Substrate Process Data Labeling for Operation Name Normalization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Inconsistent naming conventions and operation mappings in substrate process data sets hinder effective data analysis and machine learning model training, leading to inefficiencies and sub-optimal substrate production.
Innovation Solution
A software tool for bulk data labeling that normalizes operation names and mappings across data sets, enabling consistent labeling of substrate process data with common labels, linking process data to metrology data, and preparing it for data analysis and machine learning model training.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual labeling of substrate process data is performed, then data can be processed, but the process is time-consuming and inefficient
Solution Approach 1:
The system automatically performs data labeling by extracting operation names from process data and generating labels without human intervention. The processor autonomously identifies operations, determines operation names, and creates standardized labels, eliminating the need for manual labeling while maintaining data quality and consistency.
2Adaptability or versatility
If inconsistent naming conventions are used in process data sets, then data can be collected from multiple sources, but data analysis and machine learning model training are hindered
Solution Approach 1:
The system standardizes operation names across all process data by extracting and normalizing names to a consistent format. All operation names are transformed to follow a uniform naming convention, ensuring homogeneity in the data set while preserving the ability to collect data from diverse manufacturing sources with different original naming conventions.
3Adaptability or versatility
If operation mappings differ across data entries, then processes can be represented flexibly, but data normalization and common labeling become difficult
Solution Approach 1:
Instead of attempting to map various operation names to a predefined standard taxonomy, the system inverts the approach by extracting the operation name directly from each data entry and using that extracted name as the standardized label. This eliminates the need for complex mapping relationships while achieving normalization across diverse data sources.
Data Source
AI summary
A method includes obtaining a first plurality of data entries comprising process data of one or more processes performed on a plurality of substrates. The method further includes determining an operation of interest from the first plurality of data entries. The method further includes updating a first subset of data entries of the first plurality of data entries by normalizing the operation of interest across the first plurality of data entries. The method further includes labeling the first subset of data entries with a common label. The method further includes obtaining a second plurality of data entries comprising metrology data. The method further includes linking the process data of the updated first subset of data entries to the metrology data. The method further includes preparing the updated first subset of data entries for one or more data analysis operations based on the common label.