小程序
传感搜
传感圈

Unlocking the value of artificial intelligence and machine learning

2022-10-06
关注

In an era of accelerated digitalisation, artificial intelligence (AI) and machine learning (ML) have fast become part of the IT infrastructure of many businesses. Consequently, how these technologies are being used to derive meaningful insights from vast quantities of data is maturing rapidly.

Artificial intelligence
Artificial intelligence and machine learning are at the heart of solving key business problems. (Photo by cono0430/Shutterstock)

“Early on, when organisations didn’t have access to the computing power and zettabytes of data that they have today, AI was only springing up in pockets,” says Vaidya JR, SVP and global head of data and AI at IT transformation specialist Hexaware Technologies, “The approach then was to see what AI could do for a company, without truly identifying a well-defined problem. Data science solutions were just a shot in the dark.

“Organisations were struggling to put their data to effective use, which led to limited value generated and ineffectual business results,” he adds. “You can crunch any amount of data, and create numerous models; it only adds value if there is a significant impact on the business. But the current attitude has completely changed across industries, without exception.”

From being data-rich but insight-poor, businesses are putting vast amounts of data to work – sensor data, satellite imagery, web traffic, digital apps, video images, customer behaviour metrics and much more. They are in the process of automating and democratising AI and ML, but the attitude now is one of identifying business problems to solve before implementing these technologies. This marks a significant shift: AI and ML strategies are no longer driven by tech, but by strategic business objectives. 

Free White Paper

Using AI to Optimize IoT Data in Quality Control for a Global Heavy Vehicle Manufacturer

By Hexaware-Technologies

Enter your details to receive the free paper:

By downloading this whitepaper, you acknowledge that New Statesman Media Group may share your information with our white paper partners/sponsors who may contact you directly with information on their products and services.   Visit our privacy policy for more information about our services, how New Statesman Media Group may use, process and share your personal data, including information on your rights in respect of your personal data and how you can unsubscribe from future marketing communications.  Our services are intended for corporate subscribers and you warrant that the email address submitted is your corporate email address.  

Thank you. Please check your email to download the white paper.

“Enterprises are looking for use cases to drive business goals and produce results,” observes Vaidya. “AI has to derive efficiency gains at a reduced cost with actionable insights.

“Advanced analytics, AI and ML adoption have skyrocketed across organisations that are seeking to unlock value from data,” he adds. “The race has begun. Everybody is now thinking of democratising AI because they have humongous amounts of data, and it is not humanly possible to crunch all that data manually. So, automation and democratisation are a must.”

Data drives business

In the eyes of the Hexaware executive, this marks a wider corporate trend: in the age of digitalisation, all enterprises are becoming data companies. 

Today, automation features across the entire data value chain. The emergence of AI-driven analytics solutions has aided industries in creating greater business value. 

“The manufacturing industry, for instance, is fast emerging as completely data-driven,” remarks Vaidya. “We’re talking about digital twins, robots making robots, autonomous cars, and much more. The underlying aspect that is powering the entire possibility is data and AI.

“For example, when you have high rejection rates in your production unit, you need to deploy a parts failure prediction model,” he continues. “You must convert the business problem into a data science problem. What kind of data is required to iterate on the model and improve its accuracy and reliability? How should it be deployed in the real world? These are the areas where we step in and help our customers across industries figure things out.”

Even in professional services, Vaidya says, the linear models are long gone: “It is no longer about a consulting firm approaching the client, building a team, and delivering people-based solutions. Instead, they are turning towards self-service platforms. We’re looking at data platforms customers can subscribe to, that provide automation to the extent where customers only need to identify their problem and the necessary parameters, after which these self-service platforms process their requirements. This is applicable to numerous industries, like auditing, advisory, accounting, healthcare, insurance and more.”

“Data-driven customer experience is another area many of our customers are prioritising, using omni-channels for hyper-personalisation, and customising their journeys in real time,” he adds. “AI recommendation engines for personalisation are extensively used across industries to recommend a customer’s next purchase or next show to watch based on their past purchases or watch history.”

From the acquisition of data to the gleaning of insights, be it through embedding models or operationalising, automation is increasingly prevalent across the data value chain. Enterprises are increasingly turning to automated machine learning (AutoML). This process automates time-consuming, iterative tasks of ML model development, and also serves towards democratising machine learning. 

“In fact, AutoML will eventually reach a stage where if a business use case is given, it will identify options regarding how to convert that case into a typical data science problem, all the way up to operationalising AI,” predicts Vaidya.

Building ML models is only a part of the process. Operationalising them is where MLOps – a set of practices for reliably and efficiently deploying and maintaining ML models in production, transitioning the algorithms to production systems – comes into the picture. 

Artificial intelligence partnerships power

With AI strategy driven by the C-suite, rather than the IT department, third-party partners like Hexaware have a crucial role to play in helping customers deliver digital initiatives and build connected customer-centric digital enterprises.

“From our experience, digital enterprises are characterised by certain factors – hyper-personalisation, connected enterprises, customer-centric and advanced production methods,” says Vaidya. “We understand the transformations being witnessed in the market, compare their peers in the industry and then engage with our clients to co-create value. We also help organisations understand how transforming the digital helps transform the physical. Edge computing, PaaS, servitisation, IoT and IIoT – in the B2B segment – are only a few of the technologies that help make that connection.

“Where we step in is at the conversion stage. We help build digital enterprises, by converging technologies and deriving the best from them. This includes creating a data architecture roadmap, an AI Center of Excellence (CoE) roadmap, change management and guiding enterprises on how their consultants can be reskilled. We create and build a minimum viable product quickly that they can test through the entire value chain.”

Vaidya acknowledges that change is happening at an ever-increasing pace, but cautions that businesses must be careful not to rush. A foundation must first be set, instead of jumping on the AI bandwagon for the sake of it. “We start from scratch, preparing organisations for business change,” he explains. “Right from identifying the business challenge and use cases, to converting them into data science problems, to eventually building an agile data engineering organisation, which is a fundamental piece, since your model is only as good as the data it gets.” 

There are numerous adoption challenges that enterprises might face, right from excessive costs borne owing to inaccurate models, lack of skilled SMEs or even identifying the right data sets.

Ultimately, it boils down to understanding the business, defining the business challenge at hand, and converting it into a data science problem. This requires a team encompassing the right blend of skills, roles and responsibilities, coming together to create an ecosystem that generates value repeatedly and reliably.

Topics in this article: AI, Analytics, Hexaware Technologies, Machine Learning, sponsored

参考译文
解锁人工智能和机器学习的价值
在加速数字化的时代,人工智能(AI)和机器学习(ML)已迅速成为许多企业IT基础设施的一部分。因此,如何利用这些技术从大量数据中获得有意义的见解正在迅速成熟。IT转型专业公司Hexaware Technologies高级副总裁兼数据和人工智能全球主管Vaidya JR说:“早期,当企业没有今天拥有的计算能力和zetb的数据时,人工智能只是在小范围内涌现。当时的做法是看看人工智能能为公司做什么,而不是真正识别明确定义的问题。数据科学解决方案只是瞎猜。他补充称:“各组织都难以有效利用数据,这导致了有限的价值创造和无效的业务结果。”“你可以处理任意数量的数据,并创建多个模型;只有在对业务有重大影响时,它才会增加价值。但目前的态度在各行各业都完全改变了,无一例外。“从数据丰富但洞察不足,企业正在将大量数据投入工作——传感器数据、卫星图像、网络流量、数字应用程序、视频图像、客户行为指标等等。他们正处于AI和ML自动化和民主化的过程中,但现在的态度是在实施这些技术之前先确定要解决的业务问题。这标志着一个重大转变:AI和ML战略不再由技术驱动,而是由战略商业目标驱动。“企业正在寻找用例来驱动业务目标并产生结果,”Vaidya观察到。“人工智能必须以更低的成本和可操作的洞察获得效率收益。”在寻求从数据中释放价值的组织中,高级分析、人工智能和ML的采用已经迅速增长。”“比赛开始了。现在每个人都在考虑将人工智能民主化,因为他们拥有大量的数据,而人工处理所有这些数据是不可能的。因此,自动化和民主化是必须的。在这位Hexaware高管看来,这标志着一个更广泛的企业趋势:在数字化时代,所有企业都在成为数据公司。今天,自动化的特点跨越了整个数据价值链。人工智能驱动的分析解决方案的出现帮助行业创造了更大的商业价值。“例如,制造业正迅速崛起为完全由数据驱动的行业,”Vaidya说。“我们说的是数字双胞胎、机器人制造机器人、自动驾驶汽车等等。推动这一可能性的根本因素是数据和人工智能。他继续说:“例如,当你的生产单元有很高的废品率时,你需要部署一个零件失效预测模型。“您必须将业务问题转化为数据科学问题。在模型上迭代并提高其准确性和可靠性需要什么样的数据?如何在现实世界中部署它?这些都是我们介入并帮助各行各业客户解决问题的领域。Vaidya说,即使是在专业服务领域,线性模式也早已不复存在:“咨询公司不再需要接近客户、组建团队、提供以人为本的解决方案。”相反,他们正在转向自助服务平台。我们正在研究客户可以订阅的数据平台,这些平台提供自动化,客户只需要确定他们的问题和必要的参数,然后这些自助服务平台处理他们的需求。这适用于许多行业,如审计、咨询、会计、医疗保健、保险等。” 他补充称:“数据驱动的客户体验是我们许多客户优先考虑的另一个领域,他们使用全渠道实现超个性化,并实时定制他们的旅程。”“个性化的人工智能推荐引擎被广泛应用于各个行业,根据客户过去的购买或观看历史,推荐他们的下一个购买或下一个观看节目。”从数据的获取到见解的收集,无论是通过嵌入模型还是操作化,自动化在整个数据价值链中越来越普遍。企业正越来越多地转向自动机器学习(AutoML)。这个过程将耗时、迭代的ML模型开发任务自动化,也有助于机器学习民主化。Vaidya预测:“事实上,AutoML最终将达到这样一个阶段:如果给出了一个业务用例,它将确定如何将该用例转换为典型的数据科学问题的选项,一直到人工智能的操作化。”构建ML模型只是该过程的一部分。MLOps——一套用于在生产环境中可靠而有效地部署和维护ML模型的实践,将算法转换到生产系统中——就需要对它们进行操作。由于人工智能战略由高管部门而不是IT部门驱动,像Hexaware这样的第三方合作伙伴在帮助客户交付数字计划和建立以客户为中心的互联数字企业方面发挥着关键作用。Vaidya说:“根据我们的经验,数字企业具有某些特征——超个性化、企业互联、以客户为中心和先进的生产方法。”“我们了解市场正在发生的变化,与同行进行比较,然后与客户合作,共同创造价值。我们还帮助企业理解数字化是如何帮助变革实体的。边缘计算、PaaS、服务化、物联网和工业物联网——在B2B领域——只是帮助建立这种连接的少数技术。“我们的介入是在转换阶段。我们通过融合技术并从中汲取精华,帮助建立数字企业。这包括创建数据架构路线图、AI卓越中心(CoE)路线图、变更管理和指导企业如何提高顾问的技能。我们快速创建和构建一个最小可行的产品,他们可以通过整个价值链进行测试。Vaidya承认,变化正在以越来越快的速度发生,但警告说,企业必须小心,不要急于求成。首先要打好基础,而不是为了人工智能而盲目跟风。他解释说:“我们从零开始,让组织为业务变革做好准备。”“从确定业务挑战和用例,到将它们转化为数据科学问题,到最终建立一个灵活的数据工程组织,这是一个基本部分,因为你的模型只和它得到的数据一样好。”企业可能会面临许多采用挑战,包括由于不准确的模型、缺乏熟练的中小企业或甚至确定正确的数据集而承担的过高成本。最终,它归结为理解业务,定义手头的业务挑战,并将其转化为数据科学问题。这就需要一个拥有各种技能、角色和责任的团队,团结起来创造一个能够反复、可靠地产生价值的生态系统。
您觉得本篇内容如何
评分

评论

您需要登录才可以回复|注册

提交评论

techmonitor

这家伙很懒,什么描述也没留下

关注

点击进入下一篇

燧原科技打造一站式人工智能算力中心

提取码
复制提取码
点击跳转至百度网盘