Semiconductor Process Modeling Tensor Data Preprocessing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In semiconductor manufacturing, raw data from sensors often has omitted elements due to different measurement regulations, leading to low accuracy in process modeling.
Innovation Solution
A semiconductor process modeling system that includes a preprocessing component to generate tensor data from raw data, modifying it based on equipment and process parameter characteristics, thereby reducing the number of omitted elements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If raw data is used directly for modeling without preprocessing, then the modeling process is simple, but the model accuracy is low due to omitted elements
Solution Approach 1:
The patent applies preliminary action by performing data preprocessing before modeling. The preprocessing component fills omitted elements in raw data using imputation techniques (mean, median, mode) and transforms the data into tensor format with proper dimensional structure. This preliminary data preparation ensures complete input data for the modeling component, thereby improving model accuracy without adding excessive complexity during the modeling phase.
2Quantity of substance
If multiple sensors with different measurement regulations are used, then more process parameters can be measured, but more elements are omitted in the raw data matrix
Solution Approach 1:
The patent applies parameter changes by transforming the raw data parameters into a standardized tensor format. The preprocessing component restructures the data with dimensions representing different sensors, measurement points, and time points, and fills omitted parameter values using statistical imputation. This transformation maintains the quantity of process parameters while reducing information loss by systematically handling missing values.
3Productivity
If raw data with omitted elements is used for modeling, then less data preprocessing is required, but the prediction accuracy of process results deteriorates
Solution Approach 1:
The patent implements preliminary action by performing comprehensive data preprocessing including filling omitted elements and transforming data into tensor format before the modeling stage. This ensures that the modeling component receives complete, structured data, thereby maintaining high prediction accuracy. The preprocessing is automated and integrated into the workflow, minimizing the trade-off between efficiency and accuracy.
Data Source
AI summary
Provided is a semiconductor process modeling system. The semiconductor process modeling system includes a preprocessing component configured to generate tensor data from raw data obtained from semiconductor manufacturing equipment, wherein, when the raw data is expressed as a raw matrix representing values of a plurality of process parameters for each of a plurality of wafers, at least one element of the raw matrix is omitted, when the tensor data is expressed as a tensor matrix representing values of a plurality of preprocessed process parameters for each of the plurality of wafers, the number of omitted elements of the tensor matrix is less than the number of omitted elements of the raw matrix, and the preprocessing component is configured to generate the tensor data by modifying the raw data based on at least one of characteristics of the semiconductor manufacturing equipment and characteristics of the plurality of process parameters.