Media note: BrainChip introduces aTENNuate in new technical document

BrainChip Holdings Ltd (ASX: BRN, OTCQX: BRCHF, ADR: BCHPY), the world’s first commercial ultra-low-power, all-digital, event-based neuromorphic AI manufacturer, today released a technical paper detailing the Company’s aTENNuate, a lightweight deep-state space autoencoder algorithm capable of edge-optimized denoising, super-resolution and dequantization of raw audio data.

“Real-time speech enhancement on raw signals using deep state-space modeling” further extends BrainChip’s leadership in neuromorphic event-based computing by presenting the company’s approach in advancing state-of-the-art deep learning audio noise reduction methods. The white paper covers topics related to state-space modeling, network architecture, event-based benchmark experiments, and future opportunities such an approach can offer.