SparkBeyond Ltd.Download PDFPatent Trials and Appeals BoardOct 29, 20212020002768 (P.T.A.B. Oct. 29, 2021) Copy Citation UNITED STATES PATENT AND TRADEMARK OFFICE UNITED STATES DEPARTMENT OF COMMERCE United States Patent and Trademark Office Address: COMMISSIONER FOR PATENTS P.O. Box 1450 Alexandria, Virginia 22313-1450 www.uspto.gov APPLICATION NO. FILING DATE FIRST NAMED INVENTOR ATTORNEY DOCKET NO. CONFIRMATION NO. 15/165,015 05/26/2016 Meir MAOR 66164 6298 67801 7590 10/29/2021 PRTSI Inc. 232 NW 42nd Ter Plantation, FL 33317 EXAMINER RIFKIN, BEN M ART UNIT PAPER NUMBER 2198 NOTIFICATION DATE DELIVERY MODE 10/29/2021 ELECTRONIC Please find below and/or attached an Office communication concerning this application or proceeding. The time period for reply, if any, is set in the attached communication. Notice of the Office communication was sent electronically on above-indicated "Notification Date" to the following e-mail address(es): usptomail@ipatent.co.il PTOL-90A (Rev. 04/07) UNITED STATES PATENT AND TRADEMARK OFFICE ____________________ BEFORE THE PATENT TRIAL AND APPEAL BOARD ____________________ Ex parte MEIR MAOR, RON KARIDI, SAGIE DAVIDOVICH, and AMIR RONEN ____________________ Appeal 2020-002768 Application 15/165,015 Technology Center 2100 ____________________ Before ELENI MANTIS MERCADER, JOHNNY A. KUMAR, and NORMAN H. BEAMER, Administrative Patent Judges. MANTIS MERCADER, Administrative Patent Judge. DECISION ON APPEAL STATEMENT OF THE CASE Pursuant to 35 U.S.C. § 134(a), Appellant1 appeals from the Examiner’s decision to reject claims 1, 2, 4–6, 8, 11–24, 26, 28, and 29. See Final Act. 1. We have jurisdiction under 35 U.S.C. § 6(b). We REVERSE. 1 We use the word “Appellant” to refer to “applicant” as defined in 37 C.F.R. § 1.42. The Appeal Brief identifies SparkBeyond Ltd., as the real party in interest. Appeal Br. 2. Appeal 2020-002768 Application 15/165,015 2 CLAIMED SUBJECT MATTER The claimed invention is directed to “systems and/or methods that automatically generate a set of pivotal classification features that process at least some data retrieved from initially unrelated secondary data sources and optionally from a primary dataset for use in an automated machine learning process.” See Spec. 8. Claim 1, reproduced below, is illustrative of the claimed subject matter: 1. A computer implemented method of computing a classifier that classifies data with an extracted set of classification features based on secondary datasets, comprising: identifying, for a first set of values of at least one primary field of a plurality of primary fields of a plurality of primary data instances of a primary training dataset each associated with a classification label, at least one second set of secondary fields of a plurality of unclassified second data instances of at least one secondary dataset, according to the first set of values matched to corresponding values in at least one respective secondary field according to a matching requirement; linking each respective matched value of the first set of values, to other secondary fields of at least one respective secondary data instance of the respective matched secondary field, wherein the primary training dataset is selected from the group consisting of: a multi-column table and a matrix of primary cells, wherein each row represents a respective data instance and each column represents a respective primary field storing at least one value in each primary cell, wherein at least one member of the at least one secondary dataset is selected from the group consisting of: a table of secondary cells wherein each row represents a secondary data instance and each column represents a secondary field and a graph comprising linked data, wherein the linking is performed by creating pointers that point between each respective matched primary cell of the primary Appeal 2020-002768 Application 15/165,015 3 dataset and matched secondary cells of the at least one secondary dataset; generating a set of classification features each including: (i) at least one condition selected from a plurality of conditions, (ii) a value selected from one of the cells or nodes storing the linked other secondary fields of the at least one respective secondary data instance of the respective matched secondary field, wherein the respective classification feature outputs a binary value computed by the at least one condition that compares between the value selected from the other linked secondary fields and a new received data instance; selecting a subset of classification features from the set of classification features according to a correlation requirement between the classification label of the primary data instance corresponding to the linked secondary cell used in the respective classification feature, and each respective classification feature; identifying a subset of secondary cells based on the secondary cells associated with each selected subset classification feature; and computing a classifier for classification of a new received data instance according to the linked subset of secondary cells and the primary training dataset used for computation of the subset of classification features on the new received data instance for classification of the new received data instance. REJECTIONS 1. Claims 1, 5, 6, 8, 11–20, 23, 24, 26, and 29 stand rejected under 35 U.S.C. § 103 as being unpatentable over Gupta (US 2009/0171956 A1; July 2, 2009), Srivastava (US 2002/0099519 A1; July 25, 2002), Geodesy Designs, “Lesson 6: An introduction to pointers,” https://web.archive.org/web/20080307093756/https://www.cprogram ming.com/tutorial/lesson6.html (2008), and Liddy (US 6,304,864 B1; Oct. 16, 2001). Final Act. 2. Appeal 2020-002768 Application 15/165,015 4 2. Claim 2 stands rejected under 35 U.S.C. § 103 as being unpatentable over Gupta, Srivastava, Geodesy Designs, Liddy, and Shivaji-Rao (US 2008/0183705 A1; July 31, 2008). Final Act. 17. 3. Claim 4 stands rejected under 35 U.S.C. § 103 as being unpatentable over Gupta, Srivastava, Geodesy Designs, Liddy, and Mei (US 2009/0274434 A1; Nov. 5, 2009). Final Act. 19. 4. Claim 21 stands rejected under 35 U.S.C. § 103 as being unpatentable over Gupta, Srivastava, Geodesy Designs, Liddy, and Rose (US 2015/0006443 A1; Jan. 1, 2015). Final Act. 20. 5. Claim 22 stands rejected under 35 U.S.C. § 103 as being unpatentable over Gupta, Srivastava, Geodesy Designs, Liddy, and Abdel-Hady (US 2014/0337005 A1; Nov. 13, 2014). Final Act. 21. 6. Claim 28 stands rejected under 35 U.S.C. § 103 as being unpatentable over Gupta, Srivastava, Geodesy Designs, Liddy, and Byrne (US 3,587,054; June 22, 1971). Final Act. 23. ANALYSIS Appellant argues inter alia that “the Examiner seems to insist that a POSITA [person of ordinary skill in the art] would replace the manner in which references are stored in Gupta with the manner in which they are stored in Srivastava[,] the question is whether a POSITA would be motivated to replace the method of storing data” (Reply Br. 5) (emphasis omitted). Appellant contends that “the Examiner failed to provide a proper motivation to combine the references as suggested by him” (Reply Br. 5). Appellant argues that the motivation to “allow the various training data of the Gupta reference to be stored in a well-known and common data Appeal 2020-002768 Application 15/165,015 5 structure” is irrelevant since Gupta uses the auxiliary datasets to enhance the input data, unlike Srivastava which uses the dataset to provide values to the parameters. Appeal Br. 22–23. We agree with Appellant’s argument. Our inquiry focuses on whether the Examiner provided “some articulated reasoning with some rational underpinning to support the legal conclusion of obviousness.” KSR Int’l Co. v. Teleflex Inc., 550 U.S. 398, 418 (2007). Regarding the combination of Gupta and Srivastava, the Examiner finds that [i]t would have been obvious to one skilled in the art of training data to combine the work of Gupta and Srivastava in order to make use of the common data structures[,] tables[,] and matrices to store data. The motivation for doing so would be to allow “the training data [to] be given in a table (or matrix) of data” ([Srivastava para 20]) or in the case of Gupta, allow the various training data of the Gupta reference to be stored in a well-known and common data structure. Final Act. 6 (emphasis added). In other words, the Examiner has cited Srivastara as teaching common data structures, tables, and matrices to store data, and used it as a motivation that the data structure is “well-known and common.” This statement of motivation is circular, and the Examiner does not find Srivastava provides a motivation, such as a benefit, for use of this data structure. “Absent some articulated rationale, a finding that a combination of prior art would have been ‘common sense’ or ‘intuitive’ is no different than merely stating the combination ‘would have been obvious.’” In re Van Os, 844 F.3d 1359, 1361 (Fed. Cir. 2017). Appeal 2020-002768 Application 15/165,015 6 The cited portion of Srivastava states that with reference to Figures 2 and 3, that the training data may be given in a table (or matrix) of data D of size (NxP), with each row corresponding to an observation and each column corresponding to a measurement. Srivastara para. 20. The cited paragraph does not describe a benefit nor does the Examiner articulate a benefit for the combination. Accordingly, we are constrained by the record before us to reverse the Examiner’s rejections of claim 1 and the rejections of claims 2, 4–6, 8, 11– 24, 26, 28, and 29 under 35 U.S.C. § 103. The additional references of Shivaji-Rao, Mei, Rose, Abdel-Hady, and Byrne do not cure the above cited deficiency. Appeal 2020-002768 Application 15/165,015 7 CONCLUSION Claim(s) Rejected 35 U.S.C. § Reference(s)/Basis Affirmed Reversed 1, 5, 6, 8, 11–20, 23, 24, 26, 29 103 Gupta, Srivastava, Geodesy Designs, Liddy 1, 5, 6, 8, 11–20, 23, 24, 26, 29 2 103 Gupta, Srivastava, Geodesy Designs, Liddy, Shivaji-Rao 2 4 103 Gupta, Srivastava, Geodesy Designs, Liddy, Mei 4 21 103 Gupta, Srivastava, Geodesy Designs, Liddy, Rose 21 22 103 Gupta, Srivastava, Geodesy Designs, Liddy, Abdel-Hady 22 28 103 Gupta, Srivastava, Geodesy Designs, Liddy, Byrne 28 Overall Outcome 1, 2, 4–6, 8, 11–24, 26, 28, 29 REVERSED Copy with citationCopy as parenthetical citation