US 2021/0342672 A1 · CROSSBAR ARRAYS FOR COMPUTATIONS IN MEMORY-AUGMENTED NEURAL NETWORKS

Abstract

In a hardware-implemented approach for operating a neural network system, a neural network system is provided comprising a controller, a memory, and an interface connecting the controller to the memory, where the controller comprises a processing unit configured to execute a neural network and the memory comprises a neuromorphic memory device with a crossbar array structure that includes input lines and output lines interconnected at junctions via electronic devices. The electronic devices of the neuromorphic memory device are programmed to incrementally change states by coupling write signals into the input lines based on: write instructions received from the controller and write vectors generated by the interface. Data is retrieved from the neuromorphic memory device, according to a multiply-accumulate operation, by coupling read signals into one or more of the input lines of the neuromorphic memory device based on: read instructions from the controller and read vectors generated by the interface.

Claims18

Dependent claims

A hardware-implemented method for operating a neural network system, the method comprising:

providing a neural network system comprising a controller, a memory, and an interface connecting the controller to the memory, wherein the controller comprises a processing unit configured to execute a neural network and the memory comprises a neuromorphic memory device with a crossbar array structure that includes input lines and output lines interconnected at junctions via electronic devices;
programming the electronic devices of the neuromorphic memory device to incrementally change states thereof by coupling write signals into one or more of the input lines based on:
(i) write instructions received from the controller and (ii) write vectors generated by the interface, according to the write instructions;
and
retrieving data from the neuromorphic memory device, according to a multiply-accumulate operation, by coupling read signals into one or more of the input lines of the neuromorphic memory device based on:
(i) read instructions from the controller and (ii) read vectors generated by the interface, according to the read instructions.

3 more elements

A neural network system comprising:

a controller including a processing unit configured to execute a neural network,
a memory comprising a neuromorphic memory device with a crossbar array structure that includes input lines and output lines interconnected at junctions via electronic devices, and
an interface connecting the controller to the memory,
wherein the system is configured to:
program the electronic devices of the neuromorphic memory device to incrementally change states thereof by coupling write signals into one or more of the input lines based on:
(i) write instructions received from the controller and (ii) write vectors generated by the interface, according to the write instructions;
and
retrieve data from the neuromorphic memory device, according to a multiply-accumulate operation, by coupling read signals into one or more of the input lines of the neuromorphic memory device based on:
(i) read instructions from the controller and (ii) read vectors generated by the interface, according to the read instructions.

6 more elements

Dependencies are inferred from the claim text.

Crossbar arrays for computations in memory-augmented neural networks

Life of this patent

Crossbar arrays for computations in memory-augmented neural networks

Life of this patent

Abstract

Claims18