- BrainChip Holdings Ltd (BRCHF, Financial) CTO, Dr. M. Anthony Lewis, to present at the Embedded Vision Summit, focusing on architectural innovation for low-power AI.
- Presentation will contrast State-Space Models with Transformers for energy-efficient edge AI applications.
- Attendees can explore BrainChip's Edge AI technology advancements at booth #716 during the Summit.
BrainChip Holdings Ltd (ASX: BRN, OTCQX: BRCHF, ADR: BCHPY), the pioneering company in ultra-low power, fully digital, event-based neuromorphic AI, has announced that its Chief Technology Officer, Dr. M. Anthony Lewis, will be presenting at the Embedded Vision Summit scheduled for May 20–22, 2025, in Santa Clara, CA. This prestigious conference is a significant event for innovators in computer vision and edge AI technologies.
Dr. Lewis's presentation, "State-Space Models vs Transformers at the Extreme Edge: Architectural Choices for Low Power AI," will delve into the design trade-offs and opportunities for AI models operating under severe energy and computational constraints. He will discuss how State-Space Models (SSMs), including BrainChip's proprietary Temporal Event-based Neural Networks (TENNs), outperform traditional Transformer architectures in edge scenarios by reducing energy consumption and chip area through read-only operations and fewer MAC units.
The session will also cover strategies for transitioning from Transformer-based models, such as LLaMA, to SSMs via distillation, maintaining accuracy while enhancing efficiency. Attendees will have the opportunity to visit BrainChip at booth #716 for live demonstrations of the company's latest innovations in on-chip language processing and ultra-low power inference, powered by Akida solutions.
"State-space models are transforming the potential of real-time, low-power intelligence at the edge," commented Dr. Lewis. "Our presentation will demonstrate BrainChip's commitment to delivering scalable, efficient, and interactive AI solutions for the most challenging embedded environments today."
BrainChip's participation in the event underscores its leadership in the booming realm of real-time streaming Edge AI, offering highly efficient computational solutions across various industries, including aerospace, automotive, robotics, consumer electronics, and wearables.
