View source for 17825087. COMPUTING DEVICES FOR PREDICTING ELECTRICAL TESTS, ELECTRICAL TEST PREDICTION APPARATUSES HAVING THE SAME, AND OPERATING METHODS THEREOF simplified abstract (Samsung Electronics Co., Ltd.)

=COMPUTING DEVICES FOR PREDICTING ELECTRICAL TESTS, ELECTRICAL TEST PREDICTION APPARATUSES HAVING THE SAME, AND OPERATING METHODS THEREOF=


==Organization Name==

[[:Category:Samsung Electronics Co., Ltd.|Samsung Electronics Co., Ltd.]]


[[Category:Samsung Electronics Co., Ltd.]]

==Inventor(s)==

[[:Category:Seyoung Park of Hwaseong-si (KR)|Seyoung Park of Hwaseong-si (KR)]][[Category:Seyoung Park of Hwaseong-si (KR)]]

[[:Category:Gwangnae Gil of Yongin-s (KR)|Gwangnae Gil of Yongin-s (KR)]][[Category:Gwangnae Gil of Yongin-s (KR)]]

[[:Category:Sola Woo of Suwon-si (KR)|Sola Woo of Suwon-si (KR)]][[Category:Sola Woo of Suwon-si (KR)]]

[[:Category:Jonghyun Lee of Hwaseong-si (KR)|Jonghyun Lee of Hwaseong-si (KR)]][[Category:Jonghyun Lee of Hwaseong-si (KR)]]

==COMPUTING DEVICES FOR PREDICTING ELECTRICAL TESTS, ELECTRICAL TEST PREDICTION APPARATUSES HAVING THE SAME, AND OPERATING METHODS THEREOF - A simplified explanation of the abstract==


This abstract first appeared for US patent application 17825087 titled 'COMPUTING DEVICES FOR PREDICTING ELECTRICAL TESTS, ELECTRICAL TEST PREDICTION APPARATUSES HAVING THE SAME, AND OPERATING METHODS THEREOF

==Simplified Explanation==

==Abstract==
A method is disclosed for operating an electrical test prediction apparatus. The method involves determining a relationship between first electrical test (ET) data and electrical die sorting (EDS) data. The first ET data corresponds to a subset of semiconductor chips on a wafer, while the EDS data is obtained by measuring the state of each chip on the wafer using a testing device. The method further includes predicting second ET data for a region of the wafer other than the subset by performing machine learning on the determined relationship.

==Bullet Points==
* The patent application describes a method for predicting electrical test data for a region of a wafer using machine learning.
* The method involves determining a relationship between first electrical test data and electrical die sorting data.
* The first electrical test data corresponds to a subset of semiconductor chips on the wafer, while the electrical die sorting data is obtained by measuring the state of each chip on the wafer.
* By performing machine learning on the determined relationship, the method predicts second electrical test data for a region of the wafer other than the subset.

==Potential Applications==
* Semiconductor manufacturing: The method can be applied in the semiconductor industry to predict electrical test data for regions of a wafer that have not been directly tested. This can help in optimizing manufacturing processes and improving yield.
* Quality control: The predicted electrical test data can be used for quality control purposes, allowing manufacturers to identify and address potential issues in the production of semiconductor chips.
* Process optimization: By analyzing the relationship between electrical test data and electrical die sorting data, the method can provide insights into the manufacturing process and help optimize it for improved efficiency and performance.

==Problems Solved==
* Limited testing coverage: Traditional testing methods may only cover a subset of semiconductor chips on a wafer, leaving other regions untested. The method described in the patent application solves this problem by predicting electrical test data for the untested regions.
* Time and cost efficiency: Predicting electrical test data using machine learning can save time and cost compared to physically testing each chip on the wafer individually.
* Data analysis and interpretation: The method addresses the challenge of analyzing and interpreting the relationship between electrical test data and electrical die sorting data, allowing for accurate predictions.

==Benefits==
* Improved yield: By predicting electrical test data for untested regions, manufacturers can identify and address potential issues early on, leading to improved yield and reduced waste.
* Cost savings: Predicting electrical test data using machine learning can save costs associated with physical testing equipment and resources.
* Process optimization: The method provides insights into the relationship between electrical test data and electrical die sorting data, enabling manufacturers to optimize their processes for better efficiency and performance.


==Original Abstract Submitted==

A method of operating an electrical test prediction apparatus includes determining a relationship between first electrical test (ET) data, corresponding to at least one shot region comprising a subset of a plurality of semiconductor chips of a wafer, and electrical die sorting (EDS) data, obtained by measuring a state of each chip on the wafer by a testing device, and predicting second ET data, corresponding to an region of the wafer other than the at least one shot region by performing machine learning on the relationship.



[[Category:G01R31/26]]
[[Category:G06N20/00]]