The impact of AI and machine learning on low power wireless IoT devices and their target markets

Welcome to this live expert panel debate, brought to you by Nordic Semiconductor‘s industry-leading Wireless Quarter publication.

We aim to provide you with a wide-ranging discussion of what the ability to do AI and machine learning on wireless IoT edge devices means and why it matters. What’s happening today and what’s likely to happen in the future.

Panelists:

Steve Roddy, VP of Product Management, Arm Machine Learning Group

Zach Shelby, Co-founder and CEO, Edge Impulse

Svein-Egil Nielsen, CTO/EVP R&D and Strategy, Nordic Semiconductor

Moderators:

Andrew Woolls-King and Steve Keeping

Background

That the application of AI and ML to the IoT is a good thing is perhaps beyond debate. The data generated by billions of wireless sensors monitoring every aspect of society is far too great for humans to analyze. Today, Cloud servers absorb this data and clever AI and ML algorithms extract patterns and anomalies that can be used for forecasts and decision making. But the challenge for engineers is that as the IoT scales, the data increases exponentially. Sending all that raw data to the Cloud becomes too expensive and energy intensive to manage in the long term. The solution is to process the data at the edge of the sensor network in order to decide what’s important enough to forward to the servers. The challenge is that the AI and ML routines are intensive yet even the most powerful edge devices have limited computing and power resources compared to Cloud servers.

The expert panel will debate this problem and will explore some solutions, including streamlined forms of AI and ML that are optimized for low power wireless IoT devices. These new solutions have the added advantage that they don’t require a high degree of software expertise to incorporate onto edge devices.