A new white paper by BrainChip Holdings Ltd, the world’s first commercial producer of ultra-low power, fully digital, event-based, neuromorphic AI IP, unveils Temporal Event-based Neural Networks (TENNs) – an innovative way to bring greater accuracy and efficiency to complex models on compact edge devices. TENNs are extremely effective in accelerating 3D and 1D time series applications including video, vision, audio, and vital signs in healthcare to name a few.
“While CNNs have long been the backbone of image classification in AI and ML, they are not as efficient in handling spatiotemporal data and applications such as video object detection from video streams, and Time series data, thereby limiting its usage in cost-effective, thermally constrained edge devices”
The paper, “Temporal Event-based Neural Networks: A new approach to Temporal Processing,” shows how TENNs enable intelligent, energy-efficient edge solutions and details how BrainChip’s 2nd Generation Akida™ IP platform supports this innovation. TENNs provide radical, innovative ways to reduce complexity, size, and compute requirements, while still delivering the accuracy expected for the desired, intelligent, responsive experience at the edge.
AiThority: The 3 Building Blocks to Make AI Accessible
“While CNNs have long been the backbone of image classification in AI and ML, they are not as efficient in handling spatiotemporal data and applications such as video object detection from video streams, and Time series data, thereby limiting its usage in cost-effective, thermally constrained edge devices,” said Anil Mankar, co-founder and CDO at BrainChip. “The TENN is a new approach that exploits the temporal correlations much more efficiently, revolutionizing AI at the edge.
Read More: How ChatGPT Will Transform Customer Service
[To share your insights with us, please write to sghosh@martechseries.com]
Comments are closed.