November will mark the fourth annual running of Edge AI Summit, a conference organized by U.K.-based event management firm Kisaco Research.

Taking place both virtually Nov. 15-18.

“It was launched off the back of our AI Hardware Summit when we realized that there were a lot of companies working on increasing computing capabilities on devices and edge nodes to try and push machine learning processing closer to the sources of data,” says Kisaco Research’s strategic advisor Ed Nelson.

“In the first year the summit was very focused on compute architectures for devices and the ‘extreme edge’ (microprocessors and sensors), but over the years the event has incorporated much more software-focused content.”

Nelson adds that in addition to looking at how to improve compute and memory capacity at the edge, “we now cover techniques such as algorithm development to reduce the size and complexity of ML models, and ingenuities in software and application development, unearthing use cases from a variety of vertical industries.  

“In essence, the Edge AI Summit is still a technology-focused event rather than an applications-focused event, but the goal is to strike a balance between looking at how Edge AI can change the world and discussing how to actually make it work.”

Meanwhile, conference producer Moses Makin, says there will be several main themes this year all with accompanying speakers. These include:

  • Clarifying the benefit and costs of deploying AI on the edge vs. on the cloud, or taking a hybrid approach for end users, so that they can make a robust business decision about where to deploy their AI. Keynote speaker Ganesh Harinath, vice president and chief technology officer of Verizon Media will be speaking on this topic.


  • Helping edge computing, AI hardware and TinyML software vendors understand what challenges their customers – which are generally original equipment manufacturers, device manufacturers and ISVs – have and where their tools may assist in helping to deploy AI at the edge. Speakers include members of AI teams from LG Electronics, Ford as well as a number of start-ups looking to scale.


  • Helping original equipment manufacturers, ISVs and enterprise users understand which edge computing, AI hardware and TinyML software tools are best to use to ensure high-performing AI on resource constrained IoT and edge devices.


“We will also be bringing the whole value chain together to talk about overcoming interoperability challenges which often prevent edge AI proof-of-concepts being turned into real-world applications,” says Makin.

“Some interesting companies speaking include, Intel, the innovative neuromorphic computing companies like Brainchip, Innatera and SynSense, and the edge computing providers who facilitate AI deployment on IoT devices, like Edgeworx. They are explaining how they work with partners to ensure high performing, economical, secure and trustworthy AI systems at the edge.”

As for what prompted organizers to implement a hybrid approach of both virtual and live, Zach Butler, managing director of tech, media and telecom with Kisaco, says the past 18 months have “reminded us all of the power of and the benefits of doing business through physical interactions.

“Many in our community have been excited to hear of our plans to return to Silicon Valley this November, but we understand that for some it's still a little too soon to return to an in-person event.

“Our hybrid event format will unite the power of a live onsite event; through the main conference, exhibition area, roundtables, and meetings, alongside enabling a global virtual audience to consume the same content and network with both virtual and in-person attendees via our AI-powered virtual event platform.”

Paul Barker