Inverted ML Models for Manufacturing Input Prediction

Solution Overview

Problem

Traditional manufacturing process modeling methods are inefficient as they require numerous experiments and simulations to identify optimum input parameters, especially with increasing complexity and multiple interrelated steps, which leads to resource wastage and prolonged process engineering time.

Innovation Solution

The use of a set of machine learning models, including inverted models, that predict manufacturing inputs based on expected outputs, reducing the need for physical experiments and simulations by combining candidate solutions from multiple models to achieve target attributes, thereby enhancing interpretability and reducing resource consumption.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If traditional manufacturing process modeling methods are used to identify optimum input parameters, then the process can be optimized, but numerous experiments and simulations are required leading to resource wastage and prolonged process engineering time

Engineering Contradiction:

Improveoptimum input parameters identificationVSAvoidprocess engineering time

Core Design Contradiction:

Manufacturing precisionVSLoss of time

Solution Approach 1:

The patent applies preliminary action by pre-training multiple machine learning models with different architectures and hyperparameters before the actual manufacturing process optimization. These pre-trained models are then used to generate candidate input parameters, eliminating the need for time-consuming physical experiments and simulations during the actual optimization process.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent creates virtual copies of the manufacturing process through multiple machine learning models that simulate different aspects of the process. Instead of performing physical experiments, the system uses these digital twins (models) to predict outcomes and identify optimal parameters, significantly reducing resource consumption and time.

Inventive Principle:

Principle #26Copying

2Manufacturing precision

If traditional manufacturing process modeling methods are used to identify optimum input parameters, then the process can be optimized, but numerous experiments and simulations are required leading to resource wastage

Engineering Contradiction:

Improveoptimum input parameters identificationVSAvoidresource consumption

Core Design Contradiction:

Manufacturing precisionVSLoss of energy

Solution Approach 1:

The patent creates virtual copies of the manufacturing process through multiple machine learning models that simulate different aspects of the process. Instead of performing physical experiments, the system uses these digital twins (models) to predict outcomes and identify optimal parameters, significantly reducing resource consumption and time.

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent replaces the mechanical system of physical experiments and simulations with an information-based system using machine learning models. The models process data and generate predictions computationally, substituting physical resource consumption with computational processing that is far more efficient.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Reliability

If multiple machine learning models are used to predict manufacturing inputs, then the accuracy and robustness improve, but the model complexity increases

Engineering Contradiction:

Improveprediction accuracyVSAvoidmodel complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the manufacturing process modeling into multiple specialized machine learning models, each trained on different aspects or subsets of the process data. This segmentation allows each model to focus on specific patterns, improving overall prediction reliability while maintaining manageable individual model complexities.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent merges the predictions from multiple machine learning models by combining their candidate input parameter suggestions. This combination approach leverages the strengths of different models, producing more robust and accurate predictions while distributing the computational complexity across multiple simpler models rather than one complex model.

Inventive Principle:

Principle #5Merging (Combining)

Data Source

PatentUS11829873B2Predictive modeling of a manufacturing process using a set of trained inverted models

Publication Date: 2023.11.28 APPLIED MATERIALS INC

AI summary

Disclosed herein is technology for performing predictive modeling to identify inputs for a manufacturing process. An example method may include receiving expected output data for a manufacturing process, wherein the expected output data defines an attribute of an output of the manufacturing process; accessing a plurality of machine learning models that model the manufacturing process; determining, using a first machine learning model, input data for the manufacturing process based on the expected output data for the manufacturing process, wherein the input data comprises a value for a first input and a value for a second input; combining the input data determined using the first machine learning model with input data determined using the second machine learning model to produce a set of inputs for the manufacturing process, wherein the set of inputs comprises candidate values for the first input and candidate values for the second input.