International business machines corporation (20240161792). COMPENSATION FOR CONDUCTANCE DRIFT IN ANALOG MEMORY simplified abstract

From WikiPatents
Jump to navigation Jump to search

COMPENSATION FOR CONDUCTANCE DRIFT IN ANALOG MEMORY

Organization Name

international business machines corporation

Inventor(s)

Charles Mackin of San Jose CA (US)

COMPENSATION FOR CONDUCTANCE DRIFT IN ANALOG MEMORY - A simplified explanation of the abstract

This abstract first appeared for US patent application 20240161792 titled 'COMPENSATION FOR CONDUCTANCE DRIFT IN ANALOG MEMORY

Simplified Explanation

The abstract describes a system that can compensate for activation drift in analog memory-based artificial neural networks by mapping output activation vectors from different points in time.

  • The system involves inputting a set of activation vectors into a crossbar array at two different points in time.
  • Output activation vectors are read from the crossbar array at each point in time.
  • A function is determined to map the output activation vectors from the second point in time to those from the first point in time.
  • This function is then applied to subsequent output activation vectors to compensate for activation drift.

Potential Applications

This technology could be applied in various fields such as:

  • Artificial intelligence
  • Machine learning
  • Robotics

Problems Solved

This technology helps address the issue of activation drift in analog memory-based artificial neural networks, ensuring more accurate and reliable results.

Benefits

The system provides a method to compensate for activation drift, improving the performance and reliability of artificial neural networks.

Potential Commercial Applications

The technology could be utilized in industries such as:

  • Healthcare
  • Finance
  • Automotive

Possible Prior Art

One possible prior art could be the use of calibration techniques in analog memory-based artificial neural networks to address activation drift.

What are the specific calibration techniques used in this system?

The specific calibration techniques used in this system involve determining a function to map output activation vectors from different points in time.

How does this system compare to existing methods for compensating for activation drift in artificial neural networks?

This system offers a unique approach by utilizing a crossbar array and mapping functions to compensate for activation drift, which may provide more efficient and effective results compared to traditional methods.


Original Abstract Submitted

a system can compensate for activation drift in analog memory-based artificial neural networks. a set of input activation vectors can be input, at a first point in time, to a crossbar array. the first set of output activation vectors can be read from the output lines of the crossbar array. at a second point in time, which is a later time than the first point in time, the input set of activation vectors can be input to the crossbar array. a second set of output activation vectors can be read from the crossbar array. a function that maps the second set of output activation vectors to the first set of output activation vectors can be determined. the function can be applied to subsequent output activation vectors output by the crossbar array. a method thereof, can also be provided.