Neuromorphic apparatus and method with neural network

[uiux]

https://patents.google.com/patent/US20210406694A1/en

Neuromorphic apparatus and method with neural network

Abstract
A processor-implemented neural network implementation method includes: learning each of first layers included in a neural network according to a first method; learning at least one second layer included in the neural network according to a second method; and generating output data from input data by using the learned first layers and the learned at least one second layer.

---

A new patent from Samsung I am keeping my eye on throughout the application process. This can be followed here:

https://globaldossier.uspto.gov/#/result/publication/US/20210406694/1

 

[alwaysgreen]

Good or bad for us? lol

 

[uiux]

Good or bad for us? lol

Not sure

https://brainchipinc.com/brainchip-...-learned-functions-intelligent-target-device/

Key features of Patent US 11,238,342

• The patent claims protect the basic structure and function of a digital neuron consisting of multiple synapse circuits connected to a soma circuit in analogy to a biological neuron where a soma (i.e. neuron) cell receives its inputs via multiple synapses.

it will be interesting to see if it runs into our patents.

 

[JDelekto]

Not sure

https://brainchipinc.com/brainchip-...-learned-functions-intelligent-target-device/

Key features of Patent US 11,238,342

• The patent claims protect the basic structure and function of a digital neuron consisting of multiple synapse circuits connected to a soma circuit in analogy to a biological neuron where a soma (i.e. neuron) cell receives its inputs via multiple synapses.

it will be interesting to see if it runs into our patents.

It certainly seems like there is a potential for the Samsung patent to get kicked back based on the description of BrainChip's key features. I guess it comes down to whether the mathematical model for the neuron vs. the digital model of the neuron is considered different enough.

 

[Fact Finder]

It certainly seems like there is a potential for the Samsung patent to get kicked back based on the description of BrainChip's key features. I guess it comes down to whether the mathematical model for the neuron vs. the digital model of the neuron is considered different enough.

I just found the following paper. It does not reference AKIDA but does dig deeply into using Loihi. In doing so it throws up performance figures for gesture recognition, identifying objects in a video stream and training which make it look very much like a very early failed prototype in the development timeline in the creation of AKIDA:

https://arxiv.org/pdf/2107.00401

My opinion only DYOR
FF

AKIDA BALLISTA