Computer-implemented method for defect analysis, computer-implemented method of evaluating likelihood of defect occurrence, apparatus for defect analysis, computer-program product, and intelligent defect analysis system

Patent metadata

Field	Value
Publication number	US-12032364-B2
Application number	US-202017438719-A
Country	US
Kind code	B2
Filing date	Dec 3, 2020
Priority date	Dec 3, 2020
Publication date	Jul 9, 2024
Grant date	Jul 9, 2024

Abstract

A computer-implemented method for defect analysis is provided. The computer-implemented method includes calculating a plurality of weight-of-evidence (WOE) scores respectively for a plurality of device operations with respect to detects occurred during a fabrication period, a higher WOE score indicating a higher correlation between a defect and a device operation; and ranking the plurality of WOE scores to obtain a list of selected device operations highly correlated with the defects occurred during the fabrication period, device operations in the list of selected device operations having a WOE score greater than a first threshold score. A respective one of the plurality of device operations is a respective device defined by a respective operation site at which the respective device perform a respective operation.

First claim

What is claimed is:

1. An intelligent defect analysis system, comprising:
a distributed computing system comprising one or more networked computers configured to execute in parallel to perform at least one common task;
one or more computer readable storage mediums storing instructions that, when executed by the distributed computing system, cause the distributed computing system to execute software modules;
wherein the software modules comprise:
a data manager configured to store data, and intelligently extract, transform, or load the data;
a query engine connected to the data manager and configured to query the data directly from the data manager;
an analyzer connected to the query engine and configured to perform defect analysis upon received a task request, the analyzer comprising a plurality of business servers and a plurality of algorithm servers, the plurality of algorithm servers configured to query the data directly from the data manager; and
a data visualization and interaction interface configured to generate the task requests;
wherein the data visualization and interaction interface comprises an automatic task sub-interface configured to receive input of the recurring period for which the defect analysis is to be performed;
wherein the task request is an interactive task request;
the data visualization and interaction interface is configured to receive a user-defined analysis criteria, and configured to generate the interactive task request based on the user-defined analysis criteria;
upon receiving information on defects of interest, the one or more of the plurality of business servers are configured to transmit the information to the data visualization and interaction interface;
the data visualization and interaction interface is configured to display the information on defects of interest and a plurality of environmental factors associated with the defects of interest, and configured to receive a user-defined selection of one or more environmental factors from the plurality of environmental factors, and transmit the user-defined selection to the one or more of the plurality of business servers; and
the one or more of the plurality of business servers are configured to generate the defect analysis tasks based on the information and the user-defined selection;
wherein one or more of the plurality of algorithm servers is configured to perform: a computer-implemented method of evaluating likelihood of defect occurrence, comprising:
obtaining original data on defects occurred during a fabrication period;
pre-processing the original data to obtain pre-processed data;
extracting features from the pre-processed data to obtain first features;
selecting second features from the first features;
inputting the second features into a predictive model; and
evaluating likelihood of defect occurrence;
wherein extracting features from the pre-processed data comprises performing at least one of time domain analysis and frequency domain analysis on the pre-processed data;
wherein the predictive model is trained by:
obtaining training original data on defects occurred during a training fabrication period;
pre-processing the training original data to obtain pre-processed training data;
extracting features from the pre-processed training data to obtain third features;
selecting fourth features from the third features; and
tuning parameters of an initial model using the third features to obtain the predictive model for defect prediction;
wherein extracting training features from the pre-processed training data comprises performing at least one of time domain analysis and frequency domain analysis on the pre-processed training data;
wherein tuning parameters of the initial model comprises evaluating a F measure according to Equation (1):

1

F
β

=

1

1
+

β
2

⁢

(

1
P

+

β
2

R

)

;

wherein Fβ stands for a harmonic mean of precision and recall, P stands for precision, R stands for recall, β stands for a parameter that controls a balance between P and R.

2. The intelligent defect analysis system of claim 1 , wherein the data manager comprises an extract, transform, load (ETL) module configured to extract, transform, or load data from the plurality of data sources onto a data mart that is a database of non-structured query language (No-SQL) type NoSQL type; and
upon receiving an assigned task, a respective one of the plurality of algorithm servers is configured to obtain a first data from the data mart.

3. The intelligent defect analysis system of claim 2 , wherein the ETL module is further configured to extract, transform, or load data from the plurality of data sources onto a general data layer that is a distributed data storage storing information;
upon performing defect analysis, the respective one of the plurality of algorithm servers is configured to transmit a second data to the general data layer;
the ETL module is configured to generate a dynamically updated table that is automatically updated periodically; and
the general data layer is configured to store the dynamically updated table.

4. The intelligent defect analysis system of claim 3 , wherein the software modules further comprise a query engine connected to the data manager and configured to obtain the managed data from the data manager.

5. The intelligent defect analysis system of claim 4 , wherein
the analyzer further comprises a plurality of business servers;
upon receiving the task request, the one or more of the plurality of business servers are configured to transmit a query task request to the query engine;
the query engine, upon receiving the query task request from the one or more of the plurality of business servers, is configured to query the dynamically updated table to obtain information on defects of interest, and transmit the information on defects of interest to one or more of the plurality of business servers;
upon receiving defect analysis tasks, the one or more of the plurality of algorithm servers are configured to obtain the first data from the data mart to perform defect analysis; and
upon completion of the defect analysis, the one or more of the plurality of algorithm servers are configured to transmit results of the defect analysis to the general data layer.

6. The intelligent defect analysis system of claim 5 , wherein the task request is an automatically recurring task request, the automatically recurring task request defining a periodical recurring period for which the defect analysis is to be performed;
the query engine is configured to query the dynamically updated table to obtain the information on defects of interest limited to the periodical recurring period; and
upon receiving the information on defects of interest during the periodical recurring

Assignees

• Boe Technology Group Co Ltd

Inventors

• Chai Dong
• Lan Tian
• Wu Haohan
• Tang Yue
• Shen Guoliang
• Yuan Fei

Classifications

• H10P74/23

  characterised by multiple measurements, corrections, marking or sorting processes · CPC title

• G06F18/214

  Generating training patterns; Bootstrap methods, e.g. bagging or boosting · CPC title

• G05B19/4183

  characterised by data acquisition, e.g. workpiece identification · CPC title

• G06F30/398

  Design verification or optimisation, e.g. using design rule check [DRC], layout versus schematics [LVS] or finite element methods [FEM] (optical proximity correction [OPC] design processes G03F1/36) · CPC title

• G05B19/41875 Primary

  characterised by quality surveillance of production · CPC title