Calibration of electrical parameters in a deep learning artificial neural network
AssigneeSilicon Storage Technology, Inc.
InventorHieu Van Tran
Life of this patent
Estimating the term…
Abstract
Numerous examples are disclosed for performing calibration of various electrical parameters in a deep learning artificial neural network. In one example, a method comprises generating a machine learning model by applying a range of electrical parameters to a first non-volatile memory system and measuring resulting characteristics and comparing to a predetermined target range; and during operation, applying electrical parameters the first non-volatile memory system or a second non-volatile memory system using the machine learning model.
Claims18
A method comprising:
- generating a machine learning model by applying a range of electrical parameters to a first non-volatile memory system and measuring resulting characteristics and comparing to a predetermined target range;
- and
- during operation, applying electrical parameters the first non-volatile memory system or a second non-volatile memory system using the machine learning model.
A system comprising:
- a current digital-to-analog converter to generate a current in response to a digital input;
- and
- a non-volatile memory cell comprising a first terminal, a second terminal, and a third terminal, wherein the first terminal receives the current from the current digital-to-analog converter and the third terminal is coupled to ground, and wherein a voltage of the second terminal is adjusted until a voltage of the first terminal exceeds or falls below a reference voltage, the voltage of the second terminal then being applied as a bias voltage to an array of non-volatile memory cells.
A method comprising:
- calibrating an array of non-volatile memory cells;
- measuring a characteristic of the array, wherein the characteristic is a voltage, a temperature, or a time duration for a thermal event to occur;
- and
- comparing the measured characteristic against a target value, and when the measured characteristic equals the target value or is within a predetermined range, repeating the measuring and comparing, and when the measured characteristic does not equal the target value or fall within a predetermined range, repeating the calibrating, measuring, and comparing.
A method comprising:
- calibrating an array of non-volatile memory cells;
- performing a temperature check, the temperature check comprising measuring a temperature and comparing the measured temperature against a target temperature;
- when the measured temperature equals the target temperature or is within a predetermined range, repeating the temperature check;
- and
- when the measured temperature does not equal the target temperature or fall within a predetermined range, performing a voltage check, the voltage check comprising measuring a voltage in the array and comparing the measured voltage against a target voltage, and when the measured voltage equals the target voltage or falls within a predetermined range, repeating the temperature check and when the measured voltage does not equal the target voltage, repeating the calibrating.
A method comprising:
- adjusting a first reference voltage in a global digital-to-analog converter;
- adjusting a second reference voltage in the global digital-to-analog converter;
- applying the adjusted first reference voltage and the adjusted second reference voltage to a trim circuit in the global digital-to-analog converter to adjust an output of the global digital-to-analog converter;
- and
- generating a bias voltage for a plurality of rows in an array using the global digital-to-analog converter.
Dependencies are inferred from the claim text.