Infineon Technologies AG

Patent Trials and Appeals BoardApr 2, 2021

2019006903 (P.T.A.B. Apr. 2, 2021)

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. 14/029,883 09/18/2013 Markus Dobler 1012-0681 / 2013P51379US 4373 57579 7590 04/02/2021 MURPHY, BILAK & HOMILLER/INFINEON TECHNOLOGIES 1255 CRESCENT GREEN SUITE 200 CARY, NC 27518 EXAMINER SITIRICHE, LUIS A ART UNIT PAPER NUMBER 2126 NOTIFICATION DATE DELIVERY MODE 04/02/2021 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): official@mbhiplaw.com PTOL-90A (Rev. 04/07) UNITED STATES PATENT AND TRADEMARK OFFICE ____________________ BEFORE THE PATENT TRIAL AND APPEAL BOARD ____________________ Ex parte MARKUS DOBLER ____________________ Appeal 2019-006903 Application 14/029,883 Technology Center 2100 ____________________ Before ST. JOHN COURTENAY III, LARRY J. HUME, and JOYCE CRAIG, Administrative Patent Judges. HUME, Administrative Patent Judge. DECISION ON APPEAL Pursuant to 35 U.S.C. § 134(a), Appellant1 appeals from the Non- Final Rejection of claims 1–5, 8–16, 19, and 20, which are all claims pending in the application. We have jurisdiction under 35 U.S.C. § 6(b). We REVERSE. 1 According to Appellant, the real party in interest is Infineon Technologies AG. Appeal Br. 2. Appeal 2019-006903 Application 14/029,883 2 STATEMENT OF THE CASE2 The Invention Appellant’s disclosed embodiments and claimed invention relate to accurately determining failure regions of an electrical device. Spec. ¶ 1. Exemplary Claim Claim 1, reproduced below, is representative of the claimed subject matter on Appeal (emphasis added to contested prior-art limitations): 1. A method of determining failure regions of an electrical device, the method comprising: training a machine learning classifier by analyzing data points for the electrical device and recognizing patterns in the data points, each data point indicating pass or fail of the electrical device for a particular combination of factors relating to the operation of the electrical device; and using the trained machine learning classifier to predict the pass/fail state of new data points for the electrical device, each new data point corresponding to a new combination of the factors relating to the operation of the electrical device not previously analyzed by the machine learning classifier. 2 Our decision relies upon Appellant’s Appeal Brief (“Appeal Br.,” filed Feb. 19, 2019); Reply Brief (“Reply Br.,” filed Sept. 20, 2019); Examiner’s Answer (“Ans.,” mailed July 22, 2019); Non-Final Office Action (“Non- Final Act.,” mailed July 19, 2018); and the original Specification (“Spec.,” filed Sept. 18, 2013). Appeal 2019-006903 Application 14/029,883 3 REFERENCES The prior art relied upon by the Examiner as evidence is: Name Reference Date Moriuchi et al. (“Moriuchi”) US 6,436,557 B1 Aug. 20, 2002 Virkar et al. (“Virkar”) US 2010/0063948 A1 Mar. 11, 2010 Claudio M. Rocco & José Ali Moreno, “System Reliability Evaluation Using Monte Carlo & Support Vector Machine” IEEE (2003) (“Rocco”). Giles M. Foody, “The Significance Of Border Training Patterns In Classification By A Feedforward Neural Network Using Back Propagation Learning” 20 International Journal of Remote Sensing (Nov. 25, 2010). Appeal 2019-006903 Application 14/029,883 4 REJECTIONS3, 4, 5 R1. Claims 1, 10–16, 19, and 20 stand rejected under 35 U.S.C. § 103(a) as being unpatentable over the combination of Rocco and Virkar. Non-Final Act. 9, 17. R2. Claims 2, 3, 5, 8, and 9 stand rejected under 35 U.S.C. § 103(a) as being unpatentable over the combination of Rocco, Virkar, and Foody. Non-Final Act. 12. R3. Claim 4 stands rejected under 35 U.S.C. § 103(a) as being unpatentable over the combination of Rocco, Virkar, Foody, and Moriuchi. Non-Final Act. 15–16. ISSUE Appellant argues (Appeal Br. 5–11; Reply Br. 2–4) the Examiner’s rejection of claim 1 under 35 U.S.C. § 103 as being obvious over the 3 The Examiner has withdrawn the § 112(b) rejection of claims 3 and 15. Ans. 3. The Examiner has also withdrawn the § 101 rejection of claims 1– 20. Id. 4 We note the Examiner indicates that claims 6, 7, 17, and 18 are not subject to any art rejection. Non-Final Act. 16. Accordingly, under 35 U.S.C. § 6(b), claims 6, 7, 17, and 18 are not before us on appeal. 5 In addition, the Examiner fails to set forth a prima facie case of unpatentability for claims 13–16, 19, and 20. “This is a system version of the claimed method discussed above, in claims 1–3, 5, and 11 respectively, wherein all claimed limitations have also been addressed and/or cited as set forth above.” Non-Final Act. 17. We note claims 1–3, 5, and 11 are variously rejected under either Rejection R1 or Rejection R2. However, the Examiner does not articulate which specific rejection has been applied to each of claims 13–16, 19, and 20, and thus we determine that these claims have not been properly rejected. See the notice requirement set forth under 35 U.S.C. § 132(a). Accordingly, we summarily reverse the rejection of claims 13–16, 19, and 20 under 35 U.S.C. § 103. Appeal 2019-006903 Application 14/029,883 5 combination of Rocco and Virkar is in error. These contentions present us with the following issue: Did the Examiner err in finding the cited prior art teaches or suggests “[a] method of determining failure regions of an electrical device”, that includes, inter alia, the step of “using the trained machine learning classifier to predict the pass/fail state of new data points for the electrical device, each new data point corresponding to a new combination of the factors relating to the operation of the electrical device not previously analyzed by the machine learning classifier,” as recited in claim 1 (emphasis added)? ANALYSIS Based upon our review of the record, we find a preponderance of the evidence supports particular arguments advanced by Appellant with respect to claim 1 for the specific reasons discussed below. We highlight and address specific findings and arguments regarding claim 1 for emphasis as follows. Appellant contends Rocco does not teach or suggest predicting a new combination of factors that were not previously analyzed. Appeal Br. 6. Appellant argues “[c]onventionally, prediction algorithms predict the pass/fail state for the same factors that are inputted into the prediction model.” Appeal Br. 7. However, Appellant further contends: The presently disclosed technique predicts the pass fail/state of data points for a new combination of factors that were not previously analyzed. See, e.g., Specification, paragraph [0025]. Using the above mentioned factors as an example, if the model receives data points for the factors of factors of temperature, voltage, and frequency, it predicts the Appeal 2019-006903 Application 14/029,883 6 pass fail/state in a different factor domain, e.g., amplifier gain, voltage, and threshold level. Id. Appellant specifically argues Virkar’s prediction is based upon known factors, whereas “the claimed prediction technique makes a prediction of the pass/fail state of data for a new combination of factors that were not previously known.” Appeal Br. 9 (emphasis added). Therefore, Appellant argues Virkar “does not suggest the ‘new combination of the factors . . . not previously analyzed’ requirement of the claims.” Id. Appellant summarizes the argument: The claimed technique is unique in that it makes a prediction about data states in a particular factor domain without previously analyzing known data states in the same factor domain. The Office has not identified any teaching from the prior art that suggests this technique. Appeal Br. 11. The Examiner finds Rocco does not teach predicting new data points not previously analyzed by the classifier, but that Rocco teaches “corresponding to a new combination of the factors relating to the operation of the electrical device.” Non-Final Act. 10, citing Rocco, p. 482, left col. ¶ 1.0. The Examiner finds Virkar’s “new data” corresponds to the claim limitation “new data points . . . not previously analyzed,” because Virkar is using the previous training data samples and its patterns learned to estimate if new data/factors exhibit the same pattern. Therefore, these new data/factors have not been previously analyzed. Ans. 15, citing Virkar ¶ 77. The Examiner, however, conflates the two separate and distinct limitations of “new data points” and “new combination of the factors Appeal 2019-006903 Application 14/029,883 7 relating to the operation of the electrical device not previously analyzed by the machine learning classifier,” even though the Examiner cites to two different references for teaching or suggesting these two limitations. Ans. 14–15. The Examiner cites to Virkar in describing the teachings related to the disputed limitation “each new data point corresponding to a new combination of the factors relating to the operation of the electrical device” in the Answer. However, in the Non-Final Action, the Examiner cites to Rocco for teaching the “combination of factors relating to the operation of the electrical device” portion of the disputed limitation. Non-Final Act. 10. For context, we find Appellant’s Specification describes the claim term “combination of factors” as factors such as supply voltage, ambient temperature, amplifier gain, friction coefficients, threshold levels, load resistance, actuator inertia, etc. Spec. ¶ 21. We do not adopt the Examiner’s findings because the Examiner’s citation to Virkar in the Answer lacks a teaching or suggestion “to a new combination of factors . . . not previously analyzed.” Non-Final Act. 10, citing Virkar ¶ 77. Indeed, Virkar (¶ 77) actually teaches using the results to predict whether new data exhibits the same (i.e., old) categories — not a new combination of factors/categories. See Virkar ¶ 77. Therefore, we are persuaded by Appellant that the cited references do not teach or suggest the disputed limitation. Thus, for essentially the same reasons argued by Appellant,6 we reverse the Examiner’s obviousness 6 Because we agree with at least one of the dispositive arguments advanced by Appellant, we need not reach the merits of Appellant’s other arguments. See Beloit Corp. v. Valmet Oy, 742 F.2d 1421, 1423 (Fed. Cir. 1984) (finding an administrative agency is at liberty to reach a decision based on “a single dispositive issue”). Appeal 2019-006903 Application 14/029,883 8 Rejection R1 of independent claim 1, and we also reverse the rejection of dependent claims 10–12 that variously depend from independent claim 1. Moreover, on this record, the Examiner has not shown how the additionally cited Foody and Moriuchi references in rejections R2 and R3 of claims 2–5, 8, and 9 overcome the deficiencies with Rocco and Virkar, as discussed regarding Rejection R1 of claim 1. CONCLUSION The Examiner erred with respect to obviousness rejections R1–R3 of claims 1–5, 8–16, 19, and 20 under 35 U.S.C. § 103(a), and we do not sustain the rejections. Appeal 2019-006903 Application 14/029,883 9 DECISION SUMMARY7 Claim(s) Rejected 35 U.S.C. § References/ Basis Affirmed Reversed 1, 10–16, 19, 20 103 Rocco, Virkar 1, 10–16, 19, 20 2, 3, 5, 8, 9 103 Rocco, Virkar, Foody 2, 3, 5, 8, 9 4 103 Rocco, Virkar, Foody, Moriuchi 4 13–16, 19, 20 103 Rocco, Virkar, Foody, Moriuchi8 13–16, 19, 20 Overall Outcome 1–5, 8–16, 19, 20 REVERSED 7 See n.4, supra, regarding claims 6, 7, 17, and 18. 8 See n.5, supra, regarding the lack of a prima facie case for rejection of claims 13–16, 19, and 20.