Jean-Baptiste Tristan et al.Download PDFPatent Trials and Appeals BoardMay 13, 20202018008621 (P.T.A.B. May. 13, 2020) 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. 14/599,272 01/16/2015 Jean-Baptiste Tristan 50277-4632 5349 42425 7590 05/13/2020 HICKMAN PALERMO BECKER BINGHAM/ORACLE 1 Almaden Boulevard Floor 12 SAN JOSE, CA 95113 EXAMINER VINCENT, DAVID ROBERT ART UNIT PAPER NUMBER 2123 NOTIFICATION DATE DELIVERY MODE 05/13/2020 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): usdocket@h35g.com PTOL-90A (Rev. 04/07) UNITED STATES PATENT AND TRADEMARK OFFICE ____________ BEFORE THE PATENT TRIAL AND APPEAL BOARD ____________ Ex parte JEAN-BAPTISTE TRISTAN and GUY STEELE ____________ Appeal 2018-008621 Application 14/599,272 Technology Center 2100 ____________ Before JOSEPH L. DIXON, BETH Z. SHAW, and STEVEN M. AMUNDSON, Administrative Patent Judges. AMUNDSON, Administrative Patent Judge. DECISION ON APPEAL Appellant1 seeks our review under 35 U.S.C. § 134(a) from a final rejection of claims 1–19, i.e., all pending claims. We have jurisdiction under 35 U.S.C. § 6(b). We affirm. 1 We use the word “Appellant” to refer to “applicant” as defined in 37 C.F.R. § 1.42 (2017). Appellant identifies the real party in interest as Oracle International Corporation. Appeal Br. 2. Appeal 2018-008621 Application 14/599,272 2 STATEMENT OF THE CASE The Invention According to the Specification, the “invention relates to sampling from a statistical model, and, more specifically, to sampling from a latent Dirichlet allocation (LDA) model using a partially-collapsed Gibbs sampler.” Spec. ¶ 1.2 The Specification explains that “[i]n the context of Gibbs sampling, a fully-collapsed Gibbs sampler can be used to efficiently identify parameter values in that a fully-collapsed Gibbs sampler identifies parameter values with a relatively small number of iterations.” Id. ¶ 3. But the Specification also explains that inefficiencies result with certain computer architectures because the “processing of a fully-collapsed Gibbs sampler is sequential” and “not efficiently implemented on highly parallel architectures, such as a graphics processing unit (GPU).” Id. Hence, the invention provides a Markov-Chain Monte Carlo algorithm “based on a highly parallel partially-collapsed Gibbs sampler” for estimating the parameters of an LDA topic model. Spec. ¶ 19, Abstract; see id. ¶¶ 31, 43–44, 53–54. According to the Specification, the invention exploits “massive amounts of parallelism by processing the algorithm on a highly-parallel architecture, such as a GPU,” to improve upon conventional algorithms. Id. ¶ 19, Abstract. More specifically, the Specification states that “[u]sing a GPU to perform the calculations results in faster convergence 2 This decision uses the following abbreviations: “Spec.” for the Specification, filed January 16, 2015; “Final Act.” for the Final Office Action, mailed November 15, 2017; “Advisory Act.” for the Advisory Action, mailed January 24, 2018; “Appeal Br.” for the Appeal Brief, filed May 15, 2018; “Ans.” for the Examiner’s Answer, mailed July 25, 2018; and “Reply Br.” for the Reply Brief, filed August 30, 2018. Appeal 2018-008621 Application 14/599,272 3 because calculations for each document” in a document set occur “in parallel, thus taking advantage of the high level of parallelism in the GPU architecture.” Id. ¶ 76. Exemplary Claims Independent claims 1 and 8 exemplify the claims at issue and read as follows (with formatting added for clarity): 1. A method comprising: generating a topic model that identifies sets of correlated words by: receiving information for a set of documents; wherein the set of documents comprises a plurality of words; running a partially-collapsed Gibbs sampler over a Dirichlet distribution of the plurality of words in the set of documents to produce sampler result data, further comprising: calculating a mean of the Dirichlet distribution; and determining, from the sampler result data, one or more sets of correlated words; wherein the topic model comprises the one or more sets of correlated words; and analyzing text within one or more documents, other than the set of documents, based at least in part on the topic model; wherein the method is performed by one or more computing devices. 8. One or more computer-readable media storing instructions which, when executed by one or more processors, cause performance of: generating a topic model that identifies sets of correlated words by: Appeal 2018-008621 Application 14/599,272 4 receiving information for a set of documents; wherein the set of documents comprises a plurality of words; running a partially-collapsed Gibbs sampler over a Dirichlet distribution of the plurality of words in the set of documents to produce sampler result data, further comprising: calculating a mean of the Dirichlet distribution; determining, from the sampler result data, one or more sets of correlated words; wherein the topic model comprises the one or more sets of correlated words; and analyzing text within one or more documents, other than the set of documents, based at least in part on the topic model. Appeal Br. 17–18 (Claims App.). The Rejection on Appeal Claims 1–19 stand rejected under 35 U.S.C. § 101 as directed to patent-ineligible subject matter. Final Act. 2–8; Ans. 3–5. ANALYSIS We have reviewed the rejection in light of Appellant’s arguments that the Examiner erred. For the reasons explained below, we agree with the Examiner’s conclusion concerning ineligibility under § 101. We adopt the Examiner’s findings and reasoning in the Final Office Action, Advisory Action, and Answer insofar as consistent with the following analysis. See Final Act. 2–8; Advisory Act. 1; Ans. 3–8. We add the following analysis to address and emphasize specific findings and arguments. Appeal 2018-008621 Application 14/599,272 5 Introduction The Patent Act defines patent-eligible subject matter broadly: “Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.” 35 U.S.C. § 101. “Those four categories define the explicit scope and reach of subject matter patentable.” In re Nuijten, 500 F.3d 1346, 1357 (Fed. Cir. 2007). “For all categories except process claims, the eligible subject matter must exist in some physical or tangible form.” Digitech Image Techs., LLC v. Elecs. for Imaging, Inc., 758 F.3d 1344, 1348 (Fed. Cir. 2014). “A transitory, propagating signal . . . is not a ‘process, machine, manufacture, or composition of matter.’” Nuijten, 500 F.3d at 1357. Moreover, in Mayo and Alice, the Supreme Court explained that § 101 “contains an important implicit exception” for laws of nature, natural phenomena, and abstract ideas. Mayo Collaborative Servs. v. Prometheus Labs., Inc., 566 U.S. 66, 70 (2012); Alice Corp. v. CLS Bank Int’l, 573 U.S. 208, 216 (2014); see Diamond v. Diehr, 450 U.S. 175, 185 (1981). In Mayo and Alice, the Court set forth a two-step analytical framework for evaluating patent-eligible subject matter. Mayo, 566 U.S. at 77–80; Alice, 573 U.S. at 217–18. Under Mayo/Alice step one, we “determine whether the claims at issue are directed to” a judicial exception, i.e., an abstract idea, a law of nature, or a natural phenomenon. Alice, 573 U.S. at 217. Step one involves looking at the “focus” of the claims at issue and their “character as a whole.” SAP Am., Inc. v. InvestPic, LLC, 898 F.3d 1161, 1167 (Fed. Cir. 2018). Appeal 2018-008621 Application 14/599,272 6 In January 2019, the PTO issued revised guidance for determining whether claims are directed to a judicial exception. See 2019 Revised Patent Subject Matter Eligibility Guidance, 84 Fed. Reg. 50 (Jan. 7, 2019) (“2019 Guidance”). The 2019 Guidance applies to the Board. Id. at 50–51, 57 n.42; see 35 U.S.C. § 3(a)(2)(A) (investing the Director with responsibility “for providing policy direction” for the PTO). Neither the Examiner nor Appellant had the benefit of the 2019 Guidance when presenting their respective positions concerning subject-matter eligibility. The 2019 Guidance specifies two prongs for the analysis under Mayo/Alice step one (PTO step 2A). 84 Fed. Reg. at 54–55. Prong one requires evaluating “whether the claim recites a judicial exception, i.e., an abstract idea, a law of nature, or a natural phenomenon.” Id. at 54. “If the claim does not recite a judicial exception, it is not directed to a judicial exception,” and it satisfies § 101. Id. “If the claim does recite a judicial exception, then it requires further analysis” under prong two. Id. Prong two requires evaluating “whether the claim as a whole integrates the recited judicial exception into a practical application of the exception.” Id. “When the exception is so integrated, then the claim is not directed to a judicial exception,” and it satisfies § 101. Id. “If the additional elements do not integrate the exception into a practical application, then the claim is directed to the judicial exception,” and it “requires further analysis” under Mayo/Alice step two (PTO step 2B). Id. Under Mayo/Alice step two, we “consider the elements of each claim both individually and ‘as an ordered combination’ to determine whether the additional elements” add enough to transform the “nature of the claim” into “significantly more” than the judicial exception. Alice, 573 U.S. at 217–18, Appeal 2018-008621 Application 14/599,272 7 221–22 (quoting Mayo, 566 U.S. at 78–79). Step two involves the search for an “inventive concept.” Alice, 573 U.S. at 217–18, 221; Univ. of Fla. Research Found., Inc. v. Gen. Elec. Co., 916 F.3d 1363, 1366 (Fed. Cir. 2019). “[A]n inventive concept must be evident in the claims.” RecogniCorp, LLC v. Nintendo Co., 855 F.3d 1322, 1327 (Fed. Cir. 2017). Nonstatutory Subject Matter: PTO Step 1 for Claims 8–14 In the Answer, the Examiner determines that independent claim 8 and dependent claims 9–14 encompass “transitory propagating signals” and, therefore, do not fall within one of § 101’s four categories. Ans. 3–4. The Examiner explains that the “broadest reasonable interpretation of a claim drawn to a computer readable medium” encompasses “transitory propagating signals per se in view of the ordinary and customary meaning of computer readable media.” Id. at 4. Hence, the Examiner concludes that claims 8–14 encompass nonstatutory subject matter. Id. In the Reply Brief, Appellant does not contest the nonstatutory- subject-matter rejection. See Reply Br. 1–2. Rather, Appellant “request[s] that this rejection be held in abeyance until conclusion of the appeal.” Id. at 1. Because Appellant does not contest the nonstatutory-subject-matter rejection, we summarily sustain that rejection. See In re Berger, 279 F.3d 975, 980, 984–85 (Fed. Cir. 2002) (holding that the Board did not err in sustaining an indefiniteness rejection when the appellant failed to contest that rejection on appeal); see also Manual of Patent Examining Procedure (MPEP) § 1205.02 (9th ed. rev. 08.2017 Jan. 2018) (explaining that “[a]n appellant’s brief must present arguments responsive to every ground of rejection stated by the examiner in the Office action from which the appeal Appeal 2018-008621 Application 14/599,272 8 has been taken (as modified by any advisory action and/or pre-appeal brief conference decision)”). Mayo/Alice Step One: PTO Step 2A Prong One for Claim 1 Patent-ineligible abstract ideas include mathematical formulas, mental processes, and fundamental economic practices. See, e.g., Alice, 573 U.S. at 219–20 (fundamental economic practice of intermediated settlement); Bilski v. Kappos, 561 U.S. 593, 599, 611–12 (2010) (fundamental economic practice of hedging or protecting against risk in independent claim and mathematical formula in dependent claim); Parker v. Flook, 437 U.S. 584, 585–86, 596–98 (1978) (mathematical formula for calculating updated alarm limit); Gottschalk v. Benson, 409 U.S. 63, 65–67 (1972) (mental process of converting binary-coded-decimal representation to binary representation). The 2019 Guidance enumerates three groupings of abstract ideas: (1) mathematical concepts, (2) mental processes, and (3) certain methods of organizing human activity. 84 Fed. Reg. at 51–52. In Alice, the Supreme Court “did not establish any ‘precise contours’ for defining whether claims are directed to ‘abstract ideas’ or something else.” Cellspin Soft, Inc. v. Fitbit, Inc., 927 F.3d 1306, 1315 (Fed. Cir. 2019) (citing Alice, 573 U.S. at 221). Further, for Mayo/Alice step one, the Federal Circuit has noted that “[a]n abstract idea can generally be described at different levels of abstraction.” Apple, Inc. v. Ameranth, Inc., 842 F.3d 1229, 1240 (Fed. Cir. 2016). Here, the Examiner determines that “[t]he claims viewed as a whole are directed to an abstract idea e.g., a mathematical relationship/formula.” Final Act. 5; see Ans. 5, 7. The Examiner explains that the claims are “similar to organizing and manipulating information through mathematical Appeal 2018-008621 Application 14/599,272 9 correlations (Digitech)” because the claims concern “generating a model, running an[] algorithm/software and determining one or more sets of correlated words.” Ans. 5, 7; see Final Act. 6 (citing Digitech). Appellant disputes that the claims are directed to an abstract idea. See Appeal Br. 6–11; Reply Br. 2–5. Appellant asserts that the Examiner mischaracterizes claim 1 as “collecting and comparing known information” and “organizing information through mathematical operations” because those characterizations do not “reflect the inventor’s improvement” in claim 1. Appeal Br. 9–10. We disagree that the Examiner mischaracterizes claim 1. In determining that claim 1 is directed to an abstract idea, the Examiner accurately assesses the “focus” of the claim and its “character as a whole.” See Final Act. 5–6; Advisory Act. 1; Ans. 5–7; see also SAP Am., 898 F.3d at 1167. In addition, we determine that claim 1 recites abstract ideas falling within two of the three groupings of abstract ideas enumerated in the 2019 Guidance. See 84 Fed. Reg. at 51–52. Specifically, claim 1 recites mathematical concepts and mental processes. First, the 2019 Guidance describes mathematical concepts as including “mathematical relationships, mathematical formulas or equations, [and] mathematical calculations.” 84 Fed. Reg. at 52. Here, claim 1 recites “generating a topic model” that (1) “identifies sets of correlated words” and (2) “comprises the one or more sets of correlated words.” Appeal Br. 17. Generating a topic model that identifies and comprises sets of correlated words involves mathematical relationships, i.e., mathematical relationships specifying how the correlated words relate to one another. Appeal 2018-008621 Application 14/599,272 10 Claim 1 also recites “running a partially-collapsed Gibbs sampler over a Dirichlet distribution of the plurality of words in the set of documents to produce sampler result data, further comprising: calculating a mean of the Dirichlet distribution.” Appeal Br. 17. Running a partially collapsed Gibbs sampler over a Dirichlet distribution involves mathematical calculations, i.e., mathematical calculations specifying probabilities. See Spec. ¶¶ 32–37, 63–67, 73–79, 84–86, Figs. 2, 7–8. Moreover, calculating a mean of the Dirichlet distribution also involves mathematical calculations. See id. ¶¶ 19, 54, 67–68, 72–73. The mean for a set of numbers equals the sum of the numbers in the set divided by the number of numbers in the set. For these reasons, claim 1 recites mathematical relationships and mathematical calculations. The 2019 Guidance identifies mathematical relationships and mathematical calculations as mathematical concepts, and thus an abstract idea. 84 Fed. Reg. at 52. Second, the 2019 Guidance describes mental processes as “concepts performed in the human mind,” such as “an observation, evaluation, judgment, [or] opinion.” 84 Fed. Reg. at 52 (footnote omitted). Here, claim 1 recites “analyzing text within one or more documents, other than the set of documents, based at least in part on the topic model.” Appeal Br. 17. Analyzing text based at least in part on a topic model encompasses evaluations performed by a human mentally or with pen and paper. For example, someone could mentally evaluate a topic model and then rely on the results of that evaluation to mentally evaluate text within a document. Consistent with those mental evaluations, the Specification discloses using a topic model to recognize names, tag parts of speech, and search text. Spec. Appeal 2018-008621 Application 14/599,272 11 ¶ 29. The human mind may recognize names, tag parts of speech, and search text. For these reasons, claim 1 recites evaluations performed by a human mentally or with pen and paper. The 2019 Guidance identifies evaluations performed by a human mentally or with pen and paper as mental processes, and thus an abstract idea. 84 Fed. Reg. at 52 & nn.14–15; see also Intellectual Ventures I LLC v. Symantec Corp., 838 F.3d 1307, 1318 (Fed. Cir. 2016) (“Symantec”) (“[W]ith the exception of generic computer- implemented steps, there is nothing in the claims themselves that foreclose them from being performed by a human, mentally or with pen and paper.”). For the reasons discussed above, claim 1 recites mathematical concepts and mental processes, and thus recites abstract ideas. Mayo/Alice Step One: PTO Step 2A Prong Two for Claim 1 Because we determine that claim 1 recites abstract ideas, we consider whether the claim as a whole integrates the recited abstract ideas into a practical application. See 84 Fed. Reg. at 54–55. “Only when a claim recites a judicial exception and fails to integrate the exception into a practical application, is the claim ‘directed to’ a judicial exception . . . .” Id. at 51. As additional elements, claim 1 recites computer components, i.e., “one or more computing devices.” Appeal Br. 17. Further, claim 1 recites the following data-collecting limitation: “receiving information for a set of documents; wherein the set of documents comprises a plurality of words.” Id. We determine that claim 1 as a whole does not integrate the recited abstract ideas into a practical application because the additional elements do Appeal 2018-008621 Application 14/599,272 12 not impose meaningful limits on the abstract ideas. See 84 Fed. Reg. at 53–54; see also Final Act. 5–6; Ans. 7–8. Instead, the claimed computer components constitute generic computer components that perform generic computer functions. See Final Act. 5–6; Ans. 7–8; Spec. ¶¶ 21, 24, 87–99, Fig. 9. Further, the data-collecting limitation constitutes insignificant extra- solution activity. See, e.g., Mayo, 566 U.S. at 79; Bilski, 561 U.S. at 611–12; OIP Techs., Inc. v. Amazon.com, Inc., 788 F.3d 1359, 1363–64 (Fed. Cir. 2015); Ultramercial, Inc. v. Hulu, LLC, 772 F.3d 709, 716 (Fed. Cir. 2014); CyberSource Corp. v. Retail Decisions, Inc., 654 F.3d 1366, 1370 (Fed. Cir. 2011); In re Grams, 888 F.2d 835, 839–40 (Fed. Cir. 1989). As an example of insignificant extra-solution activity, in Mayo the Supreme Court decided that measuring metabolite levels for later analysis constituted purely “conventional or obvious” pre-solution activity. Mayo, 566 U.S. at 79. Similarly, the Federal Circuit has held that mere data- gathering steps “cannot make an otherwise nonstatutory claim statutory.” CyberSource, 654 F.3d at 1370 (quoting Grams, 888 F.2d at 840). The Federal Circuit has also held that (1) presenting offers to potential customers and (2) gathering statistics concerning responses were “conventional data- gathering activities that [did] not make the claims patent eligible.” OIP Techs., 788 F.3d at 1363–64. Consistent with those decisions, the MPEP identifies “gathering data” as an example of insignificant pre-solution activity. MPEP § 2106.05(g). Here, the data-collecting limitation amounts to mere data gathering and requires nothing unconventional or significant. See Classen Immunotherapies, Inc. v. Biogen IDEC, 659 F.3d 1057, 1067 (Fed. Cir. Appeal 2018-008621 Application 14/599,272 13 2011) (analogizing “data gathering” to “insignificant extra-solution activity”). In particular, claim 1’s “receiving information” limitation operates to make data available for processing like the “conventional data- gathering activities” in OIP Technologies. See OIP Techs., 788 F.3d at 1363–64. Hence, claim 1’s “receiving information” limitation does not help integrate the recited abstract ideas into a practical application. Appellant contends that claim 1 “improves the functioning of a computer” by “configuring the generation of a topic model in a way (via calculating the mean of the Dirichlet distribution) to allow massively- parallel processing of the Gibbs sampler, i.e., using calculations that are suited to be run on massively parallel hardware, such as a Graphics Processing Unit (GPU).” Appeal Br. 6–7 (citing Spec. ¶¶ 30, 65–66, 76, 78); see id. at 7–9 (citing Spec. ¶¶ 3–5, 19); Reply Br. 2–4, 8–9, 13–14. Further, Appellant asserts that “weight should be given to all improvements provided by the invention, regardless of whether the improvements themselves are explicitly mentioned in the claims.” Reply Br. 4; see Appeal Br. 8. Appellant’s arguments do not persuade us of Examiner error because claim 1 does not require “massively parallel processing” on “massively parallel hardware.” See Appeal Br. 17. Unclaimed improvements do not support patent eligibility. See ChargePoint, Inc. v. SemaConnect, Inc., 920 F.3d 759, 769 (Fed. Cir. 2019); Berkheimer v. HP Inc., 881 F.3d 1360, 1369–70 (Fed. Cir. 2018); Two-Way Media Ltd. v. Comcast Cable Commc’ns, LLC, 874 F.3d 1329, 1338–39 (Fed. Cir. 2017); Intellectual Ventures I LLC v. Erie Indem. Co., 850 F.3d 1315, 1331–32 (Fed. Cir. 2017). Appeal 2018-008621 Application 14/599,272 14 As the Examiner properly reasons, “there is no mention” of “massively parallel processing” or “massively parallel hardware” in claim 1, and “the specification is not the measure of the invention.” Ans. 6. Further, Federal Circuit case law “counsels against incorporating a feature of a preferred embodiment into the claims.” Arthrex, Inc. v. Smith & Nephew, Inc., 935 F.3d 1319, 1330 (Fed. Cir. 2019). Claim 1 requires “one or more computing devices” that “generat[e] a topic model” from “a plurality of words” by “running a partially-collapsed Gibbs sampler over a Dirichlet distribution of the plurality of words.” Appeal Br. 17. A “plurality” means “more than one.” August Tech. Corp. v. Camtek, Ltd., 655 F.3d 1278, 1286 (Fed. Cir. 2011). As broadly drafted, claim 1 encompasses one computing device with one processor core that processes a few words sequentially. Due to its breadth, claim 1 does not exploit “massive amounts of parallelism by processing the algorithm on a highly-parallel architecture” to improve computer performance. See Spec. ¶ 19, Abstract. Appellant analogizes claim 1 to the claims in Thales. See Appeal Br. 8–9, 15 (citing Thales Visionix Inc. v. United States, 850 F.3d 1343 (Fed. Cir. 2017)); Reply Br. 4. Specifically, Appellant asserts that “[t]he mere fact that a claim involves mathematical principles does not doom the claim to patent-ineligibility.” Appeal Br. 15. We disagree that claim 1 resembles the claims in Thales. There, the patent at issue disclosed “an inertial tracking system for tracking the motion of an object relative to a moving reference frame.” Thales, 850 F.3d at 1344. Each independent claim required two inertial sensors, i.e., one sensor mounted on the object and another sensor mounted on the moving Appeal 2018-008621 Application 14/599,272 15 reference frame. Id. at 1345–46, 1348. The claimed inertial sensors supplied signals for determining the object’s orientation relative to the moving reference frame. Id. at 1345–46, 1348. The Federal Circuit decided that the claims satisfied § 101 under Mayo/Alice step one because they employed “inertial sensors in a non-conventional manner to reduce errors in measuring the relative position and orientation of a moving object on a moving reference frame.” Id. at 1348–49. Unlike the claims in Thales, claim 1 does not employ conventional components in a non-conventional manner. Appeal Br. 17; see Final Act. 5–6; Ans. 7–8. The Specification describes the claimed computer components generically and evidences their conventional nature. See Spec. ¶¶ 21, 24, 87–99, Fig. 9. For example, the Specification explains that “computing devices” include “workstations, personal computers, laptop computers, personal digital assistants (PDAs), tablet computers, cellular telephony devices such as smart phones, and any other type of computing device.” Id. ¶ 21; see id. ¶ 87. Appellant also analogizes claim 1 to the claims in Core Wireless. Appeal Br. 6–7 (citing Core Wireless Licensing S.A.R.L. v. LG Elecs., Inc., 880 F.3d 1356 (Fed. Cir. 2018)). Specifically, Appellant asserts that claim 1 “is not merely a ‘method of creating a mathematical model’, or even a method that merely performs math.” Id. at 6. We disagree that claim 1 resembles the claims in Core Wireless. There, the “claims recite[d] a specific improvement over prior systems,” i.e., a “particular manner of summarizing and presenting information in electronic devices.” Core Wireless, 880 F.3d at 1362–63. The Federal Circuit decided that the claims satisfied § 101 under Mayo/Alice step one Appeal 2018-008621 Application 14/599,272 16 because the “invention improve[d] the efficiency of using the electronic device.” Id. at 1363. Unlike the claims in Core Wireless, claim 1 does not improve the efficiency of using a computing device or any other electronic device. Appeal Br. 17; see Ans. 6. As discussed above, unclaimed improvements do not support patent eligibility. See, e.g., ChargePoint, 920 F.3d at 769. A “process that start[s] with data, add[s] an algorithm, and end[s] with a new form of data [is] directed to an abstract idea.” RecogniCorp, 855 F.3d at 1327. Claim 1 recites a “process that start[s] with data, add[s] an algorithm, and end[s] with a new form of data.” Thus, claim 1 is directed to an abstract idea. The 2019 Guidance identifies exemplary considerations indicating that additional elements in claims “may have integrated the [judicial] exception into a practical application.” 84 Fed. Reg. at 55 & nn.25–29 (citing MPEP §§ 2106.05(a)–(c), 2106.05(e)). As the above analysis indicates, we have evaluated Appellant’s arguments in light of those exemplary considerations. For the reasons discussed above, however, we determine that claim 1 as a whole does not integrate the recited abstract ideas into a practical application. Thus, claim 1 is directed to a judicial exception and does not satisfy § 101 under Mayo/Alice step one. Mayo/Alice Step Two: PTO Step 2B for Claim 1 Because we determine that claim 1 is directed to a judicial exception, we “consider the elements of” the claim “both individually and ‘as an ordered combination’ to determine whether the additional elements” add enough to transform the “nature of the claim” into “significantly more” than the judicial exception. See Alice, 573 U.S. at 217–18, 221–22 (quoting Appeal 2018-008621 Application 14/599,272 17 Mayo, 566 U.S. at 78–79). Under Mayo/Alice step two, we “look with more specificity at what the claim elements add, in order to determine ‘whether they identify an “inventive concept” in the application of the ineligible subject matter’ to which the claim is directed.” Affinity Labs of Tex., LLC v. DIRECTV, LLC, 838 F.3d 1253, 1258 (Fed. Cir. 2016) (quoting Elec. Power Grp., LLC v. Alstom S.A., 830 F.3d 1350, 1353 (Fed. Cir. 2016)). An “inventive concept” requires more than “well-understood, routine, conventional activity already engaged in” by the relevant community. Rapid Litig. Mgmt. Ltd. v. CellzDirect, Inc., 827 F.3d 1042, 1047 (Fed. Cir. 2016) (quoting Mayo, 566 U.S. at 79–80). But a “non-conventional and non-generic arrangement of known, conventional pieces” may provide an “inventive concept” satisfying step two. BASCOM Glob. Internet Servs., Inc. v. AT&T Mobility LLC, 827 F.3d 1341, 1350 (Fed. Cir. 2016). As discussed above, claim 1 recites computer components, i.e., “one or more computing devices.” Appeal Br. 17. As also discussed above, the claimed computer components constitute generic computer components that perform generic computer functions. See Final Act. 5–6; Ans. 7–8; Spec. ¶¶ 21, 24, 87–99, Fig. 9. Further, as discussed above, the Specification describes the claimed computer components generically and evidences their conventional nature. See Spec. ¶¶ 21, 24, 87–99, Fig. 9. For example, the Specification explains that “the techniques described herein are implemented by” either (1) “one or more special-purpose computing devices” or (2) “one or more general purpose hardware processors programmed to perform the techniques pursuant to program instructions in firmware, memory, other storage, or a combination.” Spec. ¶ 87. The Specification describes “special-purpose Appeal 2018-008621 Application 14/599,272 18 computing devices” as including (1) “digital electronic devices such as one or more application-specific integrated circuits (ASICs) or field programmable gate arrays (FPGAs) that are persistently programmed to perform the techniques” and (2) “desktop computer systems, portable computer systems, handheld devices, networking devices or any other device that incorporates hard-wired and/or program logic to implement the techniques.” Id. Simply implementing an abstract idea using conventional machines or devices “add[s] nothing of substance.” See Alice, 573 U.S. at 226–27; see also Mayo, 566 U.S. at 84–85 (explaining that “simply implementing a mathematical principle on a physical machine” does not suffice for patent eligibility) (citing Gottschalk, 409 U.S. at 64–65, 71). Moreover, the claimed computer components operate to collect, manipulate, and communicate data. Appeal Br. 17; see Ans. 5. Court decisions have recognized that generic computer components operating to collect, manipulate, and communicate data are well understood, routine, and conventional to a skilled artisan. See, e.g., Alice, 573 U.S. at 226–27; SAP Am., 898 F.3d at 1164–65 & n.1, 1170; Apple, 842 F.3d at 1234, 1241–42; Symantec, 838 F.3d at 1316–20; Versata Dev. Grp., Inc. v. SAP Am., Inc., 793 F.3d 1306, 1334 (Fed. Cir. 2015); Ultramercial, 772 F.3d at 715–16; buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355 (Fed. Cir. 2014). For example, the system claims in Alice recited a “data processing system” with a “communications controller” and a “data storage unit.” Alice, 573 U.S. at 226. The Supreme Court decided that the system claims failed to satisfy § 101 because “[n]early every computer” includes those generic components for performing “basic calculation, storage, and Appeal 2018-008621 Application 14/599,272 19 transmission functions” and the system claims simply implemented the same abstract idea as the method claims. Id. at 226–27. The Court reasoned that (1) “the system claims are no different from the method claims in substance”; (2) “[t]he method claims recite the abstract idea implemented on a generic computer”; and (3) “the system claims recite a handful of generic computer components configured to implement the same idea.” Id. at 226. Here, the claimed computer components perform “basic calculation, storage, and transmission functions” that nearly every computer performs. See Appeal Br. 17. Consequently, the claimed computer components do not satisfy the “inventive concept” requirement. See, e.g., Customedia Techs., LLC v. Dish Network Corp., 951 F.3d 1359, 1361, 1366 (Fed. Cir. 2020) (describing the claimed “processor,” “storage device,” “programmable receiver unit,” and “remote server” as “generic computer components”); Mortg. Grader, Inc. v. First Choice Loan Servs. Inc., 811 F.3d 1314, 1324–25 (Fed. Cir. 2016) (holding that “generic computer components such as an ‘interface,’ ‘network,’ and ‘database’” did not satisfy the “inventive concept” requirement); FairWarning IP, LLC v. Iatric Sys., Inc., 839 F.3d 1089, 1095–96 (Fed. Cir. 2016) (describing the claimed “microprocessor” and “user interface” as “generic computer elements”); Prism Techs. LLC v. T-Mobile USA, Inc., 696 F. App’x 1014, 1016–17 (Fed. Cir. 2017) (describing the claimed “authentication server,” “access server,” “Internet Protocol network,” “client computer device,” and “database” as “indisputably generic computer components”). We reach a similar conclusion concerning claim 1’s data-collecting limitation, i.e., “receiving information for a set of documents; wherein the set of documents comprises a plurality of words.” As discussed above, this Appeal 2018-008621 Application 14/599,272 20 limitation amounts to mere data gathering and requires nothing unconventional or significant. Consequently, the claimed insignificant extra-solution activity does not satisfy the “inventive concept” requirement. See, e.g., Mayo, 566 U.S. at 79–80; OIP Techs., 788 F.3d at 1363–64; CyberSource, 654 F.3d at 1370. Appellant analogizes claim 1 to patent-eligible claim 4 in Berkheimer. See Appeal Br. 12–13 (citing Berkheimer v. HP Inc., 881 F.3d 1360 (Fed. Cir. 2018)); Reply Br. 7–8, 12. In our view, however, claim 1 parallels patent-ineligible claims 1–3 and 9 in Berkheimer. In Berkheimer, the Federal Circuit decided that disputed factual issues did not preclude a summary judgment of ineligibility for independent claim 1 and dependent claims 2, 3, and 9. Berkheimer, 881 F.3d at 1369–70. In contrast to those claims, the court decided that disputed factual issues precluded a summary judgment of ineligibility for dependent claims 4–7. Id. at 1370. The court said, “Whether claims 4–7 perform well-understood, routine, and conventional activities to a skilled artisan is a genuine issue of material fact making summary judgment inappropriate with respect to these claims.” Id. The court explained that “improvements [described] in the specification, to the extent they are captured in the claims, create a factual dispute regarding whether the invention describes well-understood, routine, and conventional activities.” Id. at 1369. Because claims 1–3 and 9 did not “capture the purportedly inventive concepts” described in the specification, the court affirmed a summary judgment of ineligibility for those claims. Id. at 1370. Here, claim 1 does not “capture the purportedly inventive concepts” described in the Specification. As discussed above, claim 1 does not require Appeal 2018-008621 Application 14/599,272 21 “massively parallel processing” on “massively parallel hardware.” See Appeal Br. 17; Ans. 6. Because claim 1 does not “capture the purportedly inventive concepts” described in the Specification, claim 1 parallels patent- ineligible claims 1–3 and 9 in Berkheimer. Appellant contends that claim 1 includes an “inventive concept” because claim 1 includes the following bolded features that go beyond “mathematical relationships and formulas”: 1. A method comprising: generating a topic model that identifies sets of correlated words by: receiving information for a set of documents; wherein the set of documents comprises a plurality of words; running a partially-collapsed Gibbs sampler over a Dirichlet distribution of the plurality of words in the set of documents to produce sampler result data, further comprising: calculating a mean of the Dirichlet distribution; and determining, from the sampler result data, one or more sets of correlated words; wherein the topic model comprises the one or more sets of correlated words; and analyzing text within one or more documents, other than the set of documents, based at least in part on the topic model; wherein the method is performed by one or more computing devices. Reply Br. 7–8 (emphasis by Appellant). But Appellant wrongly relies on the recited abstract ideas to satisfy the “inventive concept” requirement. A “claimed invention’s use of the ineligible concept,” e.g., an abstract idea, “cannot supply the inventive Appeal 2018-008621 Application 14/599,272 22 concept that renders the invention ‘significantly more’ than that ineligible concept.” BSG Tech LLC v. BuySeasons, Inc., 899 F.3d 1281, 1290 (Fed. Cir. 2018). Under “the Mayo/Alice framework, a claim directed to a newly discovered” abstract idea “cannot rely on the novelty of that discovery for the inventive concept necessary for patent eligibility.” Genetic Techs. Ltd. v. Merial L.L.C., 818 F.3d 1369, 1376 (Fed. Cir. 2016). Appellant asserts that “based on the lack of any art-based rejections,” the claimed features “are both novel and non-obvious, which represent a much higher bar than is presented by the ‘well-understood, routine, conventional activity’ standard.” Reply Br. 2; see id. at 13. But the absence of anticipation and obviousness “does not resolve the question of whether the claims embody an inventive concept” under Mayo/Alice step two. Symantec, 838 F.3d at 1315; see SAP Am., 898 F.3d at 1163. “The search for a § 101 inventive concept” differs “from demonstrating § 102 novelty.” Synopsys, Inc. v. Mentor Graphics Corp., 839 F.3d 1138, 1151 (Fed. Cir. 2016); see Two-Way Media, 874 F.3d at 1340. “The ‘novelty’ of any element[s] or steps” in a claim or even the claim itself “is of no relevance in determining whether the subject matter of a claim” satisfies § 101. Diehr, 450 U.S. at 188–89. Appellant cites PTO guidance about Berkheimer.3 See Reply Br. 2, 10–13. Appellant asserts that the Examiner “fails to carry [the] burden of proof” that the additional elements in claim 1 are “well-understood, routine, 3 PTO Memorandum, Changes in Examination Procedure Pertaining to Subject Matter Eligibility, Recent Subject Matter Eligibility Decision (Berkheimer v. HP, Inc.) (Apr. 19, 2018). Appeal 2018-008621 Application 14/599,272 23 conventional activity” according to that PTO guidance. Reply Br. 12–13 (emphasis omitted). But Appellant concedes that “[a] citation to one or more . . . court decisions” may demonstrate that additional elements are well understood, routine, and conventional to a skilled artisan. Reply Br. 11 (emphasis omitted). As discussed above, court decisions have recognized that generic computer components operating to collect, manipulate, and communicate data are well understood, routine, and conventional to a skilled artisan. See, e.g., Alice, 573 U.S. at 226–27; SAP Am., 898 F.3d at 1164–65 & n.1, 1170; Apple, 842 F.3d at 1234, 1241–42; Symantec, 838 F.3d at 1316–20; Versata, 793 F.3d at 1334; Ultramercial, 772 F.3d at 715–16; buySAFE, 765 F.3d at 1355. “Whether a combination of claim limitations supplies an inventive concept that renders a claim ‘significantly more’ than an abstract idea to which it is directed is a question of law.” BSG Tech, 899 F.3d at 1290. Without sufficient additional elements, “a process that employs mathematical algorithms to manipulate existing information to generate additional information is not patent eligible.” Digitech, 758 F.3d at 1351. Here, considering claim 1’s limitations as an ordered combination adds nothing to the abstract ideas that is not already present when the limitations are considered separately. See Mayo, 566 U.S. at 79; Final Act. 5–6; Ans. 7–8. Claim 1’s ordered combination of limitations amounts to nothing more than the abstract ideas implemented with generic computer components that perform generic computer functions. See Alice, 573 U.S. at 225–26; Final Act. 5–6; Ans. 7–8. Hence, we conclude that claim 1’s ordered combination of limitations does not supply an “inventive concept” Appeal 2018-008621 Application 14/599,272 24 that renders the claim “significantly more” than the abstract ideas. See Final Act. 5–6; Ans. 7–8. Thus, claim 1 does not satisfy § 101 under Mayo/Alice step two. Summary for Claim 1 For the reasons discussed above, Appellant’s arguments have not persuaded us of any error in the Examiner’s findings or conclusions for claim 1 under Mayo/Alice step one or step two. Hence, we sustain the § 101 rejection of claim 1. Dependent Claims 4 and 5 Claim 4 depends from claim 1 and specifies that “running the partially-collapsed Gibbs sampler over the Dirichlet distribution comprises computing a plurality of values, for the partially-collapsed Gibbs sampler, in parallel.” Appeal Br. 17. Claim 5 depends from claim 4 and specifies that “the plurality of values . . . are computed in a plurality of parallel Single Program Multiple Data (SPMD) units on a graphics processing unit (GPU).” Id. Appellant contends that claims 4 and 5 are “not directed to an ‘abstract’ idea, but to an improved method of generating a topic model utilizing parallel capabilities of a computing device (e.g., a GPU).” Appeal Br. 7 (emphasis omitted). Further, Appellant asserts that claims 4 and 5 “take advantage of parallel architecture of a computing device for running a highly-parallel partially-collapsed Gibbs sampler with a replaced stochastic step to efficiently and effectively create a topic model based on a sample set of documents.” Id. at 9. Appellant also asserts that “optimizations” according to claims 4 and 5 “directly improve the performance of a Appeal 2018-008621 Application 14/599,272 25 computing device generating a topic model by generating the topic model via ‘faster than state-of-the art algorithms’ while generating as good or better solutions.” Id. (emphasis omitted). Appellant’s arguments do not persuade us of Examiner error because claims 4 and 5 do not “capture the purportedly inventive concepts” described in the Specification. As explained in more detail below, claims 4 and 5 do not require “massively parallel processing” on “massively parallel hardware.” Instead, claim 4 merely narrows the abstract ideas recited in claim 1. Dependent claims that merely narrow abstract ideas “add nothing outside the abstract realm.” SAP Am., 898 F.3d at 1169. Narrowing “an abstract idea does not add ‘significantly more’ to it.” BSG Tech, 899 F.3d at 1291. And adding “a degree of particularity” as specified in claim 4 does not save a claim from abstraction. See Ultramercial, 772 F.3d at 715. As discussed above, claim 1 encompasses one computing device with one processor core that processes a few words sequentially. Similar to claim 1, claim 4 encompasses one computing device with two processor cores that process a few words in parallel. Consistent with this, the Specification distinguishes computing devices with multicore processors from computing devices with GPUs. See Spec. ¶ 30. Hence, claim 4 does not require “massively parallel processing” on “massively parallel hardware.” Due to its breadth, claim 4 does not exploit “massive amounts of parallelism by processing the algorithm on a highly-parallel architecture” to improve computer performance. See id. ¶ 19, Abstract; see also id. ¶ 30. Moreover, claim 4 employs the same generic computer components as claim 1 to perform generic computer functions, i.e., collecting, manipulating, Appeal 2018-008621 Application 14/599,272 26 and communicating data. Consequently, claim 4’s ordered combination of limitations amounts to nothing more than the abstract ideas implemented with generic computer components that perform generic computer functions. Claim 5 recites certain components, i.e., “a plurality of parallel Single Program Multiple Data (SPMD) units on a graphics processing unit (GPU).” But “claims are not saved from abstraction merely because they recite components more specific than a generic computer.” BSG Tech, 899 F.3d at 1286. Similar to claim 4, claim 5 encompasses one computing device with two SPMD units on a GPU that process a few words in parallel. Consistent with this, the Specification explains that certain calculations for each document in a document set “may be processed in parallel,” e.g., by “a particular SPMD unit” that “processes calculations specific to a particular document” in a document set. Spec. ¶ 78 (emphasis added); see Appeal Br. 17 (claim 6 depending from claim 5). A document set may consist of two documents with a handful of words between them. Although claim 5 recites SPMD units on a GPU, claim 5 does not require “massively parallel processing” on “massively parallel hardware.” Due to its breadth, claim 5 does not exploit “massive amounts of parallelism by processing the algorithm on a highly-parallel architecture” to improve computer performance. See Spec. ¶ 19, Abstract; see also id. ¶¶ 30, 78. In addition, the Specification describes a GPU generically. See Spec. ¶ 24. For example, the Specification explains that the functionality of a GPU on a server “may be performed by” any of the following: (1) “another entity running on [the] server,” (2) “a GPU or other hardware on [a] client,” and (3) “other devices that are communicatively coupled to [a] network.” Appeal 2018-008621 Application 14/599,272 27 Id. Consequently, claim 5’s ordered combination of limitations amounts to nothing more than the abstract ideas implemented with generic computer components that perform generic computer functions. For the reasons discussed above, Appellant’s arguments have not persuaded us of any error in the Examiner’s findings or conclusions for claims 4 and 5 under Mayo/Alice step one or step two. Hence, we sustain the § 101 rejection of claims 4 and 5. Independent Claims 8 and 15 and Dependent Claims 2, 3, 6, 7, 9–14, and 16–19 In the Appeal Brief, Appellant did not argue eligibility separately for independent claims 8 and 15 or dependent claims 2, 3, 6, 7, 9–14, and 16–19. See Appeal Br. 6–16. In the Reply Brief, Appellant argues that claims 2, 3, 6, and 7 recite “significant details” that “go far beyond the very broad ‘organizing and manipulating information through mathematical correlations’.” Reply Br. 8–9 (emphasis omitted). Those arguments do not, however, respond to any arguments in the Answer. See Ans. 5–8. Appellant advances those arguments for the first time in the Reply Brief. Such arguments “will not be considered by the Board” unless an appellant shows good cause. See 37 C.F.R. § 41.41(b)(2); see also Ex parte Borden, 93 USPQ2d 1473, 1475 (BPAI 2010) (informative) (discussing procedural difficulties with belated arguments). Here, Appellant has not shown good cause for the belated arguments. Hence, we decline to consider them. “Considering an argument advanced for the first time in a reply brief . . . is not only unfair to an appellee, but also entails the risk of an improvident or ill-advised opinion on the legal issues Appeal 2018-008621 Application 14/599,272 28 tendered.” McBride v. Merrell Dow & Pharm., Inc., 800 F.2d 1208, 1211 (D.C. Cir. 1986) (citation omitted). Because Appellant did not argue eligibility separately for independent claims 8 and 15 or dependent claims 2, 3, 6, 7, 9–14, and 16–19 in the Appeal Brief, we sustain the § 101 rejection of these claims for the same reasons we sustain the § 101 rejection of claims 1, 4, and 5. See Appeal Br. 6–16; 37 C.F.R. § 41.37(c)(1)(iv). CONCLUSION We affirm the Examiner’s decision to reject claims 1–19. In summary: Claims Rejected 35 U.S.C. § Reference(s)/Basis Affirmed Reversed 1–19 101 Eligibility 1–19 TIME PERIOD FOR RESPONSE No time period for taking any subsequent action in connection with this appeal may be extended under 37 C.F.R. § 1.136(a)(1)(iv). See 37 C.F.R. § 41.50(f). AFFIRMED Copy with citationCopy as parenthetical citation