Last week, we talked about Ceva, Inc. – let’s recap, if for a moment.
Ceva is comprised of teams with bona fide passion for seamlessly connecting smart edge products so they can sense and infer data more reliably, particularly when it comes to high-end consumer IoT and IIoT applications. Ceva’s portfolio of comprehensive communications and scalable edge AI IP makes great advancements in mobile, automotive, infrastructure, and even personal computing all the more real for the modern age.
In this vein, our coverage last week detailed Ceva’s partnership with STMicroelectronics; the latter having integrated Ceva’s tech into the recently introduced ST87M01 ultra-compact, low-power line of modules. In short, these combine NB-IoT data communication with resilient GNSS geo-location capabilities for IoT assets, creating opportunities in industrial condition monitoring, factory automation, smart meters and grids, smart agriculture-related efforts, environmental monitoring and more. (Read the full story here.)
Today, we’ve got additional Ceva news to discuss.
On top of what we’ve already shared, Ceva also recently announced that it has extended its Ceva-NeuPro family of Edge AI neural processing units (NPUs) with the introduction of Ceva-NeuPro-Nano NPUs.
Per the official announcement, these highly efficient (and self-sufficient) NPUs “deliver the power, performance and cost efficiencies needed for semiconductor companies and OEMs to integrate TinyML models into their SoCs for consumer, industrial, and general-purpose AIoT products.” (AIoT, of course, being the Artificial Intelligence of Things, a branch of IoT with clear focus on practical and ubiquitous AI.)
Further context below:
As mentioned, TinyML (Tiny Machine Learning) refers to the deployment of ML models on low-power, resource-constrained devices to maximize AI in the IoT space. ABI Research reports that, by 2030, “over 40% of TinyML shipments will be powered by dedicated TinyML hardware.” This signifies increasing market demand for specialized AI-centric solutions in IoT, which ties back into the usefulness of the Ceva-NeuPro-Nano NPUs.
Moreover, the Ceva-NeuPro-Nano Embedded AI NPU optimized architecture is fully programmable and efficiently executes neural networks, feature extraction, control code and DSP code, and it supports most advanced ML data types and operators (including native transformer computation, sparsity acceleration and fast quantization). This optimized architecture enables Ceva-NeuPro-Nano NPUs to deliver immense power efficiency with a smaller silicon footprint, compared to existing processor solutions used for TinyML workloads (i.e. which utilize a combination of CPU or DSP with AI accelerator-based architectures). And with Ceva-NetSqueeze AI compression tech that directly processes compressed model weighs (without the need for an intermediate decompression stage), the Ceva-NeuPro-Nano NPUs can achieve, quote, “up to 80% memory footprint reduction,” resolving (by and large) a key bottleneck inhibiting broader adoption of AIoT processors.
Quite a lot of info, we know. For many folks, it’s sort of like rocket science.
For Ceva, it’s a Tuesday.
According to Paul Schell, Industry Analyst at ABI Research, “Ceva-NeuPro-Nano represents a compelling solution for on-device AI in smart edge IoT devices. It addresses the power, performance, and cost requirements to enable always-on use cases on battery-operated devices integrating voice, vision, and sensing use cases across a wide array of end markets. From true wireless stereo (TWS) earbuds, headsets, wearables and smart speakers to industrial sensors, smart appliances, cameras and home automation devices, Ceva-NeuPro-Nano enables TinyML to give AIoT devices much-needed boosts.”
Chad Lucien, Vice president and General Manager of the Sensors and Audio Business Unit at Ceva, provided additional insight.
“Ceva-NeuPro-Nano opens exciting opportunities for companies to integrate TinyML applications into low-power IoT SoCs and MCUs and builds on our strategy to empower smart edge devices with advanced connectivity, sensing and inference capabilities. The Ceva-NeuPro-Nano family of NPUs enables more companies to bring AI to the very edge, resulting in intelligent IoT devices with advanced feature sets that capture more value for our customers.”
“By leveraging our industry-leading position in wireless IoT connectivity and strong expertise in audio and vision sensing,” Lucien added, “we are uniquely positioned to help our customers unlock the potential of TinyML to enable innovative solutions that enhance user experiences, improve efficiencies, and contribute to a smarter, more connected world.”
More Ceva-NeuPro-Nano features, ultra-low memory requirements and energy optimization specifics can be perused here.
Edited by
Greg Tavarez
