Lattice Blog


Bolstering AI PC Innovation with FPGA Edge Accelerators

Bolstering AI PC Innovation with FPGA Edge Accelerators
Posted 03/14/2024 by Hussein Osman, Segment Marketing Director

Posted in

Artificial intelligence (AI) is making its way onto a multitude of applications at a rapid pace. Seventy-three percent of U.S. companies use AI in some aspect of their business and almost a quarter (22 percent) of technology companies are aggressively pursuing integration of AI across a wide variety of devices.

As AI continues to permeate modern technology, tailoring on-device AI to meet specific use cases and device requirements becomes imperative as the need for real-time processing, analysis, and response at the Edge is growing. For instance, while generative AI and large language models (LLMs) demand higher computational and memory, certain devices with always-on capabilities necessitate low power consumption. Personal computers (PCs) that we use every day are a prime example. As industry trends shift from the cloud to the PC to improve privacy and reduce dependency on datacenters, it increases the need for enhanced on-device AI solutions that can support seamless, always-on, secure functionality.

Powering Always-On Capabilities with On-Device AI

Edge devices like PCs require AI solutions that can empower their always-on capabilities. This includes greater situational awareness, improved power efficiency, flexible and low latency sensor bridging, reduced complexity, and streamlined integration.

Powering Always-On Capabilities with On-Device AI

Enhanced Situational Awareness

A PC’s always-on capabilities mean they must constantly sense their surroundings, particularly as unauthorized physical access and visual hacking are two of the top concerns with PC security. On-device AI tools can give PCs a sense of the user and the environment, delivering security capabilities to alert the user of possible intrusion and lock the PC when the user steps away.

Efficient Power Usage

While users spend a significant amount of time every day in front of their devices, some of this time is spent focusing on other tasks. On-device AI can help conserve power by automating the adjustment of screen brightness levels based on whether users are leveraging the product.

Sensor Bridging

The wide variety of sensors being used in devices like PCs are only growing with the advancements of on-device AI. Capturing, aggregating, and processing data from multiple sensors efficiently and flexibly is critical to creating a rich user experience by creating a seamless connection between display, audio, and image with reduced data traffic.

Simplified Integration

With a plethora of different technologies requiring always-on capabilities, integrating multiple features seamlessly with other components is essential to reduce system complexities and system development cost.

Improving On-Device AI with Lattice FPGA Solutions

With AI-optimized, low power FPGAs and the Lattice sensAI solution stack with state-of-the-art computer vision attention sensing software, Lattice provides an end to end solution for designers looking to deliver high quality AI experiences. This hardware and software combination enables smart user aware devices from AI/ML model training, validation, and compilation. These tools enable ultra low power computer vision algorithms that extend battery life, improve user privacy and data security, and are completely programmable.

Why FPGA for Edge AI

Equipped with Lattice’s solutions, OEMs can future proof their smart Edge applications to adapt to the rapid pace of AI innovation and gain access to features including smart sensing, attention tracking, onlooker detection, and enhanced hardware security. Specifically, by leveraging Lattice sensAI technology, PCs can automatically power on/off when a user approaches or departs, warn users against onlookers, and enable longer battery life by dimming the screen while the user is distracted.

Since PCs operate at the Edge, on-device AI must be secure and flexible to keep up with the evolving landscape. Lattice’s AI solutions add a layer of security and privacy, without sending anything to the cloud. When using Lattice’s smart sensing solutions, Edge processing is done next to the sensor on a hardware-secured implementation that cannot be altered as hardware designs are encrypted. In sensing mode, only the metadata is shared, and the video path is completely cut off. Additionally, upgrades are authenticated and unauthorized access to the design is not permitted.

Further, as developers work to keep up with rapidly changing technology innovations, Lattice’s FPGAs provide a programmable and future proof platform to deploy and adapt evolving AI models and experiences. The hardware inside Lattice FPGAs is reprogrammable in the field, allowing for AI algorithms to be modified to meet the latest technology trends or requirements over the product’s lifecycle.

Renowned PC providers, such as Dell, Google, Lenovo, and LG, leverage these solutions from Lattice to provide users with the AI-enabled features they need to enhance both security and convenience. Most recently, Dell collaborated with Lattice to build and deliver the next generation of Computer Vision AI solutions powering the new Latitude 7350, 7450, 7650 and 7350 Detachable AI PCs.

Bolstering AI PC Innovation with FPGA Edge Accelerators - Quote

The Future of PCs

As the PC ecosystem continues to evolve with the introduction of AI, ensuring on-device AI has the necessary capabilities to support always-on functionality is paramount. Lattice is dedicated to providing customers with AI-optimized FPGA hardware and software solutions to empower transformative innovation, enhanced user experiences, and top-of-the-line security and privacy on modern PCs and other Edge devices.

To learn more about how Lattice can help you enable on-device AI with its award-winning smart sensing portfolio, reach out to speak with the team at Lattice.