The B2B Proposition: AI/Machine Learning Technologies Enable Global Industrial Transformation

Urgent demand for a new era of industrial automation based on the IIoT (Industrial Internet of Things) is now a primary driver for Digital Transformation (DT).  The benefits that will be realized are significant, and like any technology / digital operating model evolution there will be challenges towards optimizing these capabilities.  In regard to IIoT, there are currently limitations in terms of data scaling and speed of operation that impact its success. It’s no wonder that into this space has stepped ‘Machine Learning and Training’ and ‘Inferencing’, which together enable decision-making at scale and at speeds approaching real time. Collectively, this is called Artificial Intelligence or ‘AI’.

First coined in 1956 by John McCarthy at Dartmouth College, AI is the theory and development of using computer systems to perform tasks that typically require human intelligence. Examples of high-order human tasks include visual perception, decision-making and language translation. While the human brain effortlessly processes many of these functions simultaneously, computers have historically struggled to achieve human-like cognition.

Fast forward 55 years however, and especially in the last decade, enabling technologies behind AI such as computer science, chip advances, increased computing power, and the connectivity needed to move and process the abundance of communications in the modern world, have changed the narrative – AI no longer struggles in cognitive tasks.

A branch of AI increasingly applied in the B2B field is Machine Learning (ML), a set of technologies that features algorithms like a needle weaving fabric: stitching multiple pieces together to accelerate a computer’s ability to learn. ML does so by self-correcting decision-making and doubling down on successful strategies while cutting losing less-efficient models over time. It is a powerful Digital Transformation tool that brings all the potential of Industry 4.0 – IIoT on steroids – closer to reality.

Such technologies are very much global in nature, and the global industrial potential of advances within AI/ML is significant. Clearly, those nations that can get ahead in Artificial Intelligence and Machine Learning (AI/ML) are going to capture market share in the industrial evolution. But, where do we sit as a baseline? Who’s ahead? Who are the main contenders, and what are the key drivers in a move towards an Industry 4.0 future?

Recent research1 illustrated that while the US/EU have been leading the era of discovery, tech companies in Japan and South Korea have some of the highest number of AI patent filings2, while China is at the forefront of implementation1.  The US is focused in such key areas as investment in start-ups, research and development funding and chip design for AI systems, but Asia has the huge populations and corresponding industry base to facilitate the development and digital transformation of industrial automation through ML.

The high-level national ‘Made in China 2025’ strategy was founded upon this premise. When initially conceived and defined, the strategy was expected to deliver slower results than Industry 4.0, but AI and ML technology progress is definitively accelerating the transformation of industrial automation, with Industry 4.0 and its manipulation of vast datasets being an increasingly realistic goal.

So, while China may be grabbing headlines in the public AI space, what is the reality, as things stand? In Taiwan, the island’s ICT and semiconductor industries are a solid foundation for intelligent technology development. In other East Asian locations, the emphasis is on scaling applications in industrial and home robotics, self-driving cars, smart healthcare, smart manufacturing, and smart city projects, like the one planned in the foothills of Mount Fuji by one large automotive developer.

Related developments in China are robust and advanced, backed by a central and entrepreneurial determination to dominate the area globally. More specifically, the narrowing of the gap between IT and OT is a key factor in how imminent or successful Industry 4.0 is likely to be, both in China and elsewhere, and AI/ML is important here as industrial companies of various sizes seek to blend their operational and data programs with efficiency. Put simply, careful implementation of AI/ML will lead to improved operational effectiveness and faster processing, through data processing, analytics, and pattern/trend analysis. The beauty of AI/ML and Industry 4.0 is of course their applicability to most of the diverse industry types in China (and globally), underlining their potential and power.

China has seized on AI and announced ambitious goals. It will be a national imperative to establish a significant presence in the technology – and then move ahead to global leadership. The government has not been shy about its aggressive stance to drive prioritization and exponential leadership.

So, grabbing the ball and running with it seems to be what China is currently doing with AI/ML, as it looks ahead to the ultra-efficient production goals of Industry 4.0 with help from China’s established triumvirate of internet leaders, Baidu, Alibaba and Tencent, collectively known as the BAT. They have all been aggressive in venturing into AI technology. 

Online search specialist Baidu is involved in three major AI-related ventures: Apollo, an ambitious open-source project in autonomous driving, global in scope; DuerOS, a voice-enabled digital assistant; and Baidu ABC, a ‘cloud’ platform for businesses.

In September 2019, Alibaba announced an AI accelerator chip, the Hanguang 800. Built in a 12nm process, the Hanguang 800 contains 17 billion transistors and is capable of processing 78,563 images per second (IPS) and 500 IPS/W (images-per-second-per-Watt), when benchmarked in ResNet-50.

Tencent, well known as the intelligence behind its hugely popular and versatile WeChat social media platform and for being the world’s largest video game publisher, launched an AI Lab in China in 2016 and then in 2017 opened an AI R&D centre in the US, in Bellevue, Washington, to work on speech recognition and natural language processing (NLP).

It could be said that the key ingredient or building-block behind AI/ML is data, while connectivity is the ‘glue’ that takes the data in a meaningful and powerful direction. Without connectivity driving data, the successful push towards an Industry 4.0 landscape would be restricted.

As both a concept and a reality, Machine Learning (ML) offers hugely efficient economic modeling at both the global and local levels. Datasets can be built at speed and extremely large scale via ML, with algorithmic ‘training’ then yielding the ‘inferencing’ of insights in support of strategic decision-making. Advances in communications at very close to real time, as is the case with the rapid development and roll-out of 5G, then boost the breathtaking speed at which Artificial Intelligence (AI) can be applied.

These potential applications can be both granular and highly local, as is the case with the development of autonomous vehicles, and much broader in scope, financial technology (fintech), and Supply Chain Management (SCM). Innovative leaps forward in automation and productive efficiency are the goal.

In terms of scalability, AI as a Service (AIaaS) and ML as a Service (MLaaS) are already established, providing more accurate solutions to address customer demands compared to SaaS (Software as a Service). A combination of software and hardware advances will lead to increased computational power, learning potential and, crucially, data management, which in turn will enable AI/ML to be seamlessly scalable both in terms of depth and variety. China’s 5G push, for example, for advanced connectivity, lower latency and more data potential, along with its advances in areas such as quantum computing, give it a computational power advantage to enhance its AI/ML achievements.

Indeed, recent market analysis3 of the impact of AI/ML on SCM concluded that AI-enabled supply chains are up to 60% more effective with reduced risk and lower overall costs. Cloud-based AIaas for SCM will reach US$1.9B by 2025 globally. AI in SCM involving context-aware computing will reach US$1.3B by 2025 globally. AI SCM in edge computing for IoT-enabled solutions will reach US$3.2B by 2025 globally.

At Molex we see that the value of AI/ML could potentially grow exponentially. Not only will computers identify patterns, but they will also justify and explain decisions and make new suggestions. And these capabilities have great potential to recommend new product features while also predicting product demand prior to a production ramp to ensure improved customer experience in the process.

However, while AI/ML will improve the efficiency and accuracy of decision-making in B2B scenarios, it is important to realize that key factors such as data services, power consumption and training models are different from human intelligence, judgment, and input. In many cases, emotional and creative input are required more in B2B decision-making than in other areas.  Critical to this is enabling UI functionality that enables the proper focus and execution towards achieving the benefits AL/ML will bring forth.

AI/ML is at a global inflection point in gathering data and putting it to use at scale. AI/ML is the future. Nevertheless, the cloud presents latencies in communication, and increasingly AI/ML processing is migrating to the ‘edge’ – in cars for example, on smartphones and laptops and all manner of local devices and venues that require AI/ML at speed. And it is Industry 4.0, the IIoT transformed, that will require the speed of AI/ML at the edge, where a myriad of local devices must interconnect, inter-network, and inter-operate at speeds close to real time – without losing or destroying data packets.

This is simply because Industry 4.0 seeks ultra-efficiency in production through a constant process of data feedback derived from sensor arrays deployed at every point of the production process. This results in huge datasets that provide the basis for machine learning, followed by the inferencing of AI as it contributes to B2B decision-making.

It’s with the migration to the local, to the edge, to ubiquitous computing that the vendors and distributors of the electronic and semiconductor universe will ensure that Digital Transformation delivers the huge potential of AI/ML technology for corporate strategy and make AI/ML a transformative toolbox and skillset at the B2B level.

All the way through the design and implementation chain, connectivity has become increasingly fundamental, as these Digital Transformation technologies redefine the deployment and future uses for systems and control infrastructure. Advancements in modular technologies are making machines more connected than ever before. Simply, everything needs to be able to ‘talk’, communicate and interact efficiently with each other, and efficient connectivity plays a vital role in a successful migration to Industry 4.0. Molex plays a vital role in AI’s evolution, supporting development and applying AI-enabled capabilities to our customers’ product innovation, quality, efficiency and more, ensuring that our customers, with effective data and connectivity at the heart of their strategic goals, can look positively towards Industry 4.0, and beyond.

1. Study by the Information Technology and Innovation Foundation – https://www.scmp.com/news/china/science/article/3119115/us-leads-world-artificial-intelligence-china-catching-study – Published 25th 2021
2. Study by World Intellectual Property Organization – https://www.forbes.com/sites/insights-ibmai/2020/05/21/where-asia-is-taking-the-world-with-ai/?sh=4cf541257947 – Published May 2020
3. “AI in the Global Supply Chain Management Market, 2020-2025 – Cloud-based AIaaS for SCM Will Reach $1.9B by 2025, Globally” (Research and Markets). – https://www.globenewswire.com/news-release/2020/03/16/2000842/0/en/AI-in-the-Global-Supply-Chain-Management-Market-2020-2025-Cloud-based-AIaaS-for-SCM-Will-Reach-1-9B-by-2025-Globally.html