Back to results
US 2020/0272893 A1Published application
Analog to digital converter using memristors in a neural network
AssigneeTechnion Research & Development Foundation Limited
InventorsLoai DANIAL, Shahar KVATINSKY
Life of this patent
Estimating the term…
Abstract
An analog to digital converter comprises an input for receiving an analog input signal; a plurality of outputs for outputting parallel bits of a digital signal that represents said analog input signal; and a neural network layer providing connections between each of said outputs respectively, each connection having an adjustable weighting. The synapses of the neural networks may be memristors and training may use online gradient descent.
Claims20
An analog to digital converter comprising:
- an input for receiving an analog input signal;
- a plurality of outputs for outputting parallel bits of a digital signal that represents said analog input signal;
- and
- a trainable neural network layer, the neural network layer comprising connections between each of said outputs respectively, each connection having a weighting, the weighting being adjustable for training.
1 more element
An analog to digital conversion method comprising:
- receiving an analog input signal;
- outputting parallel bits of a digital signal that represents said analog input signal at outputs;
- providing connections between each of said outputs respectively;
- and
- providing an adjustable weighting to each connection, thereby to provide a trainable neural network connecting said outputs.
2 more elements
Dependencies are inferred from the claim text.