BrainChip Holdings Ltd. unveiled 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. 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.