Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Replacement Final
This is to replace the previous Final rejection.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
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.
1. Claims 1, 3-9, 11-17 and 19-20 are rejected under 35 U.S.C. 101 as being directed to patent-ineligible subject matter.
The amended claims are directed to the abstract idea of mathematical optimization and recommendation selection, including:
receiving recommendation scores for a plurality of users and items,
generating a binary recommendation array,
maximizing a sum of entries of an elementwise product of the binary recommendation array and an array containing the recommendation scores,
applying target recommendation frequencies and other constraints,
updating the binary recommendation array based on evaluation of an objective function subject to constraints, and
recommending items based on the updated binary recommendation array.
See amended claim 1; corresponding system and medium claims 9 and 17; and dependent claims 3-8, 11-16, and 19-20.
The claims recite data gathering, mathematical manipulation, and optimization-based selection. The claim language frames the core advance as a mathematical allocation of recommendation outputs according to scores and constraints, which is an abstract idea. The recited processor, memory, and non-transitory computer-readable medium do not add significantly more than generic computer implementation of that abstract idea.
The Office Action of record already recognized that the claim set “plainly recites an abstract idea as it includes various mathematical formulas / relationships in prose,” although it found the invention not wholly directed to an abstract idea. The present amendment does not change that conclusion in a way that integrates the claim into a practical application. Rather, the amended language continues to focus on score-based optimization, binary selection, and constraint-based recommendation distribution. See Office Action at p. 4.
Accordingly, claims 1, 3-9, 11-17 and 19-20 are rejected under 35 U.S.C. 101.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
2. Claims 1, 3, 4, 9, 11, 12, 17, 19 and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Lam (US PG Pub. No. 2009/0006398 A1) in view of Chen (US PG Pub. No. 2011/0035379 A1).
3. Claims 5-8 and 13-16 are rejected under 35 U.S.C. 103 as being unpatentable over Lam and Chen, as applied above, and further in view of Smith (US Patent No. 8,117,085 B1).
As per claims 1, 9 and 17, Lam discloses a method, system and program, the method comprising:
-receiving, for a plurality of users, recommendation scores of a plurality of items generated using a recommender system based on one or more metrics, wherein the one or more metrics include a click through rate or an expected amount of purchase;
(Par. [0024], “For example, a particular recommender might retrieve the user’s purchase history data." The rate is not yet required; uses information about users’ purchases which would reflect the expected appetite of a particular buyer.)
-generating a binary recommendation array based on the recommendation scores;
(Par. [0027], “For instance, several recommenders may be used to recommend a particular war movie because 1) a user recently rated several war movies, 2) this is the best selling movie in the war movie category, and 3) this movie was nominated for two Academy Awards."; basing a recommendation on whether a user has recently rated similar items is considered a recommender system; a display with two parts is considered a binary array.)
-recommending, to each user of the plurality of users, one or more items of the plurality of items based on the binary recommendation array.
(Par. [0099], "Each recommender then generates a list of candidate items for the user, together with associated scores and reasons." In this example, candidate items are recommendations put forth by each individual recommender system before they are processed and put forth in the final recommendation list.)
To the extent that Lam does not fairly suggest the recited (math) function where each grade/score or value/worth is constant/steady or changeable/variable, respectively, Chen, in the field of e-commerce, teaches:
-wherein the binary recommendation array is generated based on an objective function that maximizes a sum of entries of an elementwise product of the binary recommendation array and an array containing the recommendation scores based on target recommendation frequencies for the plurality of items,
([0057]; [0067] item recommendation)
wherein the recommendation scores are treated as constants and each value in the binary recommendation array is treated as a variable
([0094-5]; [0101]; [0122] item variable)
It would have been obvious to one of ordinary skill in the art at the time of filing to modify Lam to use the function of Chen. One would have been motivated to make the combination to tweak the (buyer) recommendations in a meaningful way to achieve (seller) goals for volatile items which might be scarce due to market conditions, and thus merit less recurrent promotion.
Finally, the broadest reasonable interpretation of the claim could be that the middle step generates both x) a binary array, and y) a different (non-binary and/or more generic) array—“an array” (see 2d line of underlining). To the extent that the second array is not just an intermediate step, and is persistent, the examiner finds that it would have been obvious to one of ordinary skill in the art at the time of filing to ignore / disregard the second array in the capstone recommendation phase. Or, alternatively incorporate / use that array in view of the comprising transitional phrase, and as detailed above for the (math) function.
Claim 3 further recites that maximizing the sum of entries is based on a set of constraints including either a number of items to be recommended to each user or a lower and upper bound on a likelihood that each item is desired to be recommended. Lam teaches selecting a number of recommendations and ranking or filtering recommendations based on scores. See Office Action at p. 11, citing Lam para. [0040]. Chen teaches tuning recommendation output through optimization and constraints. It would have been obvious to apply such constraints to Lam in view of Chen to control the number and distribution of recommendations.
Claim 4 further recites that the lower bound and upper bound of the likelihood are scaled with respect to the number of items to be recommended. Lam teaches normalized scoring and recommendation range processing. See Office Action at pp. 9-11, citing Lam paras. [0076]-[0078]. Chen teaches tuning recommendation output using objective-based control. It would have been obvious to scale the lower and upper bounds relative to the number of items to be recommended in order to control recommendation distribution in a predictable manner.
Claim 11 is the system counterpart of claim 3 and recites the same constraint-based limitation in system form. For the reasons discussed with respect to claim 3, Lam and Chen render claim 11 obvious.
Claim 12 is the system counterpart of claim 4 and recites the same scaled-likelihood limitation in system form. For the reasons discussed with respect to claim 4, Lam and Chen render claim 12 obvious.
Claim 19 is the computer-readable medium counterpart of claim 3. Lam and Chen render claim 19 obvious for the same reasons discussed with respect to claim 3.
Claim 20 is the computer-readable medium counterpart of claim 4. Lam and Chen render claim 20 obvious for the same reasons discussed with respect to claim 4.
Claims 5-8 and 13-16 are rejected under 35 U.S.C. 103 as being unpatentable over Lam and Chen, and further in view of Smith.
Lam and Chen teach the core recommendation scoring, selection, and optimization framework discussed above. Smith teaches additional techniques including excluding certain customer choices from consideration, giving less weight to certain data, and filtering or sampling subsets of recommendation-related data. See Office Action at p. 13, citing Smith col. 9, lines 55-65.
It would have been obvious to modify Lam in view of Chen and Smith to implement the recited user-grouping, aggregation, sampling, exclusion, and iterative recommendation features in order to improve targeting, reduce noise, and control recommendation distribution.
Claim 5 recites identifying one or more groups of users based on attributes or clustering, identifying a particular group, and excluding recommendation scores corresponding to users not in the particular group. Lam teaches recommendation processing based on user-related information and score selection. Chen teaches controlling recommendations through optimization. Smith teaches excluding certain customer choices from consideration or giving them less weight. It would have been obvious to identify user groups and exclude scores corresponding to users outside a selected group in order to improve targeting and reduce noise.
Claim 6 recites identifying groups of users, aggregating associated recommendation scores for each group, replacing the recommendation scores with the aggregated associated recommendation scores, and assigning an updated binary recommendation array to each user based on the group to which the user belongs. Lam and Chen teach the core recommendation scoring and optimization framework, and Smith teaches filtering, weighting, and excluding data from consideration. It would have been obvious to combine these teachings to implement group-based aggregation and assignment of recommendation arrays for efficiency and recommendation control.
Claim 7 recites sampling, for each group, a subset of users and excluding recommendation scores corresponding to users not in the subset. The same combination of Lam, Chen, and Smith renders this limitation obvious, because Smith teaches selective exclusion and weighting of recommendation-related data, and it would have been obvious to sample subsets of users and exclude corresponding scores to improve efficiency and reduce noise.
Claim 8 recites that the binary recommendation array indicates at most one item to be recommended, determining that two or more items are to be recommended to each user, identifying a first updated binary recommendation, generating an additional constraint that the first updated binary recommendation and its associated recommendation score remain zero, updating the binary recommendation array subject to the additional constraint, identifying a second updated binary recommendation, and aggregating the first and second updated binary recommendations. Lam teaches candidate recommendation selection and ranking. Chen teaches constrained optimization. Smith teaches excluding previously selected items from further consideration. It would have been obvious to iteratively generate multiple binary recommendations by excluding prior selections and aggregating successive outputs to recommend more than one item while preserving a one-item-at-a-time selection framework.
Claim 13 is the system counterpart of claim 5. For the same reasons discussed with respect to claim 5, Lam, Chen, and Smith render claim 13 obvious.
Claim 14 is the system counterpart of claim 6. For the same reasons discussed with respect to claim 6, Lam, Chen, and Smith render claim 14 obvious.
Claim 15 is the system counterpart of claim 7. For the same reasons discussed with respect to claim 7, Lam, Chen, and Smith render claim 15 obvious.
Claim 16 is the system counterpart of claim 8. For the same reasons discussed with respect to claim 8, Lam, Chen, and Smith render claim 16 obvious.
Response to Arguments
Applicant’s arguments with respect to all pending claim(s) have been considered but are moot because the new ground of rejection does not rely on any rejection of record specifically challenged in the argument.
The amended claims continue to recite abstract score processing, optimization, and constrained recommendation selection, and do not recite a specific technological improvement. Further, Lam, Chen, and Smith collectively teach or suggest the recited tunable recommendation distribution concept. Accordingly, the amendment does not overcome the rejections under 35 U.S.C. 101 or 103.
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MICHAEL FUELLING whose telephone number is (571)270-1367.
The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/MICHAEL FUELLING/ Supervisory Patent Examiner
Prosecution Insights