Todd Vierra, Vice President of Customer Engagement at BrainChip, demonstrates the company’s latest edge AI and vision technologies and products at the 2024 Embedded Vision Summit. Specifically, Vierra demonstrates the efficient processing of generative text using Temporal Event-based Neural Networks (TENNs) compared to ChatGPT.
The TENN, an innovative, light-weight neural network architecture, combines convolution in the 2D spatial domain with the 1D time domain, resulting in a more efficient 3D convolution. TENNs are particularly effective for LLMs, isolating signal from noise and reducing or eliminating the need for DSP pre-processing or filtering, making them a great solution for all types of sensors and targeting audio de-noising, high-performance video, various vital sign detectors in healthcare and more.
