《ADVANCE.AI:在业务中运人工智能的力量(英文版)(21页).pdf》由会员分享,可在线阅读,更多相关《ADVANCE.AI:在业务中运人工智能的力量(英文版)(21页).pdf(21页珍藏版)》请在三个皮匠报告上搜索。
1、1Leveraging the power of AIfor yourbusiness2022ContentsForeword:current AI commercialisation landscapeFour drivers for AI commercialisation1.Macro-environment2.Policy3.Data availability4.Technology improvementTwo paths for AI commercialisation1.AI commercialisation as a technologyAI+:AI-centric busi
2、ness opportunities led by engineers and scientists2.AI commercialisation in industries+AI Focus on value creation led by conventional companies in traditional industriesThe future challenges of AI commercialisation1.How to reduce data usage waste2.How to overcome algorithmic bias 3.How to establish
3、a standardised market access mechanismReference3446668811141618 1819Foreword:current AI commercialisation landscapePut forward as a concept at the Dartmouth Summer Research Project on Artificial Intelligence(DSRPAI)in 1956,and experiencing the following ever-evolving innovation,production and applic
4、ation for decades,Artificial intelligence(AI)has been already beyond abstraction but a prevalent technology that is transforming peoples ordinary life.Moreover,with the rapid development of information technology and the popularisation of the Internet,AI technology has ushered in the third high-spee
5、d growth marked by the deep-learning model proposal in 2006.Research and development,production and application,commercialisation and industrialisation,have been the there main stages of the growth of AI as new technology.Nevertheless,they are not in a constant linear way,as the three steps tend to
6、be performed simultaneously or repeatedly during the whole process.PwC,the professional services firm,estimates that AI could add as much as$15.7 trillion to the global economy by 2030.Nowadays,the basic technology of AI is gradually mature to reach the commercial conditions so that it can play its
7、value in a wider range of scenarios,which is the premise of its commercialisation.In addition,as the Internet industry has entered a reshuffle period in recent years,the capital market has shown more rational investment in the arena of AI.Therefore,projects with mature technology and strong commerci
8、al landing ability continue to attract the attention of capital,which to some extent promotes the industry from the early stage of general emphasis on technological advantages to more emphasis on products,solutions and other commercial capabilities of the development process.Meanwhile,the outbreak o
9、f the COVID-19 pandemic worldwide also provided an opportunity for AI solutions to prove their necessities and advantages when in inevitable limited conditions in various industries.34Social-economic changesAs the demographic dividend disappears and economic growth slows down,operating costs become
10、an increasingly important factor for most industries to consider.Using new technology to liberate the human labour force and achieve cost reduction and 1.Macro-environmentFour drivers for AI commercialisationMacro-environment,policy,data,and technology are the main driving forces for commercialising
11、 AI technology at the present stage.Like every technological revolution in history,the application value of AI in saving the labour force and improving production efficiency is also the focus of the market.In specific scenarios,AI provides new solutions for enterprises to reduce costs,boost efficien
12、cy,and provide new products to meet the needs of their consumers in the new era.Thus,the significant business opportunities and market demands are also a vital driving force for AI technologies to be applied.15Capital promotionCurrent industry generally pays attention to technology application abili
13、ty.This makes technology companies that find the landing scene usually have stronger viability and self-hematopoiesis ability and become more attractive to the capital.Market competitionWith a growing number of giants and startups coming in,the competition in the AI industry is intensifying.Differen
14、t from general consumable products,AI-enabled products and services are more likely to form stable and long-term cooperative relationships it makes gaining the trust of customers as the first-mover advantage more vital.Therefore,AI-based technology companies are scrambling to step out of the pace of
15、 commercialisation to consider seizing the market.23efficiency has become a major appeal of enterprises,which has also brought new opportunities for the commercialisation of AI and become a factor in promoting its landing.6The expansion of data presents new challenges to algorithm design and computi
16、ng power.However,those new challenges also push the commercialisation of AI technology.Advantages in data,such as large amounts of databases,the rich diversity of data,and more open access to data usage,will significantly accelerate commercialisation.A large amount of data provides basic materials f
17、or the training of algorithms and AI application.However,it brings the pressure of data processing at the same time.Nevertheless,it also drives the market to introduce new technologies and methods to further data value exploration.AI is just such a kind of new technology which is interdependent and
18、mutually promoted with data.3.Data availability 4.Technology improvement2.PolicyAI plays an important role in the global economy from the overall strategic layout of countries.Major economies have focused on the driving role of AI in the process of global economic growth and transformation.As a resu
19、lt,they have issued guidance documents to guide and promote the healthy development of the AI industryfrom the national strategic level.In terms of algorithm design,the deep-learning algorithm based on a multi-layer neural network,which is now the mainstream application,constantly strengthens the ma
20、chines ability to self-conclude object features from the massive database,as well as the ability to extract,describe and restore multi-layer features of the latest innovations.As a result,the recognition accuracy of deep learning-based machine vision,speech representation,biometrics and other AI tec
21、hnologies will be constantly improved so that practical problems can be solved in a wider range of scenarios.This is the most direct condition to promote the commercialisation of AI.In terms of computing power,the continuous innovation of graphics processing units(GPUs),field-programmable gate array
22、s(FPGAs),and application-specific integrated circuits(ASICs)and other AI chips has improved the overall computing power.Furthermore,in the face of massive data and complex scenes,the algorithm training and landing application of AI has gained more powerful patent support so that the results can be o
23、btained more quickly and accurately.This is an important push for the commercialisation of AI in terms of both technical implementation and user experience.NOTE:ASICs specific to AI uses go by many names,such as tensor processing units(TPUs),neural processing units(NPUs),and intelligence processing
24、units(IPUs).78Two paths for AI commercialisation1.AI commercialisation as a technologyAI+:AI-centric business opportunities led by engineers and scientistsThe improvement of AI will make intelligent systems even more brilliant.The following examples are spotlights,and topical issues of current AI co
25、mmercialisation that will also play main roles in future Innovation in these aspects will unlock more significant value of AI in the next step.The integration of AI,big data,and cloud computing realises AI models accessing information from the cloud and train themselves from the cloud while applying
26、 new insights into the cloud.This fusion significantly improves the calculation power and the capability of treating large amounts of data and intelligence.Extended reality,or XR,like AI,is one of the disruptive tools changing the way that we innovate and interact,and it will be a popular and powerf
27、ul tool in the way that the enterprise worksAI+Cloud computingAI+Extended realityXRCurrently,the two major types of XR are virtual reality and augmented reality.XR refers to the spectrum of experiences that blurs the line between the real world and the simulated world.The technology immerses the use
28、rs through visuals,audio,and potentially olfactory and haptic cues.When XR appears alongside AI and machine learning,AI-powered extended reality experiences could open the door to a new world of work and innovation although either AI or XR has its own value as separate components they can create exc
29、iting new opportunities and become even more impressive when combined.For instance,with AI+XR solutions,teams could work together in a virtual environment on building new products and prototypes without wasting resources.Meanwhile,thanks to AI technology,it could be possible to determine which solut
30、ions are more effective by analysing previous tests and data.Intelligent immersive experiences will be essential to improve worker productivity and the customer experience in the future.In addition,the rise of the global COVID-19 pandemic in 2020 drove even greater demand for these new solutions.910
31、The Internet of Things(IoT)describes the network of physical objects things that are embedded with sensors,software,and other technologies for the purpose of connecting and exchanging data with other devices and systems over the Internet.With more than 7 billion connected IoT devices today,IoT and d
32、ata statistics are staggering,to the point of appearing fantastical.Experts are expecting this number to grow to 22 billion by 2025.However,realising the future and full potential of IoT devices will require an investment in new technologies.The convergence of AI and IoT-AIoT-can redefine the way of
33、 economies,industries,and businesses.Why do AI and IoT blended matter the technology commercialisation?Generally,all IoT related services inevitably follow five basic steps:create,communicate,aggregate,analyse,and act,and the value of the act depends on how is the step of analyse going.Thus it can b
34、e seen that the precise value of IoT is determined at its analysis step.When AI technologies integrate into IoT platforms,common use cases such as root cause analysis,predictive maintenance of machinery or outlier detection will impact almost every detailed step of implementing an IoT service,especi
35、ally the step of analyse.Besides,advances in neural networks have brought natural-language processing(NLP)to IoT devices(such as iPhones Siri)and made them appealing,affordable,and viable for broader use.Hence,while IoT deals with devices interacting using the Internet,AI makes the devices learn fro
36、m their data and experience.This is where AI technology portrays a crucial role.AI-powered IoT creates intelligent devices that will generate new business opportunities for AI+Internet of things11Automotive+AI2.AI commercialisation in industries+AI Focus on value creation led by conventional compani
37、es in traditional industriesWith an AI-powered software system in vehicles,cars will understand their immediate environment and navigate it safely.As a result,autonomous vehicles can save time,limit energy consumption,and dramatically reduce the more than 1.25 million deaths attributed to road traff
38、ic accidents each year.it is a vital improvement for traffic security.the future digital world.AIoT istransforming the way people interact withthe devices at home,at work and throughout thecities,unleashing the power of data better and faster than ever.AI has been applied in various industries,trans
39、forming the way people live,work,travel,and do business,which contributes greatly to the increase in global economic growth and productivity.1212Arts and culture+AIFinance+AIBioScience+AIEducation+AIAI is changing the arts,enriching peoples daily experiences,preserving culture and making art more ac
40、cessible to those unable to visit a gallery or historic site for themselves.It is also a creative force able to compose music,write novels and paint pictures.The banking and financing fields have been beneficial from AI adoption significantly.Traditional financial institutions cooperate with technol
41、ogy companies to promote AI adoption and change the rules of the financial service industry.As a result,they improve commercial efficiency across the sector:from remote customer AI empowers researchers to conduct experiments more efficiently and helps them find more vegetation and animal patterns th
42、ey never knew existed.AI is taking bioScience to the next level making more breakthrough ecological discoveries,exploring life expectancies,and creating a better quality of life globally.AI is also being used across the educational landscape to develop new tactics for helping people learn.For exampl
43、e,due to the lockdown requirement during the COVID-19 pandemic,many students can not join classroom classes,and AI-powered e-learning became the solution,and students can have classes even when at home.13Retail+AIHealthcare+AIIn the retail industry,AI adoption has shifted from the individual shop to
44、 the aggregate industry.Traditional retail enterprises form partnerships with technology startups to build application scenarios around people,goods,stores and chains.AI blooms in various retail links,the technologys commercial impact is also helping retailers with strategies for e-commerce and bric
45、k and mortar stores fragmented use scenarios finally have entered a large-scale experimental period.AI is revolutionising healthcare and also improving access to life-saving therapies.Whether its being used to discover genetic codes,to power surgical robots or even to maximise hospital efficiency,AI
46、 has been a boon to the healthcare industry.The potential for both AI and robotics in healthcare is vast.As a result,AI and robotics are increasingly a part of the national healthcare ecosystem in many countries.PwC has summarised eight ways that showcase how AIs transformation in healthcare is curr
47、ently underway,including AI adoption in the medical training and research,early medical detection,diagnosis and treatment,end of life care,assistance in keeping well,and medical decision making.onboarding to fraud detection and prevention,from digital banking to better customer advisory service.Nowa
48、days,AI application scenarios have been gradually expanding from focusing on transaction security to reforming the whole process of financial operation.Over the years,the rise of AI has fundamentally transformed the very meaning of ideas,innovation,and inventions.Nowadays,AI capability plays a vital
49、 role for each business entity and brings their future commerces a unique possibility to move forward and grow.As we witness businesses across industries undergo a dramatic and profound shift in the relative balance of intelligence power,there are still many challenges that need to be paid attention
50、 to or solved in the future.For example,how to operate AI safely?Whether it can be autonomous and controllable in some underlying technologies?Whether the regulatory restrictions on the usage of AI are equal and comprehensive?Besides,AIs impact on work is profound.Some occupations for some skills Th
51、e future challenges of AI commercialisation14will decline,while others grow and many changes as people work alongside ever-evolving and increasingly capable machines.Nevertheless,challenges are also new opportunities that AI adoption offers to various industries and will facilitate AIs commercialisa
52、tion to the next level for example,the demand for wide construction of AI infrastructure;the requirements of more personalised human-computer interaction interface to enhance user experience;needing to promote AI industrialisation by cultivating AI developer ecology and serve the development of all
53、walks of life.The following presents some common challenges that industries are facing in the AI transformation today and key concerns of AI commercialisation:1516While access to information is universal,what is not common is how each business uses that information for what purposes and goals,and wh
54、ether they can make the most of the value of the information.Although emerging technology levels the playing field to a degree for businesses across industries in their ability to access intelligence from the increasing digital data and information,many of them lack the necessary technical tools or
55、appropriate strategies to unlock the value this information can deliver.This challenge is apt to be more common in the manufacturing industrys AI transformation.On the one hand,the complexity and customisation requirements of AI solutions in the manufacturing arena are high,such as product quality i
56、nspection,sorting and 1.How to reduce data usage waste1719predictive maintenance it always needs to analyse the vast amount of data and information across industries.But on the other,on the road to developing Industry 4.0,some traditional database systems hit the bottleneck computing ability.If data
57、 analysis speed can not chase the rapid information changes it will result in the usage waste of a large number of time-effective industrial data.Over time,this mismatch will become more and more serious.18Algorithmic bias in AI is a pervasive problem.While eliminating bias in AI is hard to say easy
58、,but it is essential to know how to work actively to prevent it.Meanwhile,learning how to mitigate bias in AI systems stems from understanding the training data sets that are used to generate and evolve models.It is because AI systems can be biased based on who builds them,how they are developed,and
59、 how they are ultimately used.Algorithmic bias has damaging effects in many areas,especially in law enforcement and the financial service industry,where people who are underrepresented will be further marginalised and deprived of service.2.How to overcome algorithmic bias 19The application of AI is
60、developing rapidly;nevertheless,there is an urgent need to establish a standardised market access mechanism for AI products in this field and strengthen the building of the industrys database.For example,AI technology can help the medical and healthcare industry solve many livelihood problems caused
61、 by the shortage of medical resources and uneven distribution.However,as medical treatment is a highly regulated industry,whether AI can be widely adopted as expected will also depend on how medical and data regulatory standards are set during the product commercialisation process.3.How to establish
62、 a standardised market access mechanismIf algorithmic bias can not be addressed,it will lead to the amplification of human biases.As a report by the Financial Stability Board(FSB)says,AIs lack of auditability could lead to unintended consequences.A guiding principle of good usage of data in commerci
63、alisation should be companies ought to be able to explain to a customer how a data-related decision was made.Because of the nature of algorithmic,the technology needs to be handled appropriately no one would like to see models that develop bias that goes uncorrected.ReferenceThe History of Artificia
64、l Intelligence,Harvard University,The Graduate School of Arts and Sciences,Aug 2017The History of Deep Learning:Top Moments That Shaped the Technology,Built In,Apr 2020PwCs Global Artificial Intelligence Study:Sizing the prize,PwC,2017Accentures annual Technology report finds all five major trends i
65、n 2018 being data centric,Packt,Sep 2018 XR,Machine Learning&AI:The ThreePillarsof FutureEnterprise,XR Today,Dec 2020What is the Internet of Things(IoT),OracleConnecting applications,devices and AI/ML functions,CN GroupArtificial intelligence is changing these 9 industries,Business Insider,2017How t
66、o Reduce Bias in AI,Appen,Jul 2020What is algorithmic bias,TechTalks,Mar 2018When AI becomes too big to fail,Financial Times,2017 Take Full Advantage of Technology with AI Guardian,ADVANCE.AI,2020 1.2.3.4.5.6.7.8.9.10.11.12.2021ADVANCE.AI is a leading AI company that provides digital transformation,
67、fraud prevention,and process automation solutions for enterprise clients.A leader in Artificial Intelligence,risk management and digital lending solutions,it currently partners over 700+enterprise clients across banking,financial services,fintech,payment,retail and e-commerce sectors.ADVANCE.AI is p
68、art of Advance Intelligence Group,one of the largest independent technology startups based in Singapore.Founded in 2016,the Group has presence across South and Southeast Asia,Latin America and Greater China.The Group is backed by top tier investors SoftBank Vision Fund 2,Warburg Pincus,Northstar,Vision Plus Capital,Gaorong Capital,Pavilion Capital,GSR Ventures and Singapore-based global investor EDBI.About ADVANCE.AIwww.advance.aisalesadvance.ai