been a top priority of SenseTime since day one. We invest a lot of resources to build a training system that fits the career development paths and the life cycle of employees. To support employees’ career advancement, our career development center provides professional training programs for employees at different stages and in different positions. For talent hired through campus recruitment, we offer various training programs such as AI Pilot and AI Pioneer, and integrate general training with on-the-job training to help employees improve their job competency. For new hires recruited through other means, we provide various online learning programs and offline centralized training.

公司組織頂尖專家團隊，開發(fā)各專業(yè)方向微課，同時，輔助公司的通用力培訓(xùn)體系，匹配內(nèi)外專家講授的軟技能課程，全面提升員工的綜合能力。面向不同發(fā)展階段的管理者，基于商湯自身的領(lǐng)導(dǎo)力模型，結(jié)合豐富的內(nèi)外部資源，我們設(shè)計了面向不同層級管理者的領(lǐng)導(dǎo)力發(fā)展體系。同時建設(shè)線上學(xué)習(xí)平臺，通過持續(xù)不斷的賦能員工，商湯著力在新時代下打造學(xué)習(xí)型組織，促進員工和公司的共同發(fā)展。

We organize a team of top experts to develop mini-courses in various disciplines, which, along with the general training and soft skills courses taught by experts, can comprehensively enhance the capabilities of employees. Besides, we have designed a leadership development system for managers at different levels based on SenseTime’s leadership models and abundant internal and external resources. We have also developed an online learning platform. By continuously empowering employees, we strive to build a learning organization in the new era and promote the development of employees and the company.

《IP PEOPLE》：人工智能是目前全球科技界最炙手可熱的名詞之一，眾多全球科技巨頭都在AI領(lǐng)域投入了巨大的研發(fā)力度。在您看來，目前中國的AI技術(shù)發(fā)展在全球來看處于怎樣的水平？

IP PEOPLE: AI is currently one of the top buzzwords in the tech world. Many global tech giants have made tremendous investments in the R&D of AI. How would you characterize China’s AI development when it is benchmarked against the highest level in the world?

王曉剛：從整個人工智能發(fā)展歷史來看，雖然中國屬于后來者，但在諸多政策的積極推動下，國內(nèi)人工智能呈現(xiàn)出基礎(chǔ)技術(shù)研究發(fā)展迅速、人工智能研究人群快速擴大等特點。以商湯為例，從量化產(chǎn)出來看，截至 2021 年底，商湯累計擁有 11494 件全球?qū)＠Y產(chǎn)，相較于 2020 年底增長 96%，其中 78% 為發(fā)明專利。同時，作為創(chuàng)新副產(chǎn)品，商湯團隊發(fā)布了600多篇計算機視覺深度學(xué)習(xí)的頂尖論文，全球排名第一。

Wang Xiaogang: Although a latecomer in AI development, China has introduced a series of policies to support the development of AI. The AI industry in China is characterized by rapid growth in basic technology research and fast expansion of the group of AI researchers. Take SenseTime for example. In terms of quantitative output, we had 11,494 global patents as of the end of 2021, a 96% increase compared to 2020, of which 78% were invention patents. SenseTime has published more than 600 top papers on deep learning for computer vision (the “by-products of innovation”), ranking first among companies around the world.

在AI技術(shù)落地方面，國內(nèi)海量的市場規(guī)模為“AI+”，帶來了得天獨厚的環(huán)境，提供了豐厚的應(yīng)用土壤，這是國內(nèi)的又一大優(yōu)勢。與此同時我們也應(yīng)該看到，我們在體系化、基礎(chǔ)理論及源頭創(chuàng)新上仍有不足，國產(chǎn)化硬件生態(tài)及人工智能技術(shù)工具鏈與國外仍有差距。

In terms of AI technology implementation, the huge domestic market provides favorable conditions for “AI+”—the integration of AI and industries—and generates considerable needs for AI applications. Nonetheless, we should realize that we still fall short of systematization, basic theory, and original innovation, and that we still fall behind our peers overseas in the hardware ecosystem and AI technology toolchain.

《IP PEOPLE》：我們未來還需要繼續(xù)做出哪些努力？

IP PEOPLE: What efforts should we continue to make in the future?

王曉剛：要秉持長線思維，重視核心人才培養(yǎng)，強調(diào)原創(chuàng)學(xué)術(shù)研究；頂級人才的聚集和后續(xù)專業(yè)人才的儲備，對中國原創(chuàng)AI產(chǎn)業(yè)的發(fā)展有著舉足輕重的作用。同時探索更多場景，已落地應(yīng)用倒逼原創(chuàng)技術(shù)前進：中國市場對新技術(shù)、新應(yīng)用的接受速度快，能夠快速衍生出豐富的場景需求。往往應(yīng)用的推進，也會對這項技術(shù)提出更多超前的、突破性的問題。

Wang Xiaogang: We should uphold long-term thinking, take seriously the development of key talent, and underline the originality of research. Bringing together top talent and maintaining a talent pool is essential to the development of China’s AI innovation. We should also explore application scenarios and use existing applications to drive the development of original technology. The Chinese market can quickly adopt new technologies and applications, and spawn new demands for scenarios within a short time. As we facilitate the application of a technology, we would generate more demands for further advances in the technology.

《IP PEOPLE》：多年前您曾說過，AI會給人類帶來新機會，發(fā)展至今，您覺得未來AI技術(shù)還將會如何影響人類？

IP PEOPLE: Years ago, you said that AI would bring new opportunities to humanity. And as the industry evolves, how will AI technology impact humanity in the future?

王曉剛：經(jīng)過學(xué)術(shù)界、產(chǎn)業(yè)界、政府側(cè)等相關(guān)方過去數(shù)年的共同努力，人工智能正快速深入企業(yè)的生產(chǎn)制造環(huán)節(jié)及普羅大眾的生活，帶來了諸多智能化變革。借助AI大裝置和大數(shù)據(jù)技術(shù)突破，AI加速了我們對于真理的探索。正是人工智能所帶來的創(chuàng)新范式的轉(zhuǎn)換，讓人類第一次擺脫了對自我認知的依賴。

Wang Xiaogang: Through the joint efforts of academia, enterprises, government, and other relevant parties in the past few years, AI is rapidly penetrating into manufacturing and people’s everyday life, driving numerous intelligent transformations. Based on AI infrastructure and breakthroughs in big data technology, AI has accelerated our pursuit of truth. It is the AI-enabled change in the innovation paradigm that has freed humans from their dependence on self-awareness for the first time.

同時我們也要注意到，科技創(chuàng)新來源于對過往認知的顛覆，因此要給科技發(fā)展以規(guī)矩，用倫理的框架守住技術(shù)底線、設(shè)置發(fā)展的方向，激勵正向的技術(shù)創(chuàng)新。

We should also notice that technological innovation often comes from the demolition of existing knowledge. It is, therefore, necessary to set rules for technological development, establish an ethical framework for technology, and guide the direction of development to stimulate technological innovations that bring positive changes.

人工智能要帶來普適、普惠、公平，首先要將當(dāng)今時代的生產(chǎn)要素成本大幅降低，使得原本我們獲取不到的教育資源、醫(yī)療資源等，更好、更廉價、更公平地開放給全社會。同時，要有意識地去關(guān)懷弱勢群體。只有這樣，技術(shù)才能更好地造福人類，實現(xiàn)AI向善，普惠眾人。

If we want AI to bring universal benefits, inclusiveness, and fairness, we must substantially reduce the cost of the factors of production, so that resources such as education and healthcare that were previously not accessible can be made available to the whole society in a better, more affordable, and fairer manner. At the same time, we must care for the disadvantaged. Only in this way can we use technology to better benefit all human beings and leverage AI for the greater social good.

"Record those who has made remarkable contributions to Intellectual Property"

（原標(biāo)題：對話商湯絕影王曉剛：科技領(lǐng)航，AI普惠生活）

來源：IPRdaily中文網(wǎng)（iprdaily.cn）

《知產(chǎn)人物 IP PEOPLE》英文國際期刊

編輯：IPRdaily王穎校對：IPRdaily縱橫君

注：原文鏈接：《知產(chǎn)人物 IP PEOPLE》對話商湯絕影王曉剛：科技領(lǐng)航，AI普惠生活（點擊標(biāo)題查看原文）

「關(guān)于IPRdaily」

IPRdaily是全球領(lǐng)先的知識產(chǎn)權(quán)綜合信息服務(wù)提供商，致力于連接全球知識產(chǎn)權(quán)與科技創(chuàng)新人才。匯聚了來自于中國、美國、歐洲、俄羅斯、以色列、澳大利亞、新加坡、日本、韓國等15個國家和地區(qū)的高科技公司及成長型科技企業(yè)的管理者及科技研發(fā)或知識產(chǎn)權(quán)負責(zé)人，還有來自政府、律師及代理事務(wù)所、研發(fā)或服務(wù)機構(gòu)的全球近100萬用戶（國內(nèi)70余萬+海外近30萬），2019年全年全網(wǎng)頁面瀏覽量已經(jīng)突破過億次傳播。

（英文官網(wǎng)：iprdaily.com 中文官網(wǎng)：iprdaily.cn）

本文來自IPRdaily中文網(wǎng)（iprdaily.cn）并經(jīng)IPRdaily.cn中文網(wǎng)編輯。轉(zhuǎn)載此文章須經(jīng)權(quán)利人同意，并附上出處與作者信息。文章不代表IPRdaily.cn立場，如若轉(zhuǎn)載，請注明出處：“http://m.jupyterflow.com

《知產(chǎn)人物 IP PEOPLE》對話商湯絕影王曉剛：科技領(lǐng)航，AI普惠生活

《知產(chǎn)人物 IP PEOPLE》對話商湯絕影王曉剛：科技領(lǐng)航，AI普惠生活