US 12,074,073 B2
Prescriptive analytics in highly collinear response space
Sidharth Bhatia, Santa Cruz, CA (US); Jie Feng, Milpitas, CA (US); and Dermot Cantwell, Sunnyvale, CA (US)
Assigned to Applied Materials, Inc., Santa Clara, CA (US)
Filed by Applied Materials, Inc., Santa Clara, CA (US)
Filed on May 6, 2022, as Appl. No. 17/738,937.
Application 17/738,937 is a continuation of application No. 16/599,962, filed on Oct. 11, 2019, granted, now 11,328,964.
Claims priority of provisional application 62/779,097, filed on Dec. 13, 2018.
Prior Publication US 2022/0328366 A1, Oct. 13, 2022
Int. Cl. G05B 13/04 (2006.01); G06N 3/08 (2023.01); G06Q 50/04 (2012.01); H01L 21/02 (2006.01); H01L 21/66 (2006.01); H01L 21/67 (2006.01); G01B 7/06 (2006.01); G01N 21/41 (2006.01); G06T 7/00 (2017.01); G11C 11/16 (2006.01); H01J 37/32 (2006.01)
CPC H01L 22/12 (2013.01) [G06N 3/08 (2013.01); G06Q 50/04 (2013.01); H01L 21/02 (2013.01); H01L 21/67253 (2013.01); G01B 7/105 (2013.01); G01N 21/41 (2013.01); G06T 7/0004 (2013.01); G11C 11/1675 (2013.01); H01J 37/32458 (2013.01); Y02P 90/30 (2015.11)] 17 Claims
OG exemplary drawing
 
1. A method comprising:
identifying first film property data of first substrates produced by manufacturing equipment based on manufacturing parameters;
selecting, by a processing device, a set of data points of the first film property data that are associated with a first line that is substantially orthogonal to a second line associated with target film property data, the target film property data being provided to a trained machine learning model that was trained using input data comprising historical manufacturing parameters associated with producing historical substrates and target output comprising historical film property data of the historical substrates, wherein the selecting of the set of data points is via the trained machine learning model; and
determining, based on the set of data points, updated manufacturing parameters, wherein second substrates are to be produced by the manufacturing equipment based on the updated manufacturing parameters to meet the target film property data, the method further comprising obtaining an inverted solution comprising one or more outputs from the trained machine learning model, wherein the determining of the updated manufacturing parameters comprises extracting, from the one or more outputs, the updated manufacturing parameters.