Interview With Antonio Fernandez, Vice President Technical Development At EBV
Systems with Artificial Intelligence require a wide range of electronic components. This is an area where EBV already has a vast variety of components in its product portfolio, from sensors to FPGAs. However, the new technology poses a challenge for the electronics distributor, as Antonio Fernandez, Vice President of Technical Development at EBV, told us. In order to provide customers with expert assistance when it comes to software development and technical support, the company has joined important manufacturers in developing a common AI roadmap. It is also investing in specialised software support.
AI is making its way into more and more applications. What does that mean for you as an electronics distributor?
Antonio Fernandez: For us, Artificial Intelligence represents both a huge business opportunity and an unprecedented challenge in terms of technical support. That’s our chance to stand out, demonstrating the value of our partnership to our customers.
Is your product portfolio changing?
A.F.: Artificial Intelligence can be deployed in the cloud, at the edge of the network, or most commonly in a combined scenario between edge and cloud. We already have FPGAs, MPSoCs and processors capable of implementing AI in the different application scenarios and a promising roadmap with our key manufacturers and partners of dedicated AI hardware and software.
Alongside AI chips, which components are important for supporting customers in the development of AI systems?
A.F.: That starts with sensors for collecting all the variables that affect a process, then also includes processors to effectively implement the sensor fusion and communication protocols, security chips to safeguard the communications, and much more. Then, to transfer the data, wireless connectivity transceivers – for instance, for Wi-Fi, Bluetooth, Thread or LTE – and various network products are also needed.
Does this also present challenges in consulting and other services offered by EBV?
A.F.: Sure, but our technical specialisation and close collaboration with our manufacturer partners and AI cloud service preferred third parties enable us to take on the challenge.
In particular, AI is also a software topic. How can EBV support customers here?
A.F.: EBV continues to invest in specialised software support resources and our manufacturing partners are developing specific software-development kits that will significantly accelerate the adoption of machine learning in multiple applications. A good example is the Xilinx “reVISION” stack for edge machine learning in embedded vision applications.
AI solutions are also becoming increasingly interesting for companies that have so far had little or no contact with the world of electronics. Can you also support companies like those, and if so, how?
A.F.: Our customers develop and manufacture electronics systems and cover all kinds of applications. When integrators approach us, we try to understand their needs and refer them to some experienced customers, design houses, and contract manufacturers that can solve their requirements and deliver a final product.
What technology trends in AI are you currently seeing among your suppliers, the electronics manufacturers?
A.F.: There are several interesting developments here. For instance, AI specific functions and peripherals are being integrated into the MCUs. This allows small microcontroller-based nodes to interact more intelligently with deep neural networks. Another observation would be an increase in the power of the processors and MPSoCs to implement edge inference engines with very low latency. In addition, the manufacturers are currently investing very heavily in software libraries for sensor fusion and machine learning.
In what areas do you see need for further development in electronic components for AI?
A.F.: Sensors are a critical area for the development of Artificial Intelligence. We still do not have commercial sensors for smell or taste, for instance … Artificial Intelligence is a very complex topic, so specific development kits and training will be required to bring its benefits to the mass market.
In your view, which applications are currently especially interesting or offer particularly high potential for AI systems?
A.F.: Embedded vision, autonomous systems, predictive maintenance, natural-language human-machine interface, robotics, measurement and diagnostics.
Elon Musk is often quoted as saying that AI is one of the biggest risks facing humanity. What is your opinion on this?
A.F.: Artificial Intelligence can enable disruptive innovations that could represent a significant benefit for humanity, increasing our productivity and quality of life; but there is a risk of machines replacing people, which could increase unemployment, the risk of losing privacy with Big Brother controlling everything, and of systems taking unsupervised decisions and doing more harm than good. As usual, it all depends on the use of the technology and we most probably need new rules to regulate the ethical use of Artificial Intelligence.
