Ex Parte Kuo et al

Patent Trial and Appeal BoardNov 9, 2012

11698507 (P.T.A.B. Nov. 9, 2012)

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. 11/698,507 01/26/2007 Calvin Kuo 12729/254 (Y01866US02) 5769 56020 7590 11/09/2012 BRINKS HOFER GILSON & LIONE / YAHOO! OVERTURE P.O. BOX 10395 CHICAGO, IL 60610 EXAMINER AHMED, AFFAF ART UNIT PAPER NUMBER 3622 MAIL DATE DELIVERY MODE 11/09/2012 PAPER 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. PTOL-90A (Rev. 04/07) UNITED STATES PATENT AND TRADEMARK OFFICE ____________________ BEFORE THE PATENT TRIAL AND APPEAL BOARD ____________________ Ex parte CALVIN KUO and CRAIG AUZENNE ____________________ Appeal 2011-004821 Application 11/698,507 Technology Center 3600 ____________________ Before: HOWARD B. BLANKESHIP, MICHAEL W. KIM, and THOMAS L. GIANNETTI, Administrative Patent Judges. KIM, Administrative Patent Judge. DECISION ON APPEAL Appeal 2011-004821 Application 11/698,507 2 STATEMENT OF CASE Appellants appeal under 35 U.S.C. § 134 from a rejection of claims 1- 211. We have jurisdiction under 35 U.S.C. § 6(b). We REVERSE. The claims are directed to a system for generating scores that may indicate the likelihood of the service provider partner generating large revenues (Spec., para. [0002]). Claim 1, reproduced below, is illustrative of the claimed subject matter: 1. A method for generating scores related to interactions with a service provider partner, comprising: identifying a historical dataset describing characteristics of a plurality of service provider partners wherein at least one of the characteristics comprises a profitability of each of the plurality of service provider partners when provided with advertisements of a category of advertisements to serve to users; processing the historical dataset to generate a feature vector comprising a combination of at least two variables, wherein the combination of at least two variables for data of a given service provider partner relates to detecting whether the given service provider partner will be profitable when provided with the advertisements of the category of advertisements to serve to the users; generating a classifier model from the historical dataset and the feature vector, wherein a set of inputs of the generated classifier model comprises the combination of at least two variables of the feature vector; collecting current service provider partner data representing characteristics of a current service provider partner; 1 Our decision will make reference to the Appellants’ Appeal Brief (“App. Br.,” filed Sept. 8, 2010) and Reply Brief (“Reply Br.,” filed Jan. 4, 2011) and the Examiner’s Answer (“Ans.,” mailed Nov. 24, 2010). Appeal 2011-004821 Application 11/698,507 3 processing the current service provider partner data to generate a current service provider partner data feature vector, wherein the current service provider partner data feature vector comprises the current service provider partner data for the combination of at least two variables; and generating, by a processor, a score for the current service provider partner by applying the classifier model to the current service provider partner data feature vector, wherein the score represents a likelihood that the current service provider partner will be profitable when provided with the advertisements of the category of advertisements to serve to the users. THE REJECTIONS2 Under 35 U.S.C. § 103(a), the Examiner rejected claims 1, 2, and 4-21 as unpatentable over Desikan (US 2007/0005417 A1, pub. Jan. 4, 2007) and Zhou (US 2007/0129999 A1, pub. Jun. 7, 2007); and claim 3 as unpatentable over Desikan, Zhou, and Adjaoute (US 7,089,592 B2, iss. Aug. 8, 2006). ANALYSIS Appellants set forth various arguments asserting that the Examiner erred in rejecting independent claim 1 (App. Br. 7-11; Reply Br. 2-7). Appellants then generally reference these arguments in asserting that the Examiner erred in rejecting independent claims 11 and 16 (App. Br. 11-13; Reply Br. 7-8). As independent claim 11 appears to be the broadest claim, we choose it as representative of independent claims 1, 11, and 16. See 37 C.F.R. § 41.37(c)(1)(vii). We are persuaded the Examiner erred in determining that a combination of Desikan and Zhou teaches or suggests 2 The Examiner has withdrawn the rejections of claims 7, 15, and 19 under 35 U.S.C. § 112, first and second paragraphs (Ans. 2). As the Examiner makes no mention of the provisional non-statutory double patenting rejection of claims 1-21, we will treat those as also having been withdrawn. Appeal 2011-004821 Application 11/698,507 4 “generating, by a processor, a score for the service provider partner indicating a likelihood of the service provider partner being profitable when provided with advertisements of a category of advertisements to serve to users by applying the service provider partner feature vector to a classifier model,” as recited in independent claim 11 ). The Examiner asserts that the Bayesian networks and neural networks of Desikan correspond to the recited classifier model (Ans. 8, 10-12, 14-16). Specifically, the Examiner cites paragraphs [0055] and [0058] of Desikan, which discloses that “a manual list of Websites with a given type or class of policy violation may be used to train an expert system (e.g., neural networks, Bayesian networks, support vector machines, etc.) to classify other Websites as having the policy violation or not.” However, determining whether or not to classify a Website as having a policy violation in Desikan does not correspond to “generating, by a processor, a score for the service provider partner… by applying the service provider partner feature vector to a classifier model,” as recited in independent claim 11. In other words, a boolean “yes/no” determination of a policy violation in Desikan does not correspond to the generated score recited in independent claim 11, especially where the generated score is recited as being related to profitability. Other cited portions of Desikan do disclose generating quality scores for Websites. See, e.g., paragraphs [0022], [0038], [0051] (Ans. 7-8, 10, 12, 14-16). However, those portions of Desikan that disclose generating quality scores for Websites do not utilize the “expert system[s] (e.g., neural networks, Bayesian networks, support vector machines, etc.)” -- i.e., the “classifier model” -- of Desikan for generating the quality scores, as required by independent claim 11. Appeal 2011-004821 Application 11/698,507 5 DECISION The Examiner’s rejection of claims 1-21 is REVERSED. REVERSED mls