Skip to main content

ew 2024: Advantech Addresses Edge AI Challenges

embedded world 2024 offers a glimpse into the future of edge AI leveraging the latest tools and technologies with Thomas Kaminski, Director of Product Sales and Marketing Management; Alex Troeger, Head of SW Engineering; and Umar Ahmad, IoT Platform Field Application Engineer at Advantech, edge AI platform and AIoT solutions provider.

Transcript

Tell us about your customers and how they are staying up to date with the latest trends

Thomas Kaminski: Our primary customers are typically industrial customers, customers from the retail, from the medical devices, from the healthcare. So, some are saying like that. If you are asking me about AI applications today, I have to say it is just at the beginning. And it’s also a question, what is the definition of AI today? So thinking about a real AI application, that’s just at the beginning. The research companies are saying and according to the market that within the next two or three years, 50% of all edge devices will have AI capabilities inside. So in other words, that market is developing right now is starting right now. The platforms, the right platforms supporting AI technologies are being launched, are just launched, and as well all software tools to develop and deploy the models.

What are the most challenging aspects of edge AI applications?

Thomas Kaminski: So biggest challenge is, first of all, to bring the computing power to the edge. So to avoid capturing data at the edge, transmitting data into the cloud, into the server centers where today that AI execution is done. So the biggest weakness of the old process is you need to transmit more data into the cloud. You have the latency and you have a kind of safety and security problem with your edge device because you’re capturing the data, you’re transmitting the data. By keeping the data on the edge device themself, you’re making sure that you can execute your AI model in real time. You’re not losing any time to transmit the data. You’re keeping all that sensitive data on your local device. You’re not sending them to somewhere into the cloud and into the server center where somebody can grab them or can grab them in between which is looking for the data protection rule as well.

How is Advantech leveraging the latest Intel technologies?

Thomas Kaminski: So Intel today is our biggest partner and most important partner, I have to say first of all. So we are following all the latest generation on different form factors on our boards and our systems and our scalable systems. Thinking about AI capabilities, the newest, latest generation which is still now being launched by Intel, the code name is Meteor Lake, or what Intel is saying now, the Ultra Processor Series 1, it’s the first CPU who has an integrated VPU to really execute AI models and have the computing power around to process all that application data which is as well necessary in that application.

What does the release of the latest Intel Arc Graphics mean for your customers?

Alex Troeger: Intel’s new Arc HAI-GPU Graphic cards are basically addressing the AI solutions in terms of bringing more performance to existing systems. So it’s feasible to have a system to upgrade it, to be capable to handle break AI loads in an existing system but also for new designs. And therefore, you have different standards in the market. Basically, MXM Type A, MXM Type B, PCI express graphic cards. While at Advantech is mainly addressing currently with the former Intel Arc A370 chip set, MXM Type A GPU sets, and PCI express chip sets. With the new A380 chip set, basically also handled by the Advantech’s own design, we address the more performance orientated stuff which means we are capable to handle more AI workloads in the future with an Intel® Arc GPU chip set.

How does Advantech help its customers deploy edge AI solutions?

Umar Ahmad: So we are delivering our customers Intel® AI hardware and we complement it with software services. So when they want to move into production, this is where we come and help them. We know that they want to do AI but they lack in-house resources. So, we try to help them, setting up the software toolkit, optimizing it so they can efficiently move into production stage using Intel AI hardware and complementary software services from Advantech.

What does embedded world 2024 mean to Advantech

Thomas Kaminski: I think the most exciting during these last three days was the launch of the latest A380 chip set from Intel which is supporting AI capabilities on existing platforms. We also have seen the new product launch of integrated NPUs, so which has open the door for the future for that kind of applications. And also the Intel® Geti software which in the end is bringing AI ideas, hardware and software tools together into a solution. We had a lot of interesting discussion with customer, developing new ideas for new applications in the different industries, talking about retail, talking about medical application, where we are pretty sure that these technologies will become true in a real application within the next years.

About the Author

Brandon is a long-time contributor to insight.tech going back to its days as Embedded Innovator, with more than a decade of high-tech journalism and media experience in previous roles as Editor-in-Chief of electronics engineering publication Embedded Computing Design, co-host of the Embedded Insiders podcast, and co-chair of live and virtual events such as Industrial IoT University at Sensors Expo and the IoT Device Security Conference. Brandon currently serves as marketing officer for electronic hardware standards organization, PICMG, where he helps evangelize the use of open standards-based technology. Brandon’s coverage focuses on artificial intelligence and machine learning, the Internet of Things, cybersecurity, embedded processors, edge computing, prototyping kits, and safety-critical systems, but extends to any topic of interest to the electronic design community. Drop him a line at techielew@gmail.com, DM him on Twitter @techielew, or connect with him on LinkedIn.

Profile Photo of Brandon Lewis