Parameter search method

Patent metadata

Field	Value
Publication number	US-11934759-B2
Application number	US-202017424043-A
Country	US
Kind code	B2
Filing date	Feb 4, 2020
Priority date	Feb 15, 2019
Publication date	Mar 19, 2024
Grant date	Mar 19, 2024

Abstract

A parameter candidate for a semiconductor element is provided. A data set of measurement data is provided to a parameter extraction portion, and a model parameter is extracted. A first netlist is provided to a circuit simulator, simulation is performed using the first netlist and the model parameter, and a first output result is output. A classification model learns the model parameter and the first output result and classifies the model parameter. A second netlist and a model parameter are provided to the circuit simulator. A variable to be adjusted is supplied to a neural network, an action value function is output, and the variable is updated. The circuit simulator performs simulation using the second netlist and the model parameter. When a second output result to be output does not satisfy conditions, a weight coefficient of the neural network is updated. When the second output result satisfies the conditions, the variable is judged to be the best candidate.

First claim

The invention claimed is:

1. A parameter search method using a classification model, a neural network, a parameter extraction portion, a circuit simulator, and a control portion, comprising:
a step of providing a data set of a semiconductor element to the parameter extraction portion,
a step of extracting a model parameter of the semiconductor element by the parameter extraction portion,
a step of performing simulation by the circuit simulator using a first netlist and the model parameter and outputting a first output result,
a step of learning the first output result by the classification model, classifying the model parameter, and outputting a first model parameter,
a step of providing a second netlist and a second model parameter from the control portion to the circuit simulator,
a step of supplying a first model parameter variable included in the second model parameter from the control portion to the neural network,
a step of calculating a first action value function Q from the first model parameter variable by the neural network,
a step of updating the first model parameter variable to a second model parameter variable by the control portion using the first action value function Q and outputting a third model parameter,
a step of performing simulation by the circuit simulator using the second netlist and the third model parameter and outputting a second output result,
a step of judging the second output result by the control portion using a convergence condition given to the second netlist, and
a step of setting a reward by the control portion when the second output result is judged not to satisfy required characteristics of the second netlist and updating a weight coefficient of the neural network using the reward,
wherein when the second output result is judged to satisfy the required characteristics of the second netlist, the first model parameter variable is judged to be the best candidate for the second netlist.

2. The parameter search method according to claim 1 , wherein the first netlist includes any one of or a plurality of an inverter circuit, a source follower circuit, and a source-grounded circuit.

3. The parameter search method according to claim 1 , wherein the number of the first model parameter variables is greater than or equal to two.

4. The parameter search method according to claim 1 , wherein the number of units of output layers of the neural network is twice or more as large as the number of the model parameter variables.

5. The parameter search method according to claim 1 , wherein the first output result extracted using the first netlist includes any one of or a plurality of a leakage current, an output current, signal rise time, and signal fall time.

6. The parameter search method according to claim 1 , wherein a semiconductor element used for the first netlist is a transistor whose semiconductor layer includes a metal oxide.

7. A parameter search method using a classification model, a neural network, a parameter extraction portion, a circuit simulator, and a control portion, comprising:
a step of supplying measurement data of a semiconductor element and a data set including a process parameter to the parameter extraction portion,
a step of extracting a model parameter by the parameter extraction portion,
a step of providing a first netlist from the control portion to the circuit simulator,
a step of outputting a first output result by the circuit simulator using the model parameter and the first netlist,
a step of learning the model parameter and the first output result by the classification model, classifying the model parameter, and outputting a first model parameter,
a step of providing a second netlist and a second model parameter from the control portion to the circuit simulator,
a step of supplying a first model parameter variable included in the second model parameter from the control portion to the neural network,
a step of calculating a first action value function Q from the first model parameter variable by the neural network,
a step of updating the first model parameter variable to a second model parameter variable by the control portion using the first action value function Q and outputting a third model parameter,
a step of performing simulation by the circuit simulator using the second netlist and the third model parameter and outputting a second output result,
a step of judging the second output result by the control portion using a convergence condition given to the second netlist,
a step of setting a high reward by the control portion in the case where the second output result is close to the convergence condition and setting a low reward by the control portion in the case where the second output result is far from the convergence condition when the second output result is judged not to satisfy required characteristics of the second netlist,
a step of calculating a second action value function Q from the second model parameter variable by the neural network, and
a step of updating a weight coefficient of the neural network by the neural network using an error calculated using the reward, the first action value function Q, and the second action value function Q,
wherein when the second output result is judged to satisfy the required characteristics of the second netlist, the first model parameter variable is judged to be the best candidate for the second netlist.

8. The parameter search method according to claim 1 , wherein the first output result extracted using the first netlist includes any one of or a plurality of a leakage current, an output current, signal rise time, and signal fall time.

9. The parameter search method according to claim 1 , wherein a semiconductor element used for the first netlist is a transistor whose semiconductor layer includes a metal oxide.

10. The parameter search method according to claim 7 , wherein the first netlist includes any one of or a plurality of an inverter circuit, a source follower circuit, and a source-grounded circuit.

11. The parameter search method according to claim 10 , wherein the number of the first model parameter variables is greater than or equal to two.

12. The parameter search method according to claim 10 , wherein the number of units of output layers of the neural network is twice or more as large as the number of the model parameter variables.

Assignees

• Semiconductor Energy Lab

Inventors

• Oguni Teppei
• Osada Takeshi
• Fukutome Takahiro

Classifications

• G06N3/092

  Reinforcement learning · CPC title

• G06F18/241

  relating to the classification model, e.g. parametric or non-parametric approaches · CPC title

• G06F30/36

  Circuit design at the analogue level · CPC title

• G06F30/367 Primary

  Design verification, e.g. using simulation, simulation program with integrated circuit emphasis [SPICE], direct methods or relaxation methods · CPC title

• G06F30/27 Primary

  using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model · CPC title