Launched in 2017, the Synced Machine Intelligence Awards are the largest and most authoritative annual awards platform in China. The selections are also of great interest across the international machine learning community. The Synced Machine Intelligence Awards 2019 focus is on “the power of industry,” and the selection process keyed in on AI products, application cases and industry landings, with choices based on verifiable industry performance.
Synced Machine Intelligence Awards 2019:
- 30 Most Promising AI Startups
- Top 10 AI Rising Stars in China
- Top 10 Industry AI Research Labs in China
- 30 Industry Leading AI Transformation Business Cases
- 30 Most Innovative AI Solutions
- 10 Most Noteworthy Funding Activities
Our selection committee collected hundreds of data points on over a thousand companies and countless products, solutions and application cases as part of a thorough candidate evaluation process for the Synced Machine Intelligence Awards 2019.
30 Most Promising AI Startups
In May 2019, 4Paradigm set the ICDAR world record, the global OCR authoritative standard; and in June it set a new record on the object recognition standard PASCAL VOC. In June 2019, 4Paradigm released the first software-hardware integrated AI integration system, SageOne, using a software-defined dedicated AI system architecture to provide software and hardware full-stack capabilities for enterprise transformation. In 2019, 4Paradigm achieved fruitful results in the fields of retail, energy, media, etc. Its customers include YumChina, Yonghui Superstores, China National Petroleum Corporation, SPT Energy Group Inc., People’s Daily, China Electronics Technology Group Corporation and so on.
In the retail industry Aibee plays a leading role in the offline digital shopping mall market. In 2019 it cooperated with 11 Top 30 commercial real estate brands, such as Wanda, K11, Red Star Macalline, and more than 80 large shopping malls. Aibee has landed a new generation of smart parking solutions in large shopping malls located in Beijing, Shanghai, Guangzhou, Chongqing, etc. The company has helped 8 industries and 16 offline retail brands enter the era of digital operation. In terms of tourism, Aibee is working closely with Send Intelligent to promote the intelligent upgrade of nearly 200 well-known scenic spots in China. Aibee has reached in-depth cooperation with Capital Airport, Daxing Airport, Qingdao Jiaodong Airport, and Qingdao Liuting Airport. In Daxing for example nearly 200 stores in the airport have adopted offline retail digital solutions provided by Aibee. The company completed its A1 round with US$74 million in financing, and its accumulated financing now exceeds US$170 million.
In 2019, AInnovation dived into AI + manufacturing, retail, and financial industries with a focus on visual and decision intelligence. The company has introduced industrial solutions such as the ManuVision industrial vision platform, ABC industrial quality assurance all-in-one machine, and a retail full value chain. It has carried out commercial cooperation with more than 300 industry benchmark clients, and its revenue has increased by more than 5 times compared to 2018. AInnovation’s computer vision application market share ranks among the top 6 in China (data source: IDC). The company has won championships in four major visual contests: Wider Face, Pascal VOC, Cityscapes, and MOT Challenge. It has also published papers in international conferences and journals such as ICCV, DAC, and IEEE, and has submitted nearly 300 patents in total. In 2019, AInnovation completed two rounds of financing, A / A + and B, with a total financing amount of nearly US$150 million. The company employs nearly 400 people and has branchesin 9 cities across China. It has two industry subsidiaries, CCIDQ and RuiYunQi.
In January 2019, AISPEECH and SMIC’s semiconductor investment company China Fortune-Tech Capital jointly invested in and established “ShensiliCon”, at the same time launching the first generation of milliwatt-level AI speech dedicated chip, the TH1520. In the same year, AISPEECH’s speech synthesis technology, in-vehicle speech solutions, and intelligent customer service solutions all passed the national standard certification, and the company participated in the formulation of nearly 20 technologies or products related national and group standards. In terms of key technologies, AISPEECH launched a small data rapid customization of TTS service, a full-duplex algorithm system, a multilingual hybrid recognition engine, a long speech recognition service, and an extremely low power consumption algorithm solution. On the market side, AISPEECH expanded its business areas to new industries such as hospitality, real estate, logistics, finance, and government affairs, and has multi customer cases implemented. As of the end of December 2019, the AISPEECH artificial intelligence industry fund “AISTAR VENTURES” had invested in more than 20 upstream and downstream enterprises in the field of artificial intelligence and the Internet of Things, including healthcare, education, and corporate services.
Founded in 2003, Automation Anywhere is an American enterprise-level robotic process automation (RPA) startup. Automation Anywhere announced in November 2019 that it had completed a US$290 million Series B financing, with a post-investment valuation of US$6.8 billion led by Salesforce, followed by SoftBank and Goldman Sachs. In 2018, it completed two rounds of financing within half a year, totalling US$500 million. Currently, Automation Anywhere serves more than 3,500 corporate entities in more than 90 countries, including large medical, communications, financial, and manufacturing companies such as Google, Linkedin, Siemens, Dell, and MasterCard.
Founded in 2013, Babylon Health is a UK-based digital healthcare services startup. The company is committed to making healthcare accessible and affordable for everyone through artificial intelligence technology. Artificial Intelligence Doctor is a remote diagnosis and treatment application recently launched by Babylon Health which can provide users with round-the-clock medical consultation services. Babylon Health has partnerships with Tencent, TELUS and Samsung to provide clinical services in the UK, Rwanda and Canada. The company currently has 4.3 million worldwide users, can provide 4,000 clinical consultations every day, and has completed more than 1.2 million digital consultations. In August 2019, Babylon Health announced the completion of a US$550 million Series C financing round. Going forward, Babylon Health will expand its business scope to the United States and Asia, and further develop AI diagnostic platforms for severe chronic diseases.
In 2019, Cambricon continued to develop its two major product lines: “intelligent terminal processor IP” and “intelligent cloud server chip”. In June, Cambricon launched the Chinese brand of cloud AI chips “Siyuan”, the second-generation of the cloud AI chips (MLU270) “Siyuan 270” and board card products. The company provides customers with energy-efficient solutions in multiple areas such as intelligent video analysis, speech synthesis, recommendation systems, and AI cloud. At the same time, the company has improved new products’ theoretical peak performance for processing non-sparse deep learning models by four times that of the previous generation “Siyuan 100”. In November 2019, Cambricon launched the edge AI series products: “Siyuan 220 (MLU220)” chips and module products, achieving a comprehensive and three-dimensional coverage in cloud, edge, and terminal devices.
Founded in 2011, Munich startup Celonis is dedicated to helping companies complete digital transformation and optimization of internal processes. Its machine learning-driven intelligent system automatically analyzes and sorts out business processes such as ordering products, payment, shipping, and delivering products based on the log data created by IT systems such as SAP, Salesforce, Oracle and Netsuite. Enterprise can view the resulting process, determine which steps cause delays and check for compliance or other vulnerability issues through the Celonis dashboard. After a US$50 million funding round led by Accel and 83North in June 2018, Celonis’ valuation hit US$1 billion In November 2019, Celonis announced that it had received US$290 million in Series C financing, raising company valuation to US$2.5 billion.
In 2019 CloudMinds built a cloud intelligent platform that supports millions of scale terminals. Its cloud robots serve multiple industries such as real estate, hotels, banking, education, and healthcare. CloudMinds has also launched a Robotic Development Kit (RDK) for third-party developers, enabling real-world robots and digital twin robots to be developed simultaneously. In the first half of 2019, the revenue of CloudMinds increased by 346% year-on-year to US$148 million, exceeding the full-year income of 2018. The company made small-scale sales of robots to the US market in the second half of 2019, while preparing to enter the Japanese market in 2020.
In 2019, CloudWalk Technology made great progress in various business lines. The company employs more than 1,500 people, and has attracted and retained some of the world’s best talents. CloudWalk’s core technologies have won the international intelligent perception crown 13 times and the 158 industry POC championships with a winning rate of 78.4%. Through the use of advanced three-level research and development architecture, CloudWalk has achieved major technological breakthroughs. It produced the first domestic 3D structured light face recognition technology, and its first commercial cross-mirror tracking technology Re-ID has maintained records to date. The company’s 3D reconstruction technology can boost algorithm speed by 20 times and accuracy by 30%.
After first three years’ innovative cooperation with leading financial institutions, in 2019 CraiditX expanded its clients by several times. Along with the China Merchants Bank and Industrial and Commercial Bank of China, the company cooperates with the China Construction Bank, SPDB, Ningbo Bank and other advanced banking clients to achieve full coverage for leading financial institutions. With its independently developed non-probability engine, CraiditX offers solutions ranging from risk control to transaction anti-fraud, customer operation and intelligent customer service, providing intelligent integrated solutions for financial clients. CraiditX received strategic investments from GP Capital, Shanghai Puxin Capital which is set up by Shanghai Pudong Development Bank and Shanghai Trust.
In 2019, in terms of innovation, DataGrand launched the first domestic intelligent RPA product that integrates self-developed NLP and OCR technology, expanding the boundaries of enterprise automation. In terms of industry-university-research, the company established a joint artificial intelligence and language research centre with the Shanghai Jiao Tong University School of Foreign Languages and set up a joint laboratory of NLP and search engines with Country Garden to consolidate the technical strength of semantic understanding. DataGrand has reached in-depth cooperation with Microsoft, Ernst & Young, and Times Real Estate to promote the development of intelligent automation ecology. In terms of commercial landing, many companies including Shenzhen Stock Exchange, GF Securities, Vanke, Canon, Pacific Insurance, China Everbright Bank, China CITIC Bank, and China Construction Third Engineering Bureau have adopted DataGrand’s intelligent text processing products.
Founded in 2016, Element AI is dedicated to transforming research and industry expertise into software solutions that can learn and improve at an exponential rate, designed to help large organizations implement AI and make it truly impact their business. Element AI’s products include consulting services, AI-based tools and products. Financial services and supply chains are their areas of focus. Customers include South Korea’s LG Group, National Bank of Canada, and more. Element AI’s employees are scattered across its Montreal headquarters and offices in Toronto, London, Seoul and Singapore. In September 2019, Element AI announced US$151.4 million in Series B financing led by the Canadian Public Pensions CDPQ, followed by original shareholders such as McKinsey and the Government of Quebec.
Graphcore is an AI chip company based in Bristol, UK. In December 2018 it completed a US$200 million Series D financing with a valuation of US$1.7 billion. In 2019, the Graphcore Intelligent Processing Unit (IPU) chip was mass-produced, sold and shipped. In November Graphcore and Microsoft officially released a preview version of Graphcore IPU on Microsoft Azure. This was the first time that a major public cloud provider had provided Graphcore IPUs. Also in 2019, Graphcore released the Graphcore IPU Benchmark, and the Cirrascale IPU cloud equipped with Graphcore IPU went online for leasing. Dell EMC servers equipped with Graphcore IPU C2 cards also received widespread industry attention.
Infervision continues to lead the medical AI market with its global vision and foresight, and actively promotes globalization. It has become the only Chinese medical AI company on the track in North America, Europe, the Asia Pacific and Africa. In March 2019 Infervision launched a multi-modality all-disease research platform based on deep learning to provide hospitals and doctors with a full-stack intelligent research infrastructure. In addition, Infervision launched a chest CT multi-disease solution, leading industry development. Infervision has covered 9 countries worldwide and 32 provincial-level administrative regions across China, and is rapidly developing into a large-scale AI medical service system worldwide.
In 2019, Megvii was awarded the “New Generation of Artificial Intelligence Innovation Platform for Image Perception” by the Chinese Ministry of Science and Technology, jointly established the “BAAI-Megvii Intelligent Model Design and Image Perception Joint Laboratory” with the Beijing Academy of Artificial Intelligence (BAAI), and jointly released Objects365, the world’s largest object detection data set.
Megvii and self-developed artificial intelligence algorithm platform Brain ++, have won more than 10 world scientific and technological championships and published more than 30 papers in academic summits such as CVPR, ICCV, etc. Megvii products include the industry’s first AIoT operating system-Megvii HETU and a series of leading IoT solutions such as night scene super-quality photography. Megvii has established strategic cooperation with a number of companies and institutions such as China Telecom and China Mobile to comprehensively cultivate the three major IoT scenarios of personal devices, smart cities, and supply chains in the 5G + AIoT era.
Mininglamp Technology Group is China’s leading one-stop enterprise-level artificial intelligence products and service platform. In March 2019 Mininglamp Technology announced the Group’s brand strategy upgrade and completed a Series D financing round of CN¥2 billion. In August 2019, the Ministry of Science and Technology of China announced a plan to build the National New Generation AI Open Innovation Platform with the support of Mininglamp Technology Group in developing “marketing intelligence”. The Mininglamp Technology Group holds more than 1,200 patents. Mininglamp Technology is deeply involved in the Shanghai Railway Transport Vehicle Intelligent Operation and Maintenance System which was evaluated as a national demonstration project of the rail transit industry approved by the National Development and Reform Commission in 2019. The First Research Institute of the Ministry of Public Security of PRC and Mininglamp Technology jointly published the White Paper on Standardization of Public Security Knowledge Graph, which has actively led the industry’s standardization direction.
In March 2019, Momenta released the Mpilot Highway, which offers maximized continuity of the autonomous driving experience covering highways and city expressways with low-cost, mass production ready sensors. In July, the Mpilot Parking was released. Mpilot Parking provides an enjoyable intelligent valet parking experience via superior surround view perception, high precision in-door mapping and localization technologies. In December Momenta released the L4 level autonomous driving product MSD (Momenta Self Driving), which can realize completely autonomous driving in urban scenarios. The successive release of Mpilot and MSD establishes the company’s two-leg strategy. Data, data-driven algorithms and closed-loop automation between Mpilot and MSD are the key to achieving autonomous driving at scale.
Located in Silicon Valley, autonomous driving startup Nuro was founded in 2016 by Zhu Jiajun and Dave Ferguson, founding members of Google’s earlier self-driving car project (now Waymo). Its main product is self-driving vehicles, and the company previously reached a distribution cooperation agreement with Kroger, an American grocery chain. In February 2019, Nuro announced that it had received US$940 million in funding from the SoftBank Vision Fund. This amount exceeds the US$530 million raised by Aurora, a start-up self-driving startup that also received financing in February.
Petuum released its first industrial artificial intelligence product in March 2019, establishing an industrial AI framework that enables closed-loop “smart operation” and provides industrial customers with control capabilities for equipment and processes that far exceed today’s automation systems. The product provides predictive maintenance and operation optimization management for the cement, metallurgy, oil and gas, smelting, and mining industries. In July, Petuum released Neurobots, an artificial intelligence SaaS product that can be used to enhance the capabilities of enterprise-level robotic process automation (RPA) software in text understanding, computer vision, and dynamic decision-making. The product is targeted at customer support, call centers, virtual assistance, intelligent inventory management and more. Petuum has also published papers on a general framework for sequence generation learning algorithms, which have greatly improved the efficiency of machine translation and text summary tasks.
Founded in November, 2013, Recursion is a clinical stage biotechnology company headquartered in Salt Lake City, USA. Recursion combines experimental biology and automation with artificial intelligence in a massively parallel system to effectively discover potential drugs for a variety of indications, including genetic diseases, inflammation, immunology, and infectious diseases. In July 2019, Recursion announced the completion of a US$121 million Series C financing round. The new financing will support Recursion in continuing to build its machine learning drug discovery platform, while providing new capabilities designed to significantly accelerate new molecular entity (NME) chemistry and predictive safety pharmacology. In addition, the company will continue to advance its preclinical and clinical asset pipeline, including clinical plans for cerebral cavernous vascular malformations and neurofibromas. In 2019, Recursion partnered with Takeda to evaluate the unique indications of more than 60 preclinical and clinical compounds of Takeda, and identified new candidate therapeutics in more than 6 diseases.
UPhoton Technologies (UPhoton) is an internationally competitive provider of advanced solutions for micro-nano diffractive optics and three-dimensional sensing modules and systems. UPhoton is developing rapidly, and in 2019 carried out continuous and in-depth innovation layout in optical elements and module systems. The company has actively explored multiple market segments and achieved mass production applications while launching a variety of AI vision solutions such as 3D face payment, 3D smart door locks, and 3D vision industrial measurement applications. UPhoton has completed CN¥100 million Series B and B + funding rounds. It has received strong approval from new shareholders such as China Merchants Venture, THG Ventures and strong support from old shareholders such as Legendstar and ZhenFund.
In 2019 SenseTime accelerated its commercial landing. The company covers multiple industries such as smartphones, Internet entertainment, automobiles, smart cities, education, medical, retail, advertising, and real estate. In September 2019, Beijing Daxing International Airport was officially put into operation, and SenseTime AI technology helped serve global passengers. SenseTime has achieved outstanding results in global computer vision related competitions, 62 papers were selected for CVPR 2019, and 57 papers selected for ICCV 2019 (including 11 Oral), ranking first. The company has won multiple competitions and was selected to lead the National Face Recognition Working Group to promote AI security applications.
Founded in September 2015, Tempus is a startup based in Chicago, Illinois that uses artificial intelligence to develop precision medicine. Tempus has built the world’s largest molecular and clinical database and uses machine learning, next-generation DNA sequencing, and AI-driven image recognition technology to analyze clinical and genomic information data in the database. The company uses this to generate high-quality somatic and germline molecular data and therapeutic backgrounds, enabling physicians and researchers to make real-time data-driven decisions.
In May 2019, Tempus announced that it had completed a US$200 million Series F round of financing with a total of US$520 million raised, pushing the company’s valuation to US $3.1 billion. Tempus’ initial work focused on identifying potential cancer therapies, and in the future it will use the financing to expand its operations and its business into new areas of precision medicine such as diabetes and depression.
TuSimple is an artificial intelligence enterprise devoted to the development and application of Level-4 autonomous trucking solutions which could be put into large-scale commercial deployment in the global logistics industry. In April 2019, TuSimple was a first selection for for the artificial intelligence pilot application scenario (AI+Transportation) in Shanghai, carrying out unmanned MTR intermodal transportation between Yangshan Port and LuChaoGang Railway Station on Donghai Bridge. In May, TuSimple cooperated with the United States Postal Service (USPS) to provide driverless services on more than 1,600 kilometres of transportation lines; in August, TuSimple received a strategic investment from UPS to manage its freight routes between Phoenix and Tucson; and in September, the company completed a total of US$215 million in Series D funding from Sina Capital, CDH Investments, UPS, and Mando Corporation; and in December TuSimple successfully completed the first L4 unmanned and queue-following demonstration in China, based on C-V2X technology in the fully enclosed environment of two-way four-lane highways on the Jingli Expressway (Yanchong Beijing Section).
In 2019, UBTECH launched Iron Man MARK50 robot, AIMBOT, and the Jimu Robot series including WarriorBot, UnicronBot, ChampBot, ScoreBot, etc. The company’s large humanoid service robot, Walker, has made great breakthroughs in areas like motion performance, flexible interaction, and environmental perception. UBTECH continues to explore the application of artificial intelligence robots in fields such as education, public safety, smart health care, Party construction and government affairs, business services, and computer room operation and maintenance. At the 2019 RoboCup, the Hephaestus team, jointly formed by UBTECH and Tsinghua University, won three awards in the humanoid group, including runner-up in the AdultSize Technical Challenge, runner-up in Drop-in game and third place in the 2VS2 football match season. UBTECH Sydney Artificial Intelligence Centre research papers have been accepted by a number of top global academic conferences, such as CVPR, ICCV, and NeurIPS.
Founded in 2005, UiPath is a high-profile enterprise software company in the global robotic process automation (RPA) industry. UiPath aims to provide enterprises with a complete set of enterprise service software platforms to help them effectively automate business processes and has more than 400,000 users in more than 200 countries and regions around the world. In May 2019, UiPath received round D financing of US$568 million, which pushed its post-investment valuation above US$7 billion, making it the unicorn of the RPA industry. UiPath officially landed in China in 2018, and is currently cooperating with more than 100 universities and nearly a thousand enterprises nationwide.
In January 2019, Waymo, the autonomous driving company of Google’s parent company Alphabet, announced that it would build the world’s first factory dedicated to mass production of L4 autonomous vehicles, in Detroit, Michigan. In September 2019, Waymo CEO John Krafcik announced that it had completed the production and modification of 30 Jaguar Land Rover I-PACE vehicles in Detroit and was now testing them in California. The Waymo Robotaxi driverless taxi service received a passenger license from the California Public Utilities Commission (CPUC) in July 2019, serving 4,678 riders in the first month, and currently serves at least 1,500 a month across the United States. Although the development of the autonomous driving industry in 2019 was slower than expected, the international investment bank Morgan Stanley still expected Waymo’s valuation to reach US$105 billion in September 2019. The company’s revenue comes from autonomous taxis, logistics and software licensing.
In 2019, YITU released its self-developed cloud AI chip Questcore, establishing a core technological foundation of AI software, hardware and integrated algorithms, chips, and platforms to become a leading AI infrastructure and solution provider. At the algorithm level, YITU has achieved a number of world-class breakthroughs — it continues to lead the gold standard in face recognition industry. The company’s NLP technology based paper was published in top journal Nature Medicine, and it won the global authoritative voiceprint recognition competition VoxCeleb Speaker Recognition Challenge (VoxSRC) championship by a large margin. In 2019, the Chinese Ministry of Science and Technology announced that it will rely on YITU to build the National Open Innovation Platform for Next Generation Artificial Intelligence on Intelligent Vision to promote the combination of AI and chip design and the construction of an upper-layer AI ecosystem.
In 2019, ZHUIYI completed its Series C financing, focusing on the strategic positioning of “AI digital employees”, and consolidating the foundation of its products, technologies, services and ecosystem. It formed a full-stack AI service capability with semantic, voice, and visual multi-modality, and maintained its leading position in core AI interaction technology, setting a new conversational reading comprehension CoQA challenge record and releasing the first large-scale NL2SQL Chinese dataset. Digital employees’ family of products are constantly standardized and scaled up, and new industrial AI scenarios such as anti-money laundering and Data Middle Office and Business Middle Office approaches are being implemented. ZHUIYI is also a leader in the financial intelligent service market and has made breakthroughs in real estate, operators, and government affairs.
Top 10 AI Rising Stars in China
Founded in March 2019, CISAI Tech is a joint venture between CISDI, a subsidiary of China Minmetals Corporation, and AInnovation, an AI subsidiary of Sinovation Ventures. The company focuses on the industrial field and provides overall solutions based on industrial AI around two major application scenarios: park logistics and intelligent manufacturing. The world’s first Autonomous Locomotive Pilot created by the company in conjunction with CISDI and Zhanjiang Steel has been fully launched. It is an innovative practice of artificial intelligence technology in industrial scenarios and has a landmark significance in the process of steel intelligence. In less than a year since its establishment, CISAI Tech has served 12 large domestic enterprises and achieved profitability in the first year of operation.
Founded in July 2018, DMAI is committed to creating a new generation of strong cognitive artificial intelligence technology platforms. The company’s core technologies are human-computer interaction and cross-domain integration. It is the only enterprise team in Asia to win a ICCV 2019 Best Paper Nomination Award. The research results of founder Professor Song-Chun Zhu’s laboratory team also appeared in the journal Science Robotics recognizing a new generation of artificial intelligence. DMAI has accelerated the application of AI + education. In the future, it plans to gradually integrate with business sectors such as finance, health, new retail, pan-entertainment, government and judicial affairs, and promote innovative industry business models. It is reported that the company has obtained tens of millions of US dollars in pre-A round investments from well-known investment institutions such as Sailing, IDG, Elevation, Jiangmen, Lingang, etc.
Eigen is committed to using advanced artificial intelligence technology to improve the efficiency of content production. The founding team is mainly from Microsoft, IBM and Alibaba. As of the end of 2019, it had received tens of millions of dollars in investment from institutional ventures including SIG, Qiming Venture Partners, Sequoia and Gaorong Capital; and reached strategic cooperation/investment agreements with UBTECH and Kuaishou. In 2019, Eigen and Fengchaowulian worked closely to launch China’s most advanced automotive graphics and text intelligent editor Bumblebee. Eigen launched Modi Video on the Alibaba service market to generate short marketing videos for merchants with one click. Modi Design, an automated marketing poster design on WeChat, has also gained favour from its early users. Eigen believes that technology creates value, and hopes to become the backbone of the productivity release of rich media production in this CN¥100 billion market.
Founded in March 2018, Enflame is targeting AI training platform solutions for cloud service providers and data centres. In December 2019 Enflame launched its first AI training accelerator card “CloudBlazer T10”. Its innovative architecture, interconnection, distributed training and programming platform can be widely used in multiple AI application scenarios such as cloud data centres, supercomputing centres, the Internet, finance, and smart cities. In June 2019, Enflame announced it had received US$42 million in A-round financing. The company has been invested by major venture capitals such as Tencent Ventures, Summitview Capital, Redpoint China Ventures, and Shanghai STVC Group.
Miaobi Intelligence was founded in 2019, focusing on the development and application of natural language generation (NLG) technology. It has established in-depth technical cooperation with the top NLP team at Wang Xuan Computer Research Institute of Peking University and launched the Miaobi Intelligence intelligent creation platform, which can provide enterprises with intelligent services such as robotic writing, adaptation, video generation and publishing. The company has served in the fields of corporate marketing communications, media intelligence, e-commerce copywriting and video content generation. Miaobi won third prize in the National Finals in the first content technology entrepreneurship competition held by People’s Daily Online (ranked first in NLP related fields). At present, Miaobi has completed its angel round of financing, led by Baidu Ventures, followed by Lenovo Capital.
Founded in January 2019, the Parametrix.ai team comprises experts and scholars from both artificial intelligence and the gaming industry, as well as distinguished institutions around the world. The core members have conducted long-term research and development on cutting-edge technologies such as deep learning and reinforcement learning. Combined with their engineering ability on large-scale systems, the team successfully commercialized AI technology in the most popular games. In 2019, Parametrix.ai worked together with a number of top game companies and launched multiple AI products, including the 3D survival game AI “Orion α” and Imperfect Information Game AI. The team also expanded into additional game categories. To ensure the universal application of AI technology, Parametrix.ai has developed the Lambda AI platform and Delta AI engine. The company is committed to the innovation of gaming industry with AI. Parametrix.ai secured its Series A funding from Morningside Venture Capital and Gaorong Capital.
Pi2star is a high-tech company from Tsinghua university, founded in September 2018. The company takes high-efficiency industrial edge AI chips as its core, and builds a full-stack fast-type product system from chips, algorithms, edge acquisition/computing terminals, and APP/PC products. The full stack device management AIoT solutions have been applied in many industrial scenarios. In January 2019, Pi2star released the Edge AI chip Sticker-T; and in July released the one-dimensional signal AIoT full stack product, which consists of STARCORE, NEBULA, ASTROLOGY, and STARDUST. It has completed multiple rounds of iterations and obtained high customer approval. In 2019, the company established cooperation agreements with a number of front-loaded customers, such as pump companies and motor companies, and provided products and services to customers in various industries, such as petroleum and petrochemicals, steel and metallurgy.
Founded in July 2019, Renminzhongke is a high-tech enterprise jointly established by the People’s Daily Online and the Institute of Automation, the Chinese Academy of Sciences (CASIA). The company takes machine content understanding technology as its core business and provides customers with artificial intelligence technologies and services in different application scenarios, including video, audio, image, text semantic understanding, scene recognition, target retrieval, etc. In 2019, the company was committed to the field of content security, and under the concept of “artificial intelligence + content security” launched two major product systems: “Internet audio and video content risk control platform” and “network heterogeneous media monitoring system” to provide intelligent technical support for cyberspace governance.
Established in July 2018, Tsing Micro provides Coarse-grained Reconfigurable Architecture, CRGA. This is a new chip architecture technology which is different from the traditional Von Neumann architecture and can flexibly reconfigure hardware resources according to different application requirements. This technology has the advantage of the flexibility of general-purpose computing chips and the efficiency of application-specific integrated circuits. The voice AI soc TX210, which was mass-produced in June 2019, is the first commercial use of reconfigurable chips, with shipments reaching one million, and is used in mobile phones, homes, toys and smart wearable devices. TX510, released in September, also supports multi-modal intelligent processing such as voice and vision, and is the world’s first multi-modal intelligent computing chip.
Silexon is an emerging AI/biotechnology company which aims to create an open AI platform for strategic collaboration in order to facilitate data-driven life science research and empower drug R&D process, and ultimately provide patients greater access to innovative drugs for unmet medical needs — which is the Silexon mission statement. The company’s current AI/ML algorithms are focused on early drug discovery for small molecules and peptides, although they have other algorithms for target identification, cell-based therapy, protein expression and translational medicine. Silexon has multidisciplinary expertise and scientific acumen to develop cutting-edge algorithms to empower emerging technology in drug R&D, which makes it stand out among its AIDD peers. The company has established several strategic collaborations, with biotech companies, Chinese public companies and EU-based biotechs with billions of Euros in revenue. Research focuses on therapeutic areas such as rare diseases, infectious disease, metabolic syndrome, oncology, etc. with drug candidates spanning from traditional small molecules, phytochemicals, peptides and all the way to cell therapy. In November 2019, Silexon announced it had successfully completed a Pre-A round of financing.
Top 10 Industry AI Research Labs in China
During the past year, Alibaba Group’s DAMO Academy, T-Head (Pingtouge), Alibaba Cloud, and other research teams have published more than 270 papers covering topics including natural language processing, intelligent speech, visual computing, chip architecture, self-driving, databases, cloud computing, and other cutting-edge technology fields. Top institutions have recognized these research breakthroughs, and Alibaba scientists have been recognized as academicians or elected Fellows of top professional associations.
Beyond basic technology research, Alibaba has also notched up many ‘No. 1’ honours globally. For example, Hanguang 800, as a leading AI inference chip, greatly boosts computing capabilities while improving energy-efficiency. Alibaba’s cutting-edge technologies and products have been deployed on a large scale, and its innovative AI algorithms have been used in healthcare, transportation, industrial manufacturing, and other fields.
As a world-leading artificial intelligence platform, Baidu has been continuously advancing the state-of-the-art in the fields of speech processing, image processing, natural language processing, knowledge graphs and deep learning. The company has won 23 top-level AI competitions and had over 200 papers accepted by top-tier academic conferences such as CVPR, ICCV, NeuroIPS, MRQA. Baidu has 5,712 AI patent applications — the most in China — and has trained more than 1,000 AI professional teachers and hundreds of thousands of senior technical personnel for many enterprises. Taking the intelligent cloud as a carrier, Baidu is actively promoting the widespread application of AI in finance, customer service, smart city, healthcare, Internet, and other scenarios, and has become a leader in China’s industrial intelligence space.
In 2019 the ByteDance AI team continued to apply artificial intelligence technology to product platforms in various content forms including voice, audio, pictures, video, and games. They launched Landmark AR technology for their video sharing social media platform Tik Tok in November. ByteDance has also increased the application of AI to combat vulgar content. Its AI tool Lingquan can better identify violent and sexual content, and has already been adopted by more than three million Toutiao users and content producers. While upgrading the ability to recognize vulgar content in text, it has also added image recognition and speech recognition functions to support users in identifying vulgar content in pictures and voice content. In June, ByteDance released a general-purpose high-performance distributed training framework, BytePS, which performs significantly better than the current open-source distributed training framework.
In 2019, Hikvision promoted the concept of “Fusion of Intelligent IoT and Information Networks” and showcased its AI Cloud Data Fusion Platform to further consolidate the operations in the fields of AI, big data, and applications. As of now, the application cases on the Hikvision AI Cloud Data Fusion Platform have covered 24 provinces and nearly 100 cities across China. At the software layer, Hikvision has officially released its AI Cloud series of software products, with applications spanning a wide range of more than 20 industries, including public safety, transportation, manufacturing, retail, education, and logistics. At the hardware layer, Hikvision is continuing to improve the hardware product mix, especially in the areas such as multi-dimensional perception and comprehensive situation awareness, and has launched AI-powered multi-lens cameras and radar-assisted traffic cameras. Responding to fragmented industry demand for AI technologies, Hikvision has been effectively empowering hundreds of industries through its AI open platform.
Huawei has created one of China’s top 10 Open Innovation Platforms for next-generation AI. With leading full-stack, all-scenario AI capabilities, the platform provides AI services for governments and enterprises. As of October 2019, Huawei Cloud had provided 60 types of AI services with 160 AI functions and had applied these functions in over 500 projects in more than ten industries, including smart city, manufacturing, logistics, Internet, healthcare, and smart campus. Huawei has successfully assisted Shenzhen Traffic Police, Tianjin Eco-City, PetroChina, Sanlian Hope, Golden Stone Group, Kingmed Diagnostics Group, and other enterprises with intelligent upgrades. Huawei Cloud proposed four elements essential to accelerate AI implementation in industry and officially launched the Huawei Cloud EI Cluster Service powered by Ascend 910, the world’s most powerful AI processor. To jointly build inclusive AI, Huawei also released 43 types of Ascend-powered AI cloud services, ModelArts 2.0 (one-stop AI development and management platform), City Intelligent Twins, Industry Intelligent Twins, and the next-generation intelligent all-flash storage OceanStor Dorado.
iFLYTEK has been engaged in the research of core technologies and has maintained an international cutting-edge technology level. The company actively promotes the development of artificial intelligence products and the implementation of industrial applications, and is committed to enabling machines to listen, speak, understand and think. Based on independent intellectual property rights of its core technology, iFLYTEK has continued to expand into the industry of intelligent products and services. As a leader in the field of artificial intelligence, iFLYTEK adheres to a strategic layout driven by technological innovation, with remarkable results.
In August 2019 JD was selected to develop China’s National Open Innovation Platform for Next-Generation Artificial Intelligence in Smart Supply Chains, with NeuHub, JD’s AI Open Platform, as the core technology. JD AI continues to develop large-scale AI applications in fields such as municipal services, retail, customer service, healthcare, and more. Since 2018, the JD AI Research Institute has published more than 130 papers in top AI conferences and journals and has won 13 international AI championships. In 2019, JD AI research won 3 first prizes in CVPR academic contests. Financial partner JD Digits has comprehensive AI capabilities. It has technologies and application innovations in machine learning (mainly Auto ML, federal learning, deep learning), 3D computer vision, speech recognition and synthesis, and NLP. JD.com has launched dozens of AI research projects and has cooperated with more than ten scientific research institutions and universities, including the Chinese Academy of Sciences, Tsinghua University, Shanghai Jiaotong University, and the University of Pittsburgh.
Microsoft Research Asia
Deeply rooted in basic and applied research of computer science for 21 years, Microsoft Research Asia has always been committed to advancing the state of the art of computing, and incubating next-generation revolutionary technologies for the future. In 2019, Microsoft Research Asia continued to explore the boundaries of computer science and made a large number of scientific achievements. It published more than 230 papers at top international conferences in computer science, including ACL, EMNLP, NeurIPS, ICML, AAAI, IJCAI, CVPR, ICCV, SIGGRAPH, KDD, WWW, SIGCOMM, NSDI, and MobiCom etc., covering areas from machine learning and natural language processing to computer vision, graphics, visualization, speech, data mining, software engineering, systems, networks, databases, and security.
In 2019 Tencent had 256 papers selected by top academic conferences and journals such as ICML, CVPR, NeurIPS, and ACL. Demonstrating industry-leading research capabilities, it won China’s first Best Long Paper award at ACL, and won competitions including DSTC8. In AI + games, Tencent and King Glory developed ‘Juewu’ AI, which has reached a professional level, and launched the Kaiwu open platform to deepen ecological construction. Tencent Miying, Tencent Yidian, and other products are active in AI + medical, along with self-developed intelligent pathological microscopes. In AI + agriculture, Tencent and Wageningen University co-hosted the second greenhouse competition and successfully launched the iGrow AIoT greenhouse landing solution. In AI + public welfare, Tencent Youtu’s cross-age face recognition technology helped retrieve 14 children who had been abducted for over ten years.
Tencent has open-sourced MedicalNet, the medical deep learning pre-training model. The company’s Angel became China’s first LF AI Foundation graduation project and has since been upgraded to a full-stack machine learning platform. Tencent also released the FewRel 2.0 and DocRED datasets, and has developed more than 30 joint research projects with universities.
The WeBank AI department has scored a series of impressive achievements and applications in four main areas: Federated AI Ecosystem, Intelligent Services, Intelligent Precision Marketing, and Intelligent Asset Management. The department has applied for more than 200 patents. WeBank AI first proposed the “Federated Learning” universal solution in China to solve problems of data silos and privacy protection in the large-scale implementation of AI. The company leads and promotes the construction of a global federated AI collaboration ecosystem. WeBank AI open-sourced the world’s first industrial-grade federated learning platform “FATE” (Federated AI Technology Enabler) and donated it to the Linux Foundation. WeBank AI has organized international workshops on federated learning at IJCAI and NeurIPS; led the formulation of international standard (IEEE standard) and national standards of federated learning; and promoted industrial applications of federated learning in finance, healthcare, retail and so on. Hundreds of enterprises, universities and research institutions have already joined the effort to build this federated AI collaboration ecosystem together.
30 Industry Leading AI Transformation Business Cases
Reconova Builds a Full Process Face Scan Boarding Solution For Daxing Airport
After many large-scale comprehensive drills and much debugging, the Daxing International Airport full-process face scan boarding project completed acceptance inspection and was officially put into use on September 25, 2019. The project deployed more than 300 Reconova face recognition products at the security checkpoints and boarding gates of Daxing Airport, covering more than 90% of face recognition operations at the airport. The system includes face scan security checks, smart aviation display, manual counter face scan boarding, self-service boarding gate and various other applications.
Hikvision AI Open Platform Identifies Underground Safety Risks in Jiangzhuang Coal Mine to Create a ‘Zero-Accident’ Working Environment
In May 2019, Hikvision and Jiangzhuang Coal Mine formed a joint working team and developed an intelligent safety risks monitoring and management system for coal mines, relying on the Hikvision AI open platform. Aiming at “identifying safety risks and building closed-loop process”, the team implemented this value-added project to improve coal mine safety through intelligent applications. Jiangzhuang Coal Mine was the first company in the industry to deploy Hikvision AI edge equipment in critical areas of the mine such as coal haulage ramps. Through this intelligent safety risks monitoring and management system for coal mines and other relevant systems, Jiangzhuang Coal Mine has worked out multiple intelligent solutions to address safety challenges, and created a “zero-accident” working environment.
China National Electric Nanning Power Generation Co. Ltd. Boost Power Plant Boiler Thermal Efficiency by 0.5% with JD iCity AI Technology
JD Digits’ AI-optimized thermal power generation solution developed a boiler combustion simulator based on deep learning and a thermal power generation optimization model based on reinforcement learning. By fully analyzing and processing the historical data of thermal generator units, the system can dynamically control various links of boiler valves such as opening, air supply and water supply. The project was guided by the Information Department of the CHN Energy Investment Group, carried out by China National Electric Guangxi Power Generation Co. Ltd., jointly developed by Beijing Huadian Tianren Power Control Technology Co., Ltd. and JD iCity, and finally completed deployment and application at China National Electric Nanning Power Generation Co. Ltd., where it improved boiler thermal efficiency by 0.5%.
HiAR Apply AR to More Than 50 Haier Smart Factory Scenarios Including Remote Maintenance and Operation Guidance
HiAR cooperated with Haier’s Industrial Intelligence Research Institute to integrate AR into the Haier “Smart Factory” project, with more than 50 business scenarios including remote maintenance, operation guidance, equipment inspection, quality monitoring, etc. There are 15 inter-connected factories worldwide, including Haier’s magnetic central air-conditioner interconnected factory, Haier’s Tianjin Inter-connected Laundry Appliances Factory, and Haier’s Russia Inter-connected Laundry Appliances Factory. The two parties are jointly carrying out research in multiple aspects of intelligent manufacturing (marketing, manufacturing, management, etc.), and trying to set a benchmark for AR applications. They plan to expand to 122 Haier global factories and other manufacturing companies around the world.
Dorabot Develops Optimal Container Booking Strategy and Loading Planning Algorithm to Help Basf Improve Container Utilization Rate and Order Response Speed
The customized AI-CLP (AI-powered Container Loading Planner) developed by Dorabot for the chemical industry giant BASF is an optimal container booking strategy and loading planning solution that can replace manual operations. It is able to meet loading requirements in various complex scenarios, such as cargo load-bearing capacity, collocation specification, and weight distribution balance. It also provides multiple container booking and loading options. Built upon the Dorabot Tesseract intelligent logistics platform, this software is easy to use and highly efficient. This platform can improve the loading rate and reduce operation costs for BASF. The two parties have completed a pilot trial at BASF’s Nansha factory, setting out a firm step on supply chain innovation and intelligence.
The No.10 People’s Hospital of Shanghai Applies Synyi Intelligent System for Venous Thromboembolism Prediction, Auxiliary Diagnosis and Treatment
Six months after starting the use of the Synyi Intelligent system for Venous Thromboembolism prediction, auxiliary diagnosis and treatment in The NO.10 People’s Hospital of Shanghai, the VTE risk assessment ratio increased significantly from 53.75% to 98.8%, and the evaluation efficiency also improved greatly from 15 minutes to 1-2 minutes per patient in this hospital. It’s estimated that the system can save an average of 600 working hours for medical staff per month.
Non-Profit Health Service Organization CommonSpirit, With US$29 Billion in Annual Revenue, Deploys Notable’s Robotic Process Automation Platform to Simplify Physician Services
In December 2019, CommonSpirit Health announced a partnership with Notable Health that will improve both patient and provider experience by automating health information collection and making clinical record systems interoperable. Notable’s platform uses robotic process automation (RPA) to understand data from electronic health records and deploys artificial intelligence models that anticipate and support providers’ needs so they need to spend less time on administrative tasks and can devote their time to providing compassionate care.
Online Medical Inquiry App “WeDoctor” Applies AI Technology to Doctor-End Products to Strengthen Management of Internet Diagnosis Service Providers
In 2019 the digital health platform WeDoctor applied cutting-edge technologies such as face recognition, big data risk control and CA digital certificates to doctor-side products for the first time in the industry to strengthen the management of Internet-based diagnosis service providers. In addition to the real-person authentication in the registration process, WeDoctor also introduced a big data risk control system to control risks related to doctors’ behaviour. This system extracts user behavioural portraits from technical dimensions such as IP, equipment, and account behaviour. It then incorporates big data and machine learning to identify abnormal accounts. The risk control system has been extended to various fields such as business anti-fraud, intelligent verification codes, dynamic real-life verification, medical data risk control, and business content security.
Wuhan Union Hospital’s Operating Room Uses the TMirob High Valued Consumables Delivery and Management Solution for High Valued Consumables to Achieve Consumables Delivery and Management
Union Hospital affiliated to Tongji Medical College of Huazhong University of Science and Technology uses the TMiRob high-value consumables delivery robot and supporting consumables management software for operating room high-value consumables. This completes a closed-loop traceability from delivery of cargo to storage and usage, and greatly reduces the wait time for consumables during surgery, thus managing the entire consumables life cycle. Delivery error and missed account rates have been significantly reduced. In addition, a Second-class Warehouse management software with Product Unique QR Code is used to achieve precise management of the delivery process and cost control, helping hospital management groups effectively control the proportion of high-value consumables.
DiDi Launches AR Navigation Feature for Better Travel Efficiency at Transit Hubs
To improve navigation at transit hubs, DiDi took the lead in applying AR technology to mobile travel scenarios. Through technical innovations in computer vision positioning and 3D scene reconstruction, DiDi rolled out in-app AR navigation features to help passengers navigate complex indoor environments such as airports, train stations and shopping malls. The “AR Navigation” function will provide passengers with directions to vehicle pick-up points via augmented arrows and offer information such as distance to the pick-up point in real-time. This is the first such large-scale application of AR technology in ride-hailing scenarios.
Tianrang Builds an Online City Brain Operating System for Shanghai, Helping the Government to Achieve a “One Screen View” and “One Network Management” Megalopolis Governance Model
From May to November 2019, through all sources data integration across the entire network and domains, the Shanghai City Brain Operating System developed by Tianrang Intelligence incorporated public security, transportation, communication, emergency commissions, electric power, environment, and other government commissions and bureaus into a unified online platform system to achieve a “one screen view” city operation monitoring system and a “one network management” city operation management model. The system promotes the flat organization and intelligent refinement of Shanghai’s governance and has effectively improved the rapid collaborative processing capacity of city public services and events.
Microsoft Research Asia Collaborates With Orient Overseas Container Line (OOCL) to Develop Coopetitive Multi-Agent Reinforcement Learning, Achieving Whole Optimization of Shipping Network Operations.
Microsoft Research Asia explored a new set of solutions by applying the latest AI technologies such as deep learning and reinforcement learning in cooperation with OOCL to build a Coopetitive Multi-Agent Reinforcement Learning system. Each port and vessel is modelled into an agent to partition complex problems and solve them in a decentralized way. An efficient communication mechanism was established between various agents to promote cooperation. The system coordinated the distribution and transfer of benefits between agents, while at the same time coordinating competition between them, ultimately achieving the goal of global optimization. In 2019, OOCL officially launched this artificial intelligence solution in part of its shipping business networks, bringing significant improvements to business efficiency.
Cardinal Operations PonyPlus Intelligent Transportation Platform Helps an International Beer Giant Reduce Urban Transportation Costs by 5%
Cardinal Operations technologies helped an international beer giant manage expanding logistics and delivery scale. In its customer-oriented distribution, the company faced transportation scheduling problems such as irrational vehicle resource allocation, long working shifts, and out-of-order delivery, which made it difficult to support rapid business development. PonyPlus is a smart transportation platform specifically targeted at logistics optimization by Cardinal Operations. It provided a full chain optimization service to the beer giant. The average number of vehicles was reduced by nearly 8%, and the total mileage of vehicles shortened by about 10%. Both the number of stores and the number of delivery boxes increased by more than 50%. On the whole, the company has saved 5% in transportation costs, showing a significant optimization effect.
Recurrent AI-Developed Conversation Data Based AI Sales Middle Platform Helps Koolearn Online and Other Companies Improve Customer Conversion
Human-computer interaction and human-to-human interaction are two main types of interaction between enterprises and customers. Clicks, swipes and other behavioural data generated by the former have been fully utilized, but the value of the dialogue data generated by the latter has not been fully exploited. Based on the original NLP underlying technology XLNet, Recurrent AI created an AI sales middle platform based on conversation data, helping Koolearn Online, ZhongAn Insurance and other companies with medium and large telemarketing or customer service centres to perform better quality-control from dialogue data, analyze customer voice and differences in salesperson capabilities to forecast order succeed rate, ultimately helping companies improve customer conversion.
AHI-Fintech Assists China Merchants Bank to Carry out POS Anti-Fraud Model Project
The risk of oversea POS card transactions has always been prominent, and the recent pattern of identity theft has shown a trend of shifting from skimming to data leaks and large-scale credit card fraud. It is particularly important for banks to quickly identify cases of information leakage, investigate the cards involved in a timely manner, and prevent theft and loss at the transaction end. The project mainly involves: 1. Real-time identification of fraud transactions on oversea POS card and blocking the transactions; 2. Investigation of domestic and overseas information leakage points (including offline skimming and online information leakage). Targeting these two scenarios, China Merchants Bank’s retail finance headquarters and Ahi Fintech jointly built two models to control the risks of terminal POS overseas fraud and source bank card information skimming.
Beijing Hyundai Builds “Sales + Customer Service” Full-Scenario Intelligent Centre With Ronglian AI Intelligent Voice Robot
Beijing Hyundai Motor used Ronglian’s AI intelligent voice robot to achieve full intelligence in customer service and marketing systems. The system replaces and assists human agents to implement intelligent voice consultation, intelligent outbound calls, and multiple rounds of dialogue. Additionally, by achieving deep analysis of massive data and effectively screening out prospective customers, the system can significantly reduce operating costs, greatly improve conversion rate and customer service satisfaction, and fully empower the future intelligent business development. The project was launched in February 2019. The intelligent voice robot was able to eliminate 79% of invalid users, and the conversion rate of Class A customers was 66%. The sales and return visits that previously required 50 agents to work 30 days now require only 20 robots to work 14 days. The average weekly replacement rate for robot return visits is more than 20% for New Contract Return Visit and over 35% for Former Employee Return Visit, and customer satisfaction has increased by 30%.
Guotai Junan’s Artificial Intelligence Platform Powered by Transwarp Sophon
In 2019, Transwarp Artificial Intelligence Platform Sophon helped Guotai Junan Securities to build an artificial intelligence platform based on big data, machine learning and container cloud technologies. The platform achieved a leapfrogging of information technology to Fintech, applying artificial intelligence to the entire life cycle of synthetic finance, and created an intelligent digital financial platform while, supporting business development and improving the efficiency of research and services through informatization and electronization. A total of 99% of requests were responded to within 50ms. The system also supported real-time detection of millions of online users, and the efficiency of auto prevention and control rules is more than 3 times quicker than previous manual discovery. Guotai Junan’s Artificial Intelligence Platform powered by Transwarp Sophon led business innovation and value creation through digitization and intelligentization.
Nearly 200 Commercial Stores in Daxing Airport Adopt AI-Beestore, Aibee’s Offline Retail Digital Solution, to Promote Airport Non-Aerospace Revenue
Nearly 200 commercial stores in Daxing Airport have adopted AI-BeeStore, an offline retail digital solution provided by Aibee to help airport businesses achieve a comprehensive digital upgrade on offline retail. In the past, the evaluation of airport store operations relied on the observation, statistics, and inquiries by salesclerks. This required a lot of human resources and time, and it was still difficult to output timely, comprehensive and systematic data comparable to e-commerce platforms. The AI-BeeStore-based airport retail format upgrades to digital operations, making offline operations and precise retailing possible, effectively transferring the high consumption potential of airport passengers, improving business operation efficiency, and boosting non-airline revenue.
Anheuser-Busch InBev Recovers Huge Order Bonuses Through Guandata AI Inspection Program, Achieving a 32x ROI
POCP is an inspection program for orders within AB InBev. As a top brewing manufacturer, AB InBev has hundreds of thousands of orders of varying sizes each month. There are dealers who reported fake numbers and amounts of orders to obtain high discounts and bonuses or disrupted the market by returning discounts to the market. With limited numbers of inspectors, it is an extremely difficult task to inspect and recover illegal orders nationwide. Through a high degree of mutual trust and cooperation, Guandata and Anheuser-Busch InBev used machine learning and data empowerment to implement a customized order risk assessment system for Budweiser. By mid-2019, the ROI of the POCP project was 32x and it was enabling huge AB InBev order bonuses.
Leyanbot Helps a Garment Company’s E-Commerce Division to Increase Customer Stickiness and Store Sales Conversion
Since the application of Leyanbot two years ago, the e-commerce division of a garment company with annual sales exceeding CN¥1 billion greatly improved the efficiency of its online customer service reception rate, with a daily processing capacity of 1700-1800 orders per person. This allowed 70% of the company work to focus on sales and after-sale services to provide users with more detailed and personal care to improve user stickiness and conversion of online store sales. Since the deployment, the average response time of the company’s human customer service has improved from 120 seconds in 2017 to 40 seconds in 2018, and to 20 seconds in 2019, leading among online stores of the same volume. At the same time, without increasing the number of customer service staff, division performance in 2019 increased by 20% compared to 2018, and per capita reception efficiency improved significantly.
Yunxi Helps BESTORE to Reach All Channel Contacts and Obtain Member and Corporate Data
BESTORE became the industry’s first snack brand to introduce a middle platform structure after their new generation member middle platform officially launched in September 2019. This helped the company learn consumer spending habits and taste preferences by obtaining data on 80 million members. BESTORE continues to rank first in snack brand reputation in China’s annual Double 12 shopping event. The company has also gained insights and implemented precise reach-outs to customers who have a strong willingness to repurchase through big data analysis.
VideoPlusPlus Launches AI Video Ads on Popular TV Shows for New Automobile Retailing Platform Tangeche
Targeting the branding needs of auto leasing platform Tangeche, VideoPlusPlus AI Scene Marketing Platform selects the best times to run the company’s car rental ads based on seven genre labels covering modern, costume dramas, etc. It can also recognize actors’ lines and car-related scenes to trigger appropriate ads and ensure improved and timely exposure. The company aims to achieve its marketing goals by naturally attracting user attention from a specific situation, and stimulating user imagination and subconscious desire through the products. The ad program ran from June 10th to June 18th, 2019, during the annual 618 mid-year mega sale period, and Tangeche successfully stood out in the brand war with a maximum CTR of 2.16%.
Laiye Technology Developed a Knowledge Base of Maternal and Infant Care for Yili, and Created a Smart Customer Service “Xiao Yi” to Upgrade Customer Consultation Experience
Laiye Technology has created a smart customer service robot “Xiao Yi” that research suggests is more considerate than human customer service representatives. The bot was developed based on the characteristics of Yili Group’s customer service business scene and the actual pain points encountered. The core technology covers natural language processing (NLP), intelligent dialogue interaction, etc. and allows Xiao Yi to self-complete multiple rounds of dialogue. The robot is used in smart childcare, and its human-machine collaboration skills can also improve professionalism and efficiency in customer service. An AI customer service system is connected with the enterprise CRM system on user points, gift redemption, etc. Xiao Yi is a leader in comprehensive ability in the customer service robot industry.
Japanese Fashion Giant Uniqlo Uses Mujin Industrial Robots to Pack Clothing, and the Packaging Production Line Is Nearly Fully Automated
Uniqlo currently produces 1.3 billion pieces of clothing each year and has more than 3,500 stores in 26 countries. The huge production pressures and high labour costs have pushed this fast retail company to undertake a fully automated transformation. Clothing products need lightweight and automated processing to avoid damage. Mujin smart robots use 3D imaging lenses to scan, fold and pack clothing products. Mujin co-founder and CEO Issei Takino said that in October 2018, 90% of Uniqlo’s warehouses in Tokyo were using intelligent robots. In December 2019, Uniqlo announced cooperation with industrial robot startup Mujin to fully automate all the production lines with intelligent packaging robots.
Zhihu Question Routing Recommendation System Ensures Efficient Question Distribution and Answer Rate
Knowledge sharing platform Zhihu uses a question-and-answer mechanism to achieve cost reduction and efficiency gains in knowledge acquisition. The Question Routing System is essential in promoting content production under a question-and-answer mechanism. The system uses AI technology, machine learning and deep learning algorithms to turn the “problem finding right creators” scenario into an AI recommended search application scenario, which builds a smooth and efficient bridge between knowledge seekers and knowledge output parties and speeds up the transfer of knowledge.
In response to the large number of questions on Zhihu, the Question Routing Recommendation System can distribute 100,000+ questions every day, and ensure that the answer rate reaches 70% or more within 3 days after questions are asked. The system accurately recommends questions and more than 200,000 answers are generated daily. The system has been running steadily for 2 years, attracting a total of 36 million answers and adding 8 million creators to the community.
Alibaba Digital Media & Entertainment Group Provides Artificial Intelligence Platform for Automatic Extraction and Editing of Video Materials, Empowering Media Creators
Problems such as how to generate the most suitable cover image or images or even animated images from a video or how to automatically produce rich content trailers for long videos are challenges in the drive to provide consumers with more exciting and rich entertainment content. Youku Cognitive and Intelligent Lab of Alibaba Group constructed a set of artificial intelligence platforms for automatic extraction and editing of video materials, which can complete tasks like the intelligent understanding of video structure, automatic production of video cover image, automatic editing of video highlights, horizontal to vertical versions of videos, etc. The system is not only applied to Alibaba Digital Media & Entertainment Group’s business such as Youku, but also provides technology grants to most national social media creators through Alibaba Cloud.
DGene’s Interactive 3D Video Provides Shanghai Media Group 2019 World Wushu Championships With Full Angle Live Broadcasting
At the fifteenth World Wushu Championships in October 2019, DGene provided interactive live broadcast technical support for the SMG Great Sports channel with Interactive 3D Video technology to achieve a real-time interactive viewing experience. DGene’s Interactive 3D Video uses 72 high-precision cameras to accurately capture pictures of the players in motion in milliseconds, and has 360° multi-camera smooth and real-time switching capability. Viewers can watch on regular TVs or from any angle on mobile phones. The solution’s strong points are its speed, interactive free viewing angle and light installation.
Moviebook Provides Technical Support for a TV Program Jointly Produced by Guangming Online and zgzx.com
In 2019, Guangming Online and Zgzx.com jointly designed an online program with holographic demonstrations showing how members of the Chinese People’s Political Consultative Conference perform their duties. Moviebook provided this show with exclusive technical support through intelligent image production technology. It has developed an intelligent image production system based on key technologies such as computer vision and 3D reconstruction. Based on complete engineering services, this system conducts systematic and large-scale coupled processing and analysis of multimedia from four aspects: modularization, componentization, standardization and automation. Through intelligent tracking algorithms, the system comprehensively processes models such as regions, features, and deformations, and carries out pixel-level multi-point tracking analysis on streaming media.
The Kuaishou “Livestream Studio Voice Assistant Xiaokuai” AI-Powered Voice Assistant Enables Live Streaming and Control in Noisy Environments Even on Low-End Mobile Phones
It is more difficult to build a voice assistant for live broadcasting applications than for tasks handled by ordinary mobile phone voice assistants and smart speakers. Noisy live broadcast environments and limited computing resources create higher technical requirements. The “Xiaokuai” voice robot uses voice wake-up, voice recognition, natural language understanding, dialogue management, speech synthesis and other technologies. The real time voice assistant can respond accurately in loud environments and will work on even low-end mobile devices.
Thunisoft Legal AI Helps to Build an Intelligent Criminal Investigation Auxiliary Case Handling System for Huangshi People’s Procuratorate of Hubei Province
Thunisoft Legal AI built an intelligent criminal investigation auxiliary case handling system for Huangshi People’s Procuratorate of Hubei Province. The system supports complex case handling, short case expediting and expedited procedures, covering the whole process from approving arrests to prosecution. It also has multi-functional modules such as case review, sentencing research and judgment, similar cases push, legal think tank, document writing, litigation supervision and smart court appearance. It can not only efficiently handle simple cases of “one person, one thing”, but also has “multiple people, many things” capability.
The system can be used for 22 common crimes, covering 80% of criminal cases. It has provided review guidelines for more than 100 items, and can detect illegal or defective evidence items. It has a dataset of 10,552 scenarios and 638 evidence types to assist prosecutors in review and confirmation and can automatically generate case analysis maps to show the full picture of case evidence.
*Note: The selected application cases are sorted according to their smart field
30 Most Innovative AI Solutions
Baidu Open Source Deep Learning Platform Paddlepaddle
Baidu PaddlePaddle is China’s first and currently only open source industrial-level deep learning platform. It has leading technologies such as a core framework for convenient development, support for ultra-large-scale deep learning model training, a high-performance inference engine deployed on multiple terminals and multiple platforms, and an industrial-grade open source model library.
In April 2019, PaddlePaddle released its full platform structure along with 11 new features and services. In November, it released 21 new product directions and important upgrades. It has further improved its deep learning model development capabilities, training capabilities, prediction and deployment capabilities, and is shoulder to shoulder with international mainstream frameworks such as TensorFlow and PyTorch. PaddlePaddle has served more than 1.5 million developers and has more than 65,000 enterprise users. It has released 169,000 models on customized training platforms, and the number of models has shown a significant growth trend. It is widely used in various industries such as agriculture and the service industry.
Oneflow Technology Ltd. Developed a Distributed Deep Learning Training Framework Oneflow
OneFlow, a novel distributed deep learning framework developed by OneFlow Technology Ltd., is specifically designed for large models, large data, and large-scale streaming computing. So far, it has reached the industrial application level and will soon be open sourced. In 2019 it was deployed for a few representative case customers, including leading Internet companies and core institutions such as universities, research institutes, artificial intelligence parks and government services. The cases mainly cover massive image recognition, large-scale applications of natural language processing, large-scale applications of advertising and recommendation systems, and end-to-end turn-key machine learning platforms.
Huawei Cloud Ascend AI Computing Solution
The Huawei Cloud Ascend AI computing solution was released in December 2019. Based on a wide range of Ascend-powered cloud services, this solution features an open development and service framework, provides rapid integration capabilities, shares standardized components, and achieves efficient tuning. The computing power is twice as powerful at the same cost. Device-edge-cloud AI architecture embeds AI computing in the the entire network. The full-stack capabilities are developed by Huawei and made available externally to assist in overcoming any surge. Additional computing power is provided for content review, beauty/image super resolution, face recognition, and content recommendation. The solution is already being used by 10+ top-level industrial customers.
Beagle Data AI Enabled Intelligent Checking System for Abnormal Trading of Financial Securities
The main technological innovation of Beagle’s intelligent search system is establishing an abnormal account automatic identification model to identify hidden behaviour patterns and characteristics that differ from ordinary accounts using artificial intelligence algorithms. Beagle Data uses machine learning to collect traits of brokers’ abnormal transactions into an inference-side service, which comprehensively improves the risk-auditing ability of the entire industry. Organizations throughout the industry are able to utilize industrial information to extract key features without the need to access sensitive data from the regulatory authorities. At present, the Beagle Data Intelligent Checking System has been implemented in many securities companies.
Magnum Research B2B AQUMON’s Solution
Focusing on the digital transformation needs of financial institutions under tightening regulatory environments and market transformation trends, in 2019 Magnum Research introduced the B2B AQUMON Solution to domestic and foreign financial institutions, with algorithms as the core driver, IT as the technological innovation, and financial market theory as the underlying logic. The solution provides quantitative asset allocation strategies based on ETFs, mutual fund, stock bonds, etc. The B2B AQUMON Solution is based on a quantitative model that scientifically analyzes asset data, completes portfolio optimization and construction through IT technology, and uses algorithms to monitor the market dynamically in real time. Magnum Research has cooperated on the B2B AQUMON Solution with more than 60 global financial institutions.
Mcfly Digital Agriculture Full Life Circle Solution Based on Vision&Spectrum Technique
This program innovatively realized field scale monitoring and analyzing of key agricultural indicators (KAI). It provides a high-efficiency total solution for agricultural decision-making, with digital and scientific management of agricultural activities. Through variable plant-protection services, cost effectiveness and planting efficiency are both promoted. Tracking the planting input-output ratios through the whole process enables a deterministic income increasing solution concerning both spatial scale and farming time. The solution follows the general trend of gradually increasing farmland intensification caused by the transfer of agricultural land, and aims at achieving the measurability, predictability, cost-effectiveness, and high-efficiency of crop planting for both large scale and family agriculture. The overall solution is mature and stable, it has more than 3,000 farming customers across China with an average farmland size of over 200 mu (15 ha).
Lepu AI-ECG Platform
The AI-ECG Platform is the first domestic artificial intelligence medical technology for the automatic analysis and diagnosis of electrocardiograms. Its auxiliary diagnostic capabilities focus on arrhythmia, including premature beats, tachycardia, escape beats, conduction blocks, and atrial fibrillation. Compared with traditional ECG analysis and diagnosis methods, it has the advantages in accuracy, efficiency, and stability. The comprehensive accuracy rate is 95.2%, which is comparable to the level of professional ECG doctors. It obtained US FDA approval and EU CE certification in 2018. In 2019, application scenarios covered clinical routine static ECG analysis, dynamic ECG real-time monitoring and early warning, bedside monitoring, home monitoring, exercise ECG monitoring, and many others.
BeiShen Healthcare Mobile Phone Software Biliscan for Neonatal Jaundice Screening
Jaundice is one of the first health problems faced by newborns, with an incidence of 50% -80%. Severe conditions can cause bilirubin encephalopathy, mental retardation, deafness, irreversible brain damage, and even death. After 3 years of researching and developing the neonatal jaundice screening mobile phone software BiliScan, ShenZhen BeiShen Healthcare Co. can analyze skin images taken by mobile phone cameras using AI technologies to calculate the jaundice value, as well as to monitor jaundice after a newborn is discharged from the hospital in order to warn of any high risk.
In February 2019, BeiShen Healthcare obtained a medical device registration certificate issued by the Guangdong Provincial Drug Administration, the first medical device registration certificate based on mobile phone software (App) in China and the first artificial intelligence medical device to complete GCP clinical trials. ShenZhen BeiShen Healthcare partnered hospitals have spread across the country, serving millions of newborn families and 4 million registered users. The company worked closely with more than 30 provincial and district key hospitals in China to carry out big data research on neonatal jaundice.
AccuDiagnosis Technology Intelligent Full Quantitative AI Preoperation Planning System
The intelligent full-quantitative AI preoperative planning system developed by AccuDiagnosis Technology is designed to help with 3D preoperation planning in surgery, solving problems such as low efficiency and inaccuracy of image processing and reconstruction during an ongoing procedure and providing surgeons with visual, quantifiable and controllable precise surgical decision-making and auxiliary tools. The company founders includes the team of “surgical precision” pioneer Dong Jiahong and PhD and master students from Tsinghua University. Since the product launch in June 2019, the AI preoperative planning system has participated in auxiliary analysis and treatment on more than 1500 operations, saving doctors more than 5,000 hours of time.
Thorough Images Artificial Intelligence Assisted Gastrointestinal Pathological Diagnosis Platform
Pathology is the gold standard for medical diagnosis. Pathological reports are of vital importance for clinicians to further diagnose and treat diseases. Thorough Images released a deep learning-based assisted gastrointestinal pathological diagnosis platform in January 2019 which can improve the diagnostic efficiency of doctors and prevent missed diagnoses. After multi-center large-scale testing, the platform achieved a diagnostic sensitivity of 100% and a specificity above 85%. The platform has been deployed in dozens of top hospitals, including the Chinese PLA General Hospital; Cancer Hospital, Chinese Academy of Medical Sciences; the Peking Union Medical College Hospital, etc.
MYNT AI MYNT EYE STANDARD
MYNT AI’s active binocular-related technologies and products have made breakthroughs in technical fields such as positioning and navigation (VSLAM), three-dimensional measurement, and stereo recognition. Focusing on the implementation of cleaning robots, storage AGVs, volume measurement and other solutions and applications, MYNT AI is one of the very few companies in the world to achieve technology landing and large-scale revenue in the field of 3D visual perception. At present, MYNT AI has gained more than 1,600 clients worldwide including Tier 1 companies and top universities including STO Express, Megavii, Shanghai Jiao Tong University, and Waseda University. In 2019, 600 new enterprise customers were added. The new product Xiaomi Binocular Camera released during the period was highly precise, reaching the millimeter-level accuracy standard of some industrial cameras.
Elens Visual Analytics Platform – Zhice
Zhice is an one-stop visual analytics platform for intelligence analysis. Fuelled by the latest AI technologies, Zhice is able to integrate data and create knowledge graphs intelligently, and provides a rich variety of visual analytics modules including graph analysis, spatial-temporal analysis, transaction flow analysis, etc., offering a novel and efficiently tool set for intelligence analysis. Zhice is an integrated augmented-intelligence platform that demonstrates powerful impact in computer-assisted decision making, and has been helping numerous top clients in various markets.
Pensees Dimensional and Information-Based Solution to Public Security Control and Construction
Launched in mid-2019, the solution uses artificial intelligence, the Internet of Things, data intelligence, cloud computing, mobile app, etc. as core technologies to enable dimensional and informationized security prevention and control, whole real-scene command and dispatch; and comprehensively improve management, service and application levels. Pensees created a new model of public security prevention and control, of which “cloud security” and “intelligent security” are two important innovations. “Cloud security” refers to the construction of a public security prevention and control data cloud, which implements data integration, transmission, application, operation and maintenance based on VPN private networks. It is based on the cloud and relies on multidimensional IoT perception big data collision analysis technology to develop security control services. “Intelligent security” refers to deep applications based on data, providing multi-dimensional scene-based AI applications for users. By the end of 2019, the solution had served nearly 10 clients.
RegTech “Smoke Index” Big Data Monitoring and Early Warning Platform
The RegTech “Smoke Index” big data monitoring and early warning platform integrates multi-source heterogeneous big data resources and uses knowledge maps, artificial intelligence and other technologies to calculate a company’s “Smoke Index” in real time to measure its risk level. In 2019, the platform provided comprehensive services in scenarios such as illegal fund-raising risk monitoring, supervision of key financial fields, illegal financial advertising supervision and investigation and judgment of economic crime cases; forming a comprehensive financial risk prevention and control solution from central to local, administrative to criminal prevention and control. So far it has served more than 20 provinces and more than 100 financial regulatory agencies in China, playing an important role in preventing and defusing financial risks across the country.
Ant Financial Secure Collaborative Intelligence Platform
The Ant Financial Secure Collaborative Intelligence Platform is capable of aggregating information from multiple parties for machine learning in a scenario where multiple parties including data providers and platforms participate but do not trust each other, and ensure that the privacy of each party is not leaked and data is not misused. The Ant Financial Secure Collaborative Intelligence Platform aims to break the data isolation, enable data sharing and value transfer while ensuring data security and privacy protection, and reach the goal of invisible data availability. In 2019, The Secure Collaborative Intelligence Platform improved multi-party secure computing and trusted execution environment solutions, making them more complete, easy to use, and stable. Its joint modelling and joint decision-making procedures have been widely used under credit risk control, merchant risk control, advertising marketing and other scenarios.
DataVisor Unsupervised Machine Learning Engine-Driven Comprehensive Risk Control Management Solution dCube
As a comprehensive anti-fraud management platform (AI modeling platform), dCube can quickly and effectively model fraud features on the DataVisor Feature Platform, and put these models into production and decision-making phases. dCube combines transformational AI-powered technology with a streamlined workflow, including data integration and verification, feature engineering, model development, model verification and deployment, decision engine, case management, and in-depth analysis. DataVisor provides customized, diversified, and modular services based on clients’ multiple scenarios, helping data science teams and risk control teams collaborate efficiently on the same platform, and solving the problem of disconnection between technology and business in risk control management processes in traditional industries.
Dingxiang Affiliated Network Anti-gang Fraud Solutions
Aiming at the risks characterized by new-type organized fraud under the trend of digital transformation of banking retail business, Dingxiang independently researched and developed an affiliated network system which aims to help banks establish user data portraits of the entire life cycle of retail businesses. The system can tap multi-level risk relationships and nodes with rules, strategies and supervised learning to improve the reliability and accuracy of risk management systems in financial institutions. In 2019, Dingxiang conducted the first domestic best practice of anti-organized fraud in an affiliated network, helping a commercial bank to discover dozens of credit card cashing and fraud gangs with a risk exposure of more than CN¥1 billion. The solution effectively prevents and controls organized fraud risks such as cashing, fraudulent loans, abnormal fund collection and money laundering. It also helps commercial banks establish user portraits, independent risk control and anti-organized fraud systems.
NextData Full Stack Intelligent Content Recognition Engine Tianjing
In order to solve the recognizing difficulty and high cost of audio content, NextData launched the Tianjing full-stack intelligent content recognition engine based on mature artificial intelligence algorithms such as GAN, TDNN, LSTM, RNN, etc., and developed an Automatic Speech Recognition (ASR) model with high accuracy. The company then added NLP technology to build a complex speech recognition model system. The platform can accurately identify illegal or abusive content in politics, pornography, advertisements, and check audio data for content compliance. In 2019 the company launched national anthem recognition and pornographic voice detection functions while protecting the platform’s business security in all aspects. The solution has served hundreds of clients and aims to help domestic clients expand in overseas markets in the future.
Meituan “Next-Generation Store” Solution
Released in December 2019, Meituan’s “Next Generation Store” solution includes four major dimensions: digital operation, professional production, diversified marketing, and intelligent service. It helps catering companies solve the pain points of takeaway operations and improve the efficiency and quality of merchant services. Through online and offline deep integration, Meituan helps businesses iterate software and hardware upgrades and provides a complete set of solutions: In business, relying on big data to help businesses formulate store business models and operating strategies; in production, providing automated production equipment, meal pickup equipment, and design store layout; in marketing, providing a full range of services through membership access and resource integration. In services, the solution uses intelligent technology to improve store and product information management and optimize the delivery experience. Meituan Takeaway will help millions of merchants build “next-generation stores” to better serve hundreds of millions of users.
Bailian AI Intelligent Process Automation Solutions
Released in November 2019, the Bailian AI self-developed intelligent process automation solution uses advanced natural language processing, optical character recognition (OCR), knowledge graph and other technologies. It creatively supports scenarios that traditional RPA cannot support through the identification of interface elements and the structuring of core information. The company currently provides intelligent robotic process automation (Intelligent-RPA) services for more than 10 scenarios in five industries including international trade, logistics supply chain, IT, financial insurance, and corporate security services, including intelligent documents review, invoice review, form filling, material submission, order processing, data summary, etc. Basiloan is continuing to land in additional industries and fields.
Encoo RPA Enterprise-Level Robotic Process Automation Intelligent Platform
Encoo RPA is an enterprise-level software robotic process automation intelligent platform independently developed by Encoo. Based on leading automation technology, Encoo RPA can conduct repetitive, tedious, and time-consuming tasks across systems, reducing costs and increasing efficiency for enterprises, as well as quickly achieving business innovation and large-scale human-machine collaboration. The 2020 version of Encoo RPA, released in September 2019, embeds powerful artificial intelligence technologies such as computer vision, document understanding, and natural language processing to help companies build platform-level capabilities for automated and intelligent business. At present, Encoo has served hundreds of customers in finance, energy, logistics, telecommunications, manufacturing and other industries.
DeepZero One-stop Intelligent Enterprise Data Management Platform AlphaData™
The platform is a data-driven decision-making system created by DeepZero based on 11 years of technology and data capabilities for enterprise clients. At present, it has served 200+ large and medium-sized enterprises in the automotive, retail, FMCG, beauty, and home appliance industries. The company has established a unique two-tier structure of “data + business middle platforms”. In 2019, by adding business modules such as CDP module, leads scoring model, and prediction model, the company took the lead in landing data-driven decision-making processes in the automobile and pharmaceutical industries. At present, products are applied to business scenarios such as enterprise data asset activation, one-stop intelligent data management, real-time intelligent decision output, and intelligent leads scoring management, helping companies break data silos, improve the level of digital intelligence in business, and empowering enterprises to increase sales and efficiency.
BizSeer Technology BizSeer AIOps Platforms
In 2019, BizSeer technology created the “BizSeer AIOps platforms”, the first comprehensive analysis platform in China’s IT operations field that integrates multiple vertical algorithms across data sources. Large enterprises can find system failures and locate faults accurately in massive monitoring of data without tedious configuration, improving system availability and operation and maintenance efficiency. The platform uses a large number of BizSeer-developed technologies, and its related achievements have been published in dozens of international cutting-edge papers. The BizSeer AI-enabled IT operation platform has been purchased by more than 20 head financial and corporate clients. Due to the high universality of the platform, it is expected to be widely applied to operators, energy, manufacturing, aviation logistics and other enterprises outside of the financial field in 2020.
EOITEK Sharplook AIOps Platform
The new version of Sharplook AIOps Platform was released in July 2019. It creatively provides a diversified intelligent operation product system. Since its launch, it has received unanimous praise from dozens of clients, including benchmark customers such as China UnionPay, Bank of Communications, Shanghai Pudong Development Bank, Xiamen International Bank, and Bank of Ningbo. As a representative product, Sharplook Intelligent Event Analysis is driven by machine learning algorithms, it reduces the noise of a large number of alarm events and conducts correlation analysis which assists in root cause location and precipitates knowledge of fault eliminating so as to improve the operation efficiency of enterprises and reduce operation costs. In the future, Sharplook AIOps will also empower more medium-sized clients with its situation-based, out-of-the-box product capabilities.
Sogou Vocational Avatar
Created with a concept that combines Natural Interaction and Knowledge Computing, Sogou’s Vocational Avatar technology is the world’s first AI multimodal synthesis technology that can replicate human speeches, lip movements, facial expressions, and body movements. The technology represents a breakthrough in AI-enabled speech and video synthesis. Based on Sogou’s latest advances in speech recognition, natural language processing and deep learning technology, the Vocational Avatar technology can create lifelike resemblances of human beings with a small amount of video and audio data. The Vocation Avatar technology embodies industry-leading capabilities in natural human-machine interactions and intelligent Q&A.
FUXI-Lab IP Animation Editing Platform
Launched in March 2019, Fuxi-Lab IP animation editing platform uses cutting-edge AI technology to automatically produce lifelike and high-quality animations for virtual characters. It significantly improves video animation production efficiency and reduces production time from weeks to mins, which makes producing huge amounts of high quality video animation in a short time feasible. In the field of video games, the platform produced more than 10 hours of facial expression animation for “Justice Online” and “Revelation Mobile” as well as thousands of body motion animations for “Code Combat”, to build a more immersive experience for players. Meanwhile, in the field of virtual idols, this platform supported the virtual idol “Miss Q” of NetEase News in generating 34 episodes of short video animations with vivid expression and body motion in just a month, playing an essential role in operating their IP with a high quantity of content. The continuous growth of the animation market in the future will create more demand for virtual character animation production platforms.
Quicktron Third-Generation Smart Logistics Robot M60 (Zhuque)
In 2019, Quicktron’s third-generation Smart Logistics Robot M60 (Zhuque) was launched. Its overall efficiency improved more than 30% compared with the second-generation, with operating speed exceeding 2m/s. The robot uses laser SLAM navigation, texture navigation and hybrid navigation to automatically identify the indoor warehousing environment and plan the path and realize intelligent obstacle avoidance. Laser SLAM navigation doesn’t depend on the auxiliary navigation signs, and takes natural objects in the warehouse as positioning reference objects. Texture navigation uses the ground texture information captured by the robot during its movement to compare with the texture image in the self-built map so as to precisely position the robot. At present, robot has been applied in dozens of leading companies in aviation manufacturing, third-party logistics, electronics, clothing and footwear and other industries.
Whale Dynamic Autonomous Driving Multi-Sensor Fusion Solution
Released in June 2019, the Whale Dynamic Team’s own patented early fusion kit is the leading solution for vehicle roof-top sensing. It achieves multi-sensor time synchronization at microsecond level and unified power management. The built-in 10 Gigabit server can handle unified data transmission of multiple sensors, and also has CAN connection capability. The vehicle-level cable management and interface design have enabled developers to maximize the efficiency of rapid deployment of multiple sensors on vehicle roof-tops. This solution has been delivered to multiple autonomous driving teams in China, including the China Science Academy Shenzhen Institutes of Advanced Technology.
CyberInsight MachineInsight Solution
In 2019, CyberInsight released its MachineInsight solution which provides failure warning, machine tool life prediction, as well as spindle health management. The solution is designed with Cyber-Physical System architecture and realizes offline algorithm model training and online evaluation of machine tools health condition based on high-frequency vibration data and operational data combined with intelligent analysis. By the end of December 2019, CyberInsight has served two machine tool manufacturers by enhancing product competitiveness and service value. They also helped four manufacturers to convert production and operational data into economic benefits. Among those successful cases was a manufacturing company that announced in March 2019 the solution had helped reduce unexpected equipment downtime by 60%, increase product yield rate to 99.7%, and reduce overall costs by 16%. The manufacturing company was selected to the World Economic Forum “Lighthouse Factory”.
iReadyIT iReadyInsights Digital Twin Platform
The “iReadyInsights digital twin platform” is based on new generations of cutting-edge technologies such as big data, artificial intelligence, the Internet of Things and 5G; and enables innovative applications of digital twin technologies in the fields of environment, industry, and cities. In November 2019, the cloud-based business partner version of the platform was officially launched to develop environmental brain, industrial brain, and green city brain projects. The platform uses data lake architecture and technology to achieve second-level data collection and analysis of 100,000+ devices, with an annual data volume of one trillion. The innovatively built FAAS-based algorithm model platform collects industry knowledge such as equipment models, analysis models, mechanism models, AI models; and conducts scenario-based simulation and optimization. Some 5000+ physical templates and 500+ algorithm models have been deposited. At present, the platform has accumulated 5000+ material templates and 500+ algorithm models and is helping more than 30 Fortune 500 companies, parks and governments to achieve green smart services.
*Note: Selected application cases are sorted by industry
10 Most Noteworthy Funding Activities
Driverless Delivery Startup Nuro Gets $940 Million SoftBank Investment
NVIDIA to Acquire Mellanox for $6.9 Billion
UiPath nabs $568M at a $7B valuation to bring robotic process automation to the front office
Alibaba-backed facial recognition startup Megvii raises $750 million
Alibaba invests RMB 3.6 billion in China TransInfo, furthering smart mobility push
Salesforce to buy Tableau for $15.7B, betting on business data in deal for Seattle tech mainstay
Apple buys self-driving startup Drive.ai just days before it would have died
HPE advances its intelligent data platform with acquisition of MapR business assets
Shanghai artificial intelligence industry investment fund was established with an initial target scale of 10 billion
Intel Acquires Artificial Intelligence Chipmaker Habana Labs
Congratulations to all these companies and individuals for their outstanding achievements! Click here to read the complete Synced Machine Intelligence Awards 2019 Announcement.