Industry 4.0 Meets AI

We’re at a fascinating inflection point. Industry 4.0—a long-used term in manufacturing—is now more mainstream, due to the availability of affordable IoT infrastructure, the desire to gain new business insights from data, plus the arrival of 5G.

Next technology wave

While it’s still early days, we see artificial intelligence (AI) as the next big technology wave, delivering new insights for the physical world. While the internet of things (IoT) is providing us deeper insights into individual processes, AI looks at the aggregate, delivering a holistic 360 view of the business. Together, you get the total picture.

Investing in AI

At Dell Technologies, we’re continuously blown away by the advancements in AI from innovative companies like Noodle.ai, Edico Genome, FogHorn Systems, Graphcore, Moogsoft, Zingbox, and Immuta (start-ups that we’re proud to support through our financial capital division).

And we’re also leveraging AI in our own business functions. For example, to combat unconscious bias in our recruitment process, we work with companies like Textio, who use an intelligent text editor to avoid gendered phrasing in job descriptions, while Mya Systems has developed an intelligent chatbot to “interview” and evaluate job candidates, using objective, performance-based questions.

Powering AI applications

So, how do our OEM & IoT Solutions advance AI? Simple. All these intelligent applications—designed largely by OEM-type companies—are generating new requirements to manage scale and latency, for example, compute power, storage, software-defined networking and virtualization. As a result, the industry will need to completely redesign the entire end-to-end infrastructure.

As the essential infrastructure company, we’re already providing the IT infrastructure to power AI workloads–the rocket fuel you might say. Last year, we expanded our server and workstation portfolio to accelerate AI-driven workloads as well as deepening our relationship with Intel to advance AI community innovation. Meanwhile, we offer our Ready Solutions for AI including two design options: Deep Learning with NVIDIA and Machine Learning with Hadoop. Both include pre-validated hardware/software stacks that can be rapidly implemented to accelerate AI.

Computer vision and machine intelligence working together

I see AI enhancing existing best practices in industrial automation. Of course, vision-based quality control systems have been used in manufacturing for some time but AI advancements and falling costs of acceleration hardware—like GPUs and FPGAs—are making computer-vision-based insights more accessible.

Take the existing inspection cameras on a manufacturing conveyor belt. The output of those vision-based analytics becomes even richer when combined with additional data. For example, consider the telemetry generated by PLCs and sensors located on the machines that manufacture the inspected parts. By combining and analyzing this rich mix of data, the production, quality and maintenance teams gain fresh and more detailed insights about their overall operations. Now, that’s Industrial Intelligence!

From car manufacturing to mining operations

A picture paints a thousand words so let’s look at some customer examples. We’re proud to work with Hadoop, an Italian industrial automation company, specializing in automated manufacturing systems. Comau has digitized the door assembly line for Maserati Levante. A digital dashboard identifies potential bottlenecks or breakdowns. If issues arise, the customer can print the necessary spare part – directly onsite – using a 3D printer mounted on a Comau robot NJ 60. The solution, developed in partnership with Autodesk and Continuous Composites, and powered by our technology shows how companies can produce strong, lightweight factory-ready parts on demand. For Maserati, this level of predictive intelligence delivers higher productivity and lower costs.

In the engineering world, we’re working with The Weir Group to help transform its global business model and roll out predictive maintenance as a service. More on that partnership shortly from my colleague Dermot O’Connell. And Noodle.ai is improving how the world makes and moves things—for factory floors, trucking fleets, distribution centers, and retail stores. The company helped Big River Steel, save energy, improve product quality, and identify new areas for revenue growth.

From autonomous driving to deep-sea exploration

Zenuity is breaking new ground in world-class driver assistance and autonomous-driving technologies. When the company needed to analyze data from cars, tested in a wide variety of environments across the globe, we provided an HPC and storage environment to handle large-scale data loads, reduce technology troubleshooting time, and accelerate time-to-market.

In the marine world, the Arggonauts from the Fraunhofer IOSB in Karlsruhe are using mobile robotics to revolutionize deep-sea exploration. Using our customized workstation technology, the team controls remote-operated vehicles and captures subsea data camera images. An HPC Datacenter compute renders the images and translates data into maps while AI is used to quickly classify images from the unstructured data (read the full story here).

Protecting endangered species

On the opposite end of the scale, one of our analytics partners, SAS®, is collaborating with the non‐profit Wild Track to monitor and protect endangered species like cheetahs and leopards. With the help of SAS® technology—powered by Dell PowerEdge server infrastructure—WildTrack is collecting footprint images and analyzing them with a customized statistical model to gain new insights and help protect these animals from extinction.

Just think of the huge range of societal and business challenges these AI solutions are addressing: better manufacturing quality, increased uptime, improved knowledge of our world, reduced energy consumption and the conservation of endangered species.

Looking ahead

A massive load of sensors is coming online. And, thanks to 4K video cameras, we’re already seeing the enormous amount of high-resolution data from the physical world yield new analytics-driven outcomes. As we couple this with 5G technology, we’ll have access to an entirely new range of real-time services and enhanced experiences.

However, even with 5G’s high bandwidth and low latency capabilities, the connectivity, cost, security and uptime requirements will drive an increasing amount of edge computing. This is especially true for critical operations that must avoid exposing their processes directly to the cloud. Today, training AI models for edge analytics typically happens in centralized data centers before these models are pushed to the edge for real-time inferencing on data streams. Over time, I expect to see more of this model training happening closer to the edge.

Adapt as your business needs evolve

Today, we’re merely scratching the surface of the potential for AI and IoT. So, it’s critical to have consistent, flexible infrastructure that enables you to adapt as your business needs evolve. Dell Technologies is the only company on the planet that has hardware and software solutions that play at every level of AI and IoT implementation, from the edge, to the core, to the cloud. We can deliver all the necessary assets, including scalable, secure, manageable and open infrastructure architecture, IoT and big data expertise, the ability to customize through our OEM division, the right partners plus a sophisticated global support and supply chain.

I’m excited about the potential for AI to help solve big societal and business challenges. I certainly believe that it will enhance and support our human efforts—not replace them.