AUTOMATIC ELIMINATION OF NOISE FOR BIG DATA ANALYTICS

• US 20190325321A1
• Filed: 06/29/2019
• Published: 10/24/2019
• Est. Priority Date: 06/29/2019
• Status: Active Grant

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: Swanson, John A., Singh, Vivek K., Sastry, Kumara, Parks, Helen F., Chen, I-Tzu
• Granted Patent: US 11,663,700 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 17/18 for evaluating statistical data , e.g. average values, frequency distributions, probability functions, regression analysis forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04
• CPC Class Codes: G06F 18/23 Clustering techniques
• CPC Class Codes: G06F 18/24 Classification techniques
• CPC Class Codes: G06F 18/2433 Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 20/20 Ensemble learning
• CPC Class Codes: G06N 3/126 Evolutionary algorithms, e.g. genetic algorithms or genetic programming
• CPC Class Codes: G06N 5/025 Extracting rules from data
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks
• CPC Class Codes: G06T 2207/10061 from scanning electron microscope
• CPC Class Codes: G06T 2207/30148 Semiconductor; IC; Wafer
• CPC Class Codes: G06T 5/70 Denoising; Smoothing
• CPC Class Codes: G06V 10/40 Extraction of image or video features
• CPC Class Codes: G06V 10/762 using clustering, e.g. of similar faces in social networks
• CPC Class Codes: G06V 10/764 using classification, e.g. of video objects
• CPC Class Codes: G06V 2201/06 Recognition of objects for industrial automation

Abstract

A method comprising identifying a set of target features for a plurality of data instances of an input data collection; determining feature values for the set of target features for the plurality of data instances; identifying a plurality of outlier data instances based on the determined feature values; identifying a plurality of noisy data instances from the outlier data instances based on feature values of the plurality of noisy data instances, wherein a noisy data instance is identified based on a determination that noise is present in noisy data instance; and providing an indication of the plurality of noisy data instances.

Claims

1. At least one machine readable storage medium having instructions stored thereon, the instructions when executed by a machine to cause the machine to:

identify a set of target features for a plurality of data instances of an input data collection;

determine feature values for the set of target features for the plurality of data instances;

identify a plurality of outlier data instances based on the determined feature values;

identify a plurality of noisy data instances from the outlier data instances based on feature values of the plurality of noisy data instances, wherein a noisy data instance is identified based on a determination that noise is present in noisy data instance; and

provide an indication of the plurality of noisy data instances.

2. The at least one medium of claim 1, the instructions when executed to cause the machine to:

determine a class of the input data collection; and

select the set of target features from a plurality of target features based on the class of the input data collection.

3. The at least one medium of claim 2, wherein the plurality of data instances are images of a processing layer of at least one semiconductor chip of at least one wafer of a process technology and wherein the class of the data instance corresponds to the processing layer and the process technology.

4. The at least one medium of claim 2, the instructions when executed to cause the machine to determine the class of the input data collection based on feature values determined for the plurality of target features for a subset of the plurality of data instances of the input data collection.

5. The at least one medium of claim 1, wherein identifying the plurality of outlier data instances based on the determined feature values comprises clustering the plurality data instances of the input data collection based on the determined feature values.

6. The at least one medium of claim 1, the instructions when executed to cause the machine to:

utilize a first model to generate a first prediction indicative of whether a first one of the outlier data instances should be classified as a noisy data instance;

utilize a second model to generate a second prediction indicative of whether the first one of the outlier data instances should be classified as a noisy data instance; and

determine whether the first one of the outlier data instances is a noisy data instance based on the first prediction and the second prediction.

7. The at least one medium of claim 6, wherein the first model is a random forest model and the second model is a Bayesian optimization model.

8. The at least one medium of claim 7, wherein the Bayesian optimization model is trained with a hierarchical mixture of Gaussian models coupled with Gaussian outlier detection.

9. The at least one medium of claim 7, wherein the random forest model is optimized using a multi-modal asynchronous genetic algorithm.

10. The at least one medium of claim 1, wherein the set of target features comprises one or more of alignment distance, alignment quality, a Laplacian of Gaussian-based blob measure, image entropy, merged contours, broken contours, unexpected contours, contour width distribution, contour height distribution, contour area distribution, local or global image intensity histogram, local or global image contrast variation, line-edge roughness, and line-width roughness.

11. A method comprising:

identifying a set of target features for a plurality of data instances of an input data collection;

determining feature values for the set of target features for the plurality of data instances;

identifying a plurality of outlier data instances based on the determined feature values;

identifying a plurality of noisy data instances from the outlier data instances based on feature values of the plurality of noisy data instances, wherein a noisy data instance is identified based on a determination that noise is present in noisy data instance; and

providing an indication of the plurality of noisy data instances.

12. The method of claim 11, further comprising:

determining a class of the input data collection; and

selecting the set of target features from a plurality of target features based on the class of the input data collection.

13. The method of claim 12, wherein the plurality of data instances are images of a processing layer of at least one semiconductor chip of at least one wafer of a process technology and wherein the class of the data instance corresponds to the processing layer and the process technology.

14. The method of claim 12, further comprising determining the class of the input data collection based on feature values determined for the plurality of target features for a subset of the plurality of data instances of the input data collection.

15. The method of claim 11, wherein identifying the plurality of outlier data instances based on the determined feature values comprises clustering the plurality data instances of the input data collection based on the determined feature values.

16. An apparatus comprising:

a memory to store an input data collection comprising a plurality of data instances; and

a processor coupled to the memory, the processor to;

identify a set of target features for the plurality of data instances of the input data collection;

determine feature values for the set of target features for the plurality of data instances;

identify a plurality of outlier data instances based on the determined feature values;

identify a plurality of noisy data instances from the outlier data instances based on feature values of the plurality of noisy data instances, wherein a noisy data instance is identified based on a determination that noise is present in noisy data instance; and

provide an indication of the plurality of noisy data instances.

17. The apparatus of claim 16, the processor to:

determine a class of the input data collection; and

select the set of target features from a plurality of target features based on the class of the input data collection.

18. The apparatus of claim 17, wherein the plurality of data instances are images of a processing layer of at least one semiconductor chip of at least one wafer of a process technology and wherein the class of the data instance corresponds to the processing layer and the process technology.

19. The apparatus of claim 16, further comprising an imaging tool to capture the data instances of the input data collection.

20. The apparatus of claim 16, further comprising one or more of a battery communicatively coupled to the processor, a display communicatively coupled to the processor, or a network interface communicatively coupled to the processor.