It is a big claim that Brainchip management have achieved...

It is a big claim that Brainchip management have achieved nothing.

Perhaps the partnership with TATA ELXSI to drive the adoption of AKIDA technology in medical and industrial use cases is not as significant as one might reasonably think given that TATA Consulting Services the TATA Group research ARM has only published about AKIDA in two or three areas. On the medical front the following seems like something that TATA ELXSI could run with given it has chosen to partner with Brainchip and not Intel:

Towards low power cognitive load analysis using EEG signal: A neuromorphic computing approach

Dighanchal Banerjee, Sounak Dey, Debatri Chatterjee, Arpan Pal

TCS Research, India

{dighanchal.b|sounak.d|debatri.chatterjee|arpan.pal}tcs.com

Abstract

Real-time on-device cognitive load assessment using EEG is very useful for ap- plications like brain-computer interfaces, robotics, adaptive learning etc. Existing deep learning based models can achieve high accuracy, but due to large memory and energy requirement, those models can not be implemented on battery driven low-compute, low-memory edge devices such as wearable EEG devices. In this paper, we have used brain-inspired spiking neural networks and neuromorphic computing paradigms, that promises at least 104 times less energy requirement compared to existing solutions. We have designed two different spiking network architectures and tested on two publicly available cognitive load datasets (EEG- MAT & STEW). We achieved comparable accuracy with existing arts, without performing any artifact removal from EEG signal. Our model offers ∼ 8× less memory requirement, ∼ 103× less computational cost and consumes maximum 0.33 μJ energy per inference

4 Conclusion

In this paper, we have presented an application of SNN and NC for real time in-situ cognitive load assessment using EEG signal targeted towards implementation on wearable EEG device. We have shown that the proposed method works without artifact removal from the signal and works best for LOSO validation mode. The model has very low memory and energy footprint making it eligible for implementing on battery driven EEG devices. In future, we want to test the model on few more EEG dataset to prove its robustness and implement it on real neuromorphic hardware such as Intel Loihi, Brainchip Akida etc.

Then again perhaps TATA EXLSI has decided to run down the industrial aerospace route in view of the following research by Tata Consulting Services:

Low Power & Low Latency Cloud Cover Detection in Small Satellites Using On-board Neuromorphic Processors

C Kadway, S Dey, A Mukherjee, A Pal… - … Joint Conference on …, 2023 - ieeexplore.ieee.org

… of training an SNN in a direct supervised manner, which is difficult due to the discontinuous
nature of spikes, we convert a pre-trained CNN into a compatible SNN using the Akida …

Then again perhaps these sorts of engagements just happen by accident and management have had nothing at all to do with them and just bask in the brilliance of AKIDA technology.

My opinion only DYOR

Fact Finder