DETAILED ACTION
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 .
Status of Claims
This action is in reply to the amendment filed on 12/11/2025.
Claims 1, 4, 7-8, 11, 14-15, 17, and 20 have been amended and are hereby entered.
Claims 1-20 are currently pending and have been examined.
This action is made FINAL.
Information Disclosure Statement
All references listed in the IDS dated 11/18/2025 have been considered.
Response to Applicant’s Arguments
Claim Rejections – 35 USC § 112
Regarding Claims 4, 11, and 17, the present amendments correct the antecedent basis issues for the terms “the intellectual property rights enforcement data” and “the authenticated memorabilia data,” but fail to do so for the term “the item listing system data.” As such, the previous 112(b) rejection of these claims is modified in view of the present amendments and maintained.
The present amendments to Claims 7, 14, and 20 obviate the previous 112(b) rejections thereto; therefore, these rejections are withdrawn.
Claim Rejections – 35 USC § 101
Applicant’s arguments regarding the 101 analysis have been considered and are unpersuasive.
Applicant asserts that the claims all fall within proper statutory categories as per Step 1 of the 101 subject matter eligibility analysis. This assertion is made absent any meaningful substantive discussion, and amounts to no more than an unsupported conclusory statement rather than a properly reasoned argument. While Applicant agrees that Claims 1-7 and 15-20 all properly fall within one of the enumerated statutory categories of Step 1, Examiner maintains the 101 signals per se-based rejections of Claims 8-14 issued in the Non-Final Rejection of 8/11/2025. Applicant makes no mention of these signals per se rejections in the present Remarks, and the present amendments to these claims fail to address these rejections.
None of the arguments presented by Applicant which are nominally regarding Step 2A, Prong One are indeed Step 2A, Prong One arguments. Rather, they are a string of Step 2A, Prong Two arguments, particularly regarding the consideration of an improvement to a technology. Even properly considering these arguments under the standards of Step 2A, Prong Two, none of these are persuasive arguments of technological improvements. Generally across each of these asserted improvements, Applicant fails to properly consider the distinction between judicial exceptions (here, abstract ideas) and additional elements. This is particularly important here as integration into a practical application (including by way of the improvement to a technology consideration) may only occur by way of any recited additional elements or the combination thereof (see, e.g., MPEP 2106.04(d) and 2106.05).
Regarding the first assertion, merely specifying the content of the training data analyzed in the present invention, particularly when both as claimed and as described in the specification, this information is entirely abstract in nature, does not evidence an improvement to a technology. As articulated in the Summary of the Interview of 10/15/2025 in relation to similar argument, “Examiner finds nothing inherently technological about the dataset used for training the ML model, especially given the broad manner in which they are []claimed.”
Regarding the second assertion, and similar to the abstract nature of the input data discussed above in relation to Applicant’s first Prong One assertion, specifying such data as “heterogeneous across modalities (text, image, transactional, contextual)” does nothing to make such data non-abstract, especially given examples of what this data is intended to encompass as described in the specification. Further, while the newly claimed feature-fusion data embedding in the model training step represents a non-abstract additional element, this does not embody a “[r]epresentation learning improvement” as asserted. Particularly, as conveyed to Applicant regarding this functionality in the Interview of 10/15/2025, “[c]ombining embeddings via feature-fusion is a pre-existing technique as discussed above, and as such does not constitute an improvement to a technology (or appear to evidence integration by way of any other consideration).” As such, one of ordinary skill in the art could not reasonably conclude that the incorporation of such pre-existing feature-fusion techniques would be an improvement to a technology as per Applicant’s effective filing date. “[P]atents that do no more than claim the application of generic machine learning to new data environments, without disclosing improvements to the machine learning models to be applied, are patent ineligible under § 101." Recentive Analytics, Inc. v. Fox. Corp., Fed Cir. No. 2023-2437 (Apr. 18, 2025) (slip op. at 18).
Regarding the third assertion, Applicant again fails to distinguish between abstract ideas and additional elements. While the training and use of a machine learning model constitute additional elements, the claimed data analysis and generation of results are abstract in nature. As explained to Applicant in the Interview of 10/15/2025 in response to similar arguments, the “high-level performance [of machine learning models and techniques] in proposed-amended Claim 1 does not show an improvement to a technology (such as machine learning), but rather uses a technology and pre-existing techniques associated therewith in a new environment or field of use, which is not sufficient to show subject matter eligibility.” Additionally, the “[i]nference-to-action” referenced in this argument remains abstract as claimed, as also explained to Applicant in the Interview of 10/15/2025.
Applicant’s single-sentence analogies to the Enfish and McRO cases, presented in conclusory manner absent any attempt at fact-to-fact or reasoning-to-reasoning analysis, does nothing to make the above-discussed arguments persuasive. Unlike the presently claimed invention and present arguments associated therewith, those inventions do improve a technology.
Regarding Applicant’s arguments both nominally and actually regarding Step 2A, Prong Two, these arguments suffer from some of the same failings discussed above, particularly claiming of standard machine learning techniques as somehow representing improvements to machine learning as it existed at Applicant’s date of filing, and a failure to recognize the distinction between judicial exceptions and additional elements.
Regarding the first Prong Two argument, the acquisition of multi-source inputs from authenticity verified domains as claimed is entirely abstract. While the construction of a ground-truth training set is a non-abstract additional element, this is an exceedingly standard and ubiquitous machine learning technique as of Applicant’s effective filing date, and as such does not represent an improvement to a technology or any other meaningful application of abstract ideas set forth in MPEP 2106.05(d).
Regarding the second Prong Two argument, “transform[ing] heterogeneous signals with NLP/CV/transactional encoders” is not embodied in the claims as presently drafted, and as such is entirely irrelevant. Regarding the claimed feature-fusion techniques, this fails to show integration into a practical application for the same reasons discussed above and in the Interview of 10/15/2025.
Regarding the third Prong Two argument, “[p]erform[ing] listing-level inference to produce authenticity confidence and analytics output” is an entirely abstract step, and describing such outputs as “machine-interpretable,” particularly in the extremely high-level manner in which this might be considered embodied by the claims as presently drafted, does nothing to make this otherwise. It has long been established that mere performance of abstract steps (including myriad examples of data analysis) by way of computers does not render such steps non-abstract. See, e.g., the seminal Alice case and MPEP 2106.04(a)(2)(III)(C).
Regarding the fourth Prong Two argument, the generation and communication of a brand-focused security notification associated with the analytics and authenticity representations, even as vaguely claimed as being the basis for further remedial actions in Claim 6, are entirely abstract ideas. The vague and high-level claiming of such steps occurring via computer elements does nothing to make this otherwise. Again, see e.g., the seminal Alice case and MPEP 2106.04(a)(2)(III)(C).
Applicant’s vague and conclusory analogies to the DDR Holdings and Finjan cases are no more effective than the similarly asserted references to Enfish and McRO addressed above. Unlike those cases, the presently claimed invention merely claims the performance of an entirely abstract analysis based on entirely abstract data resulting in entirely abstract outputs at a high level as performed via computer elements (including machine learning models and standard techniques associated therewith).
Regarding Step 2B, Applicant’s assertions of non-conventional activity both misapprehends this standard as well as continues to fail to distinguish between abstract ideas and additional elements in similar manner as the arguments addressed above. Further, these arguments contain several purported distinctions unembodied by the claims as presently drafted (e.g., “not the routine unlabeled/weakly labeled web crawls typical of prior systems;” “generation of multi-dimensional authenticity representations;” “a structured, machine-actionable output”), which are therefore irrelevant regardless of the equal applicability of the failures of these arguments discussed below.
The well-understood, routine, and conventional consideration of Step 2B applies only to claim elements categorized as additional elements, and further sub-categorized as insignificant extra-solution activity (see, e.g., the Step 2B analyses of Example 46, Claim 1 of the October 2019 PEG Update, and of Example 47, Claim 2 and Example 48, Claim 1 of the July 2024 PEG Update). The majority of the features referenced in Applicant’s Step 2B arguments are not additional elements but rather abstract concepts (e.g., the gathering of multi-dimensional data, including data representing heterogeneous data attributes, from a variety of sources; the use of such data in pattern-recognition-type data analyses to learn authentic brand characteristics; using output security analytics to generate a security notification; potentially taking further enforcement actions based on the content of such a notification, at least in Claim 6), and whether abstract ideas are well-understood, routine, and conventional is entirely irrelevant. Rather, the only claim element found to recite insignificant extra-solution activity is the embedding of data through feature fusion during model training, yet as discussed above (and as evidenced by the low level of detail provided for this element in Applicant’s own specification – see updated 101 rejections below for more information), such feature-fusion techniques were perfectly well-known and conventional in machine learning arts at the time of Applicant’s effective filing date, and as such could not reasonably be called an “[u]nconventional ML training approach.”
Further, Applicant returns to the improvement to a technology consideration by asserting that the purportedly unconventional elements asserted above “provide a technical solution to a technical problem (robust, cross-modal authenticity detection and automated action),” purportedly resulting in “improved accuracy, robustness, and operational reliability of brand-security systems.” However, “robust, cross-modal authenticity detection and automated action” is an abstract concept rather than a technological one, and the high-level performance of such abstract detection and action by way of computer elements does not make this otherwise (again, see Alice; see also the various Examples of the July 2024 PEG Update). Further, as similarly explained to Applicant in the Interview of 10/15/2025, “improved accuracy, robustness, and operational reliability of brand-security” are abstract results rather than technological ones, and merely couching this as a narrower improvement to “brand-security systems” rather than its proper context as improving the abstract concept of brand security more broadly does not make this otherwise.
Nothing in Applicant’s proffered arguments or the various citations to the specification (even if the full details of which were properly embodied in the claims as presently drafted, a great many of which are not) properly evidence an improvement to a technology as asserted.
Claim Rejections – 35 USC § 103
Applicant’s arguments regarding the 103 analysis have been considered and are unpersuasive.
Applicant’s arguments in view of the present amendments are moot in view of the updated 103 rejections below; see updated 103 rejections for more information.
Regarding the content of Applicant’s 103 rejections, Examiner makes some observations. Firstly, Applicant’s assertions solely consider the Venkatraman and Chase references in isolation, failing to properly consider them as modifying one another as cited in the previous and present 103 rejections. One cannot show nonobviousness by attacking references individually where the rejections are based on combinations of references (see In re Keller, 642 F.2d 413, 208 USPQ 871 (CCPA 1981); In re Merck & Co., 800 F.2d 1091, 231 USPQ 375 (Fed. Cir. 1986)). Secondly, many of Applicant’s arguments, in similar manner to several of the 101 arguments above and arguments presented in the Interview of 10/15/2025, assert a failure to disclose various features which are not embodied by the claims as presently drafted, and as such are irrelevant. For example, as stated in the Interview of 10/15/2025: “the amendments to the limitation for generating a brand-focused security notification do not narrow the content of that notification in the same way as argued (ie: ‘[t]he notification is not a generic report, but a computational output tied to model-based authenticity analytics’), but instead merely specify data which this [] notification is ‘based on’ and ‘associated with.’ Examiner does not find that this distinguishes the claim from the previously cited references.” Lastly, many of Applicant’s assertions of distinctions are simply untrue, even in view of the particular citations as provided in the previous Office Action. For example, regarding the assertion that “[n]o disclosure of a training dataset ‘generated from multiple authenticity-verified domains’ that provide ground truth inputs,” Venkatraman discloses the gathering of such data from a variety of third-party sources/subscriptions, and Chase discloses the gathering of such data from various manufacturers. Applicant is encouraged to take greater care in the future, both with consideration of what is actually embodied by the claim language and what is disclosed in the references cited against such language.
Claim Interpretation
Claims 2 and 9 contain the following limitation: “wherein the brand-focused machine learning model is integrated into a brand-focused listing management tool of the item listing system to support brand-focused listing management.” In this limitation, the language “to support brand-focused listing management” as drafted is considered intended use, and as such is not given patentable weight.
Claim Rejections – 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 4, 11, and 17 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claims 4, 11, and 17 each contain the terms “the intellectual property rights enforcement data,” “the authenticated memorabilia data,” and “the item listing system data.” These terms lack antecedent basis, as they are not previously disclosed within the respective claim strings. Examiner notes that first instances of these terms are present in Claims 3, 10, and 16; however, as Claims 4, 11, and 17 do not depend upon any of these claims, these instances cannot provide antecedent basis here. For the purposes of this examination, these terms are interpreted as “intellectual property rights enforcement data,” “authenticated memorabilia data,” and “item listing system data” respectively.
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.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Regarding Claims 1, 8, and 15, the limitations of accessing an item listing associated with an item listing service; analyzing the item listing using a brand-focused model that is trained based on a multi-dimensional authenticity analysis dataset and brand multi-dimensional authenticity features, wherein the multi-dimensional authenticity analysis dataset is generated from multiple authenticity-verified domains; wherein the brand multi-dimensional authenticity features comprise heterogeneous data attributes representing brand authenticity across multiple modalities, the heterogeneous data attributes are associated to represent authentic characteristics of a brand; based on analyzing the item listing using the brand-focused machine learning model, generating a brand-focused security notification associated with the item listing, wherein the brand-focused security notification is associated with brand-focused security analytics result data and multi-dimensional authenticity representations generated by the brand-focused machine learning model to indicate brand authenticity confidence; and communicating the brand-focused security notification, as drafted, are processes that, under their broadest reasonable interpretations, cover certain methods of organizing human activity. For example, these limitations fall at least within the enumerated categories of commercial or legal interactions and/or managing personal behavior or relationships or interactions between people (see MPEP 2106.04(a)(2)(II)).
Additionally, the limitations of accessing an item listing associated with an item listing service; analyzing the item listing using a brand-focused model that is trained based on a multi-dimensional authenticity analysis dataset and brand multi-dimensional authenticity features, wherein the multi-dimensional authenticity analysis dataset is generated from multiple authenticity-verified domains; wherein the brand multi-dimensional authenticity features comprise heterogeneous data attributes representing brand authenticity across multiple modalities, the heterogeneous data attributes are associated to represent authentic characteristics of a brand; based on analyzing the item listing using the brand-focused machine learning model, generating a brand-focused security notification associated with the item listing, wherein the brand-focused security notification is associated with brand-focused security analytics result data and multi-dimensional authenticity representations generated by the brand-focused machine learning model to indicate brand authenticity confidence; and communicating the brand-focused security notification, as drafted, are processes that, under their broadest reasonable interpretations, cover mental processes. For example, these limitations recite activity comprising observations, evaluations, judgments, and opinions (see MPEP 2106.04(a)(2)(III)).
Additionally, the limitation of analyzing the item listing using a brand-focused machine learning model that is trained based on a multi-dimensional authenticity analysis dataset and brand multi-dimensional authenticity features, wherein the multi-dimensional authenticity analysis dataset is generated from multiple authenticity-verified domains, as drafted, is a process that, under its broadest reasonable interpretation, covers mathematical concepts. For example, these limitations recite mathematical relationships and/or calculations (see MPEP 2106.04(a)(2)(I)).
If a claim limitation, under its broadest reasonable interpretation, covers fundamental economic principles or practices, commercial or legal interactions, managing personal behavior or relationships, or managing interactions between people, it falls within the “Certain Methods of Organizing Human Activity” grouping of abstract ideas. If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind or with the aid of pen and paper but for recitation of generic computer components, it falls within the “Mental Processes” grouping of abstract ideas. If a claim limitation, under its broadest reasonable interpretation, covers mathematical relationships, mathematical formulae or equations, or mathematical calculations, it falls within the “Mathematical Concepts” grouping of abstract ideas. Accordingly, the claims recite an abstract idea.
The judicial exception is not integrated into a practical application. In particular, the claim recites the additional elements of one or more computer-storage media having computer-executable instructions embodied thereon that are executable by a processor, one or more computer processors, computer memory storing computer-useable instructions that executable by the one or more computer processors, an item listing system, a brand-focused machine learning model, providing ground-truth inputs for training the brand-focused machine learning model, and embedding data through feature-fusion during model training. One or more computer-storage media having computer-executable instructions embodied thereon that are executable by a processor, one or more computer processors, computer memory storing computer-useable instructions that executable by the one or more computer processors, an item listing system, a brand-focused machine learning model, and providing ground-truth inputs for training the brand-focused machine learning model, in the context of the claims as a whole, amount to no more than mere instructions to apply a judicial exception (see MPEP 2106.05(f)). Embedding data through feature-fusion during model training, in the context of the claims as a whole, amounts to no more than insignificant extra-solution activity (see MPEP 2106.05(g)). Accordingly, these additional elements do not integrate the abstract ideas into a practical application because they do not, individually or in combination, impose any meaningful limits on practicing the abstract ideas. The claims are therefore directed to an abstract idea.
The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the judicial exception into a practical application, the additional elements amount to no more than mere instructions to apply a judicial exception, and insignificant extra-solution activity for the same reasons as discussed above in relation to integration into a practical application. The limitation found to recite insignificant extra-solution activity is further found to be well-understood, routine, and conventional in view of the standards of 112(a) as the high-level description of this functionality in the original disclosure (see at least Paragraphs 0065-0073 as filed) would be understood by one of ordinary skill in the art at the time of filing to indicate well-understood, routine, and conventional activity. These cannot provide an inventive concept. Therefore, when considering the additional elements alone and in combination, there is no inventive concept in the claims, and thus the claims are not patent eligible.
Claims 2-7, 9-14, and 16-20, describing various additional limitations to the system of Claim 1, product of Claim 8, or method of Claim 15, amount to substantially the same unintegrated abstract idea as Claims 1, 8, and 15 (upon which these claims depend, directly or indirectly) and are rejected for substantially the same reasons.
Claims 2 and 9 disclose wherein the brand-focused machine learning model is integrated into a brand-focused listing management tool of the item listing system to support brand-focused listing management (an abstract idea in the form of a certain method of organizing human activity and a mental process), which does not integrate the claims into a practical application.
Claims 3, 10, and 16 disclose wherein the multi-dimensional authenticity analysis dataset comprises intellectual property rights enforcement data, authenticated memorabilia data, and item listing system data that provide ground truth labeling and annotations associated with model training (further defining the abstract ideas already set forth in Claims 1, 8, and 15), which does not integrate the claims into a practical application.
Claims 4, 11, and 17 disclose wherein training the brand-focused machine learning model is based on machine learning techniques that fuse the brand multi-dimensional authenticity features that correspond to intellectual property rights enforcement data, authenticated memorabilia data, and the item listing system data (an abstract idea in the form of a certain method of organizing human activity, a mental process, and a mathematical concept), which does not integrate the claims into a practical application.
Claims 5, 12, and 18 disclose wherein generating the brand-focused security notification is based on identifying a predicted brand for the item listing, wherein the item listing does not include a brand input (an abstract idea in the form of a certain method of organizing human activity and a mental process), which does not integrate the claims into a practical application.
Claims 6, 13, and 19 disclose based on the brand-focused security notification, executing a listing quality analysis based in part on brand-focused security data associated with the brand-focused security notification (an abstract idea in the form of a certain method of organizing human activity and a mental process); and based on the listing quality analysis, flagging the item listing for remedial action (an abstract idea in the form of a certain method of organizing human activity and a mental process), which do not integrate the claims into a practical application.
Claims 7, 14, and 20 disclose wherein communicating the brand-focused security notification comprises communicating the brand-focused security notification to a seller associated with the item listing and a security administrator of the item listing system (further defining the abstract ideas already set forth in Claims 1, 8, and 15), which does not integrate the claims into a practical application.
Claims 8-14 are additionally rejected under 35 U.S.C. 101 because the claimed invention is directed to non-statutory subject matter. The claim(s) does/do not fall within at least one of the four categories of patent eligible subject matter because, under its broadest reasonable interpretation, Claims 8-14 encompass transitory forms of signal transmission, or signals per se (see In re Nuijten, 500 F.3d 1346, 84 USPQ2d 1495 (Fed. Cir. 2007) and MPEP 2106.03). In order to overcome these rejections, Examiner suggests amending “One or more computer-storage media having…” to read “One or more non-transitory computer-storage media having…”
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.
Claims 1-6, 8-13, and 15-19 are rejected under 35 U.S.C. 103 as being unpatentable over Venkatraman et al (PGPub 20160253679) (hereafter, “Venkatraman”) in view of Chase et al (WO 2024173092, claiming priority to US Provisional Application 63446554, filed 2/17/2023) (hereafter, “Chase”) and Chen et al (CN 114863230) (hereafter, “Chen”).
Regarding Claims 1, 8, and 15, Venkatraman discloses:
One or more computer-storage media having computer-executable instructions embodied thereon executable by a processor (¶ 0025, 0032; Figs. 1-2; server includes a processor that executes computer readable medium and instructions stored on the server (or elsewhere) to perform functions of the BAMS system; memory stores code (machine-readable or executable instructions) for an operating system);
one or more computer processors (¶ 0025; Figs. 1-2; server includes a processor that executes computer readable medium and instructions stored on the server (or elsewhere) to perform functions of the BAMS system);
computer memory storing computer-useable instructions executable by the one or more computer processors (¶ 0025, 0032; Figs. 1-2; server includes a processor that executes computer readable medium and instructions stored on the server (or elsewhere) to perform functions of the BAMS system; memory stores code (machine-readable or executable instructions) for an operating system);
accessing an item listing associated with an item listing system (¶ 0014, 0040, 0061, 0063; Figs. 3-4; data collection and brand abuse reporting system (DBAR) includes an input section comprising one or more collection engines adapted to harvest and crawl to obtain information available over internet websites and potentially other sources; collected detection data sets include marketplace listings, ecommerce websites, etc.);
based on analyzing the item listing using the brand-focused machine learning model, generating a brand-focused security notification associated with the item listing, wherein the brand-focused security notification is associated with brand-focused security analytics result data and multi-dimensional authenticity representations generated by the brand-focused machine learning model to indicate brand authenticity confidence (¶ 0029, 0035, 0052-0055, 0060, 0063-0064, 0066; Figs. 3-4, 6; BAMS system generates aggregated abuser lists and similarity scoring; a representative BAMS adapted to receive and process a received document having text content, image content, HTML content, and price content; image data or file is received by Image Matching module which may compare the received image with images stored in and accessed from image clusters storage or DB; text from document may be represented as a date file that is delivered to text match module, which may compare the text file, directly or after natural language processing and other steps are performed, with text strings or linguistic patterns stored in database or storage; document classifier processes document and key receives event ID data from element and is connected via nodes to matching module, document classifier, regular expression parser and text match module; the IDE matches the key attributes of threat vectors, enablers, and bad actors to the abuse database to see if there are matches; if a match is found the database is augmented with the relationships between the incident, the existing enablers, threat vectors, and bad actors; if a match is not found new entries are inserted into the database to capture the relationships between the incident, the enablers, threat vectors, and bad actors; a score is generated for bad actors based on the volume and breadth of incidents to indicate the level of online brand abuse they have conducted; data sets from the Attribute Matching module are delivered to Scoring Engine for further processing, including scoring; a High Value Target ID module receives a scored set of attribute matching datasets post-scoring from scoring engine for reporting by reporting module; the methods and systems of the present invention, described in detail hereafter, may be employed in providing remote users access to searchable databases and tools for setting up and receiving reports concerning brand monitoring and abuse or infringement detection; a report is generated listing the high value targets for a given brand; reporting by the reporting module, which may include particular user-defined or selected reporting formats; the reporting data set is delivered to user portal for presentation to a user such as by way of a connection of a client remote device); and
communicating the brand-focused security notification (¶ 0029, 0035, 0060, 0064; Figs. 3-4; BAMS system generates aggregated abuser lists and similarity scoring; the methods and systems of the present invention, described in detail hereafter, may be employed in providing remote users access to searchable databases and tools for setting up and receiving reports concerning brand monitoring and abuse or infringement detection; a report is generated listing the high value targets for a given brand; reporting by the reporting module, which may include particular user-defined or selected reporting formats; the reporting data set is delivered to user portal for presentation to a user such as by way of a connection of a client remote device).
Venkatraman additionally discloses analyzing the item listing using a brand-focused machine learning model that is trained based on a multi-dimensional authenticity analysis dataset, wherein the multi-dimensional authenticity analysis dataset is generated from multiple authenticity-verified domains (¶ 0014, 0025-0026, 0029, 0033, 0035, 0038-0039, 0041, 0055-0056, 0060; Figs. 1, 3-4; training/learning module; the entity identification module establishes an entity resolution layer and includes natural language processing, machine learning and image matching techniques to identify fields and patterns in collected data and categorize them using an ontology of entities relevant to brand protection; entity resolution layer—using data fusion, machine learning and image matching techniques, identifies fields and patterns in the data and classify and categorize them using an ontology of entities relevant to brand protection; the BAMS may be operated by a traditional professional services company, e.g., Thomson Reuters, wherein BAMS database corpus or set includes internal service or databases or sources of content such as TR Feeds; in addition, BAMS database set may be supplemented with external sources, freely available or subscription-based, as additional data considered by the IDE; news database or source may be a source for confirmed facts; also, government/regulatory filings database or source, USPTO, as well as other sources, provide data to the BAMS system for generating aggregated abuser lists and similarity scoring; an online information-retrieval system, such as offerings from Thomson Reuters Financial, Thomson IP, Westlaw, MarkMonitor, and other systems). Venkatraman does not explicitly disclose but Chase does disclose analyzing the item information using a brand-focused machine learning model that is trained based on a multi-dimensional authenticity analysis dataset and brand multi-dimensional authenticity features; wherein the multi-dimensional authenticity dataset is generated to provide ground-truth inputs for training the brand-focused machine learning model (¶ 0006, 0029, 0048-0049, 0056-0057, 0077, 0080, 0094, 0110; an identification algorithm of the one or more identification algorithms may be trained using a dataset of a plurality of images of golfballs, the dataset including one or more tags for each of the plurality of images, the one or more tags including at least one of a brand, a model, and a logo; one or more models may be trained using — or may otherwise retrieve this information for a condition analysis — one or more of: a plurality of images and/or scans of golfballs, which may be of different types, and which may include different conditions for comparison for identification by a model; data retrieved from manufacturer specifications and the like (such as data that indicates coloring, markings, shape, features, dimple properties, size, tolerances, material properties, and/or the like); and the like; one or more scans (and/or images) of a golf ball may be processed to identify (e.g., via machine learning models trained on voluminous images of similar objects having various attributes and features) attributes and features thereof, such as manufacturer, brand, model, damage, condition, cleanliness, advantages, disadvantages, third-party information, and so on; the database may be a golfball database including data such as a plurality of play-based attributes for each of a plurality of golfballs; such data may be created, added, removed, and/or revised in any of a number of ways, including automatically by data scraping from information provided by manufacturers or the like (e.g., web scraping to extract information from websites using automated tools or scripts to gather data from web pages), and/or manually by administrators, and so on).
Venkatraman additionally discloses wherein the brand multi-dimensional authenticity features comprise heterogeneous data attributes representing brand authenticity across multiple modalities, the heterogeneous data attributes are embedded through data-fusion during model training to represent authentic characteristics of a brand (¶ 0014, 0025-0026, 0038-0039, 0041, 0047, 0055-0056, 0060; Figs. 1, 3-4; training/learning module; the entity resolution layer uses natural language processing, machine learning and image matching techniques to identify fields and patterns in the data and categorize them using an ontology of entities relevant to brand protection; the resolution layer parses entities from unstructured text using text pattern matching techniques; this kind of entity may be resolved using supervised or unsupervised document classification techniques after transforming the text documents into numeric vectors: using multiple string fuzzy text pattern matching algorithms such as fuzzy Aho-Corasick; and using topic models such as Latent Dirichlet Allocation (LDA) and Hierarchical Dirichlet Processes (HDP); some entities represent a set of similar images, such as several variations (cropping, scaling, rotation, blur and other transformations used by infringers) of the same product picture; first, a matching algorithm computes a distance metric between any two images; second, a hierarchical clustering algorithm clusters the images in the leaves of a trie (prefix tree); each node is represented by the average of the feature vectors of all images in the children branches; a new image is matched to a cluster (leaf) by computing the distance of the new image to nodes starting by the root of the tree and going always towards the child with the smallest distance; the entity identification module establishes an entity resolution layer and includes natural language processing, machine learning and image matching techniques to identify fields and patterns in collected data and categorize them using an ontology of entities relevant to brand protection; entity resolution layer—using data fusion, machine learning and image matching techniques, identifies fields and patterns in the data and classify and categorize them using an ontology of entities relevant to brand protection). Venkatraman does not explicitly disclose but Chen does disclose wherein embedding data through data-fusion during model training is embedding data through feature-fusion (Abstract; pgs. 8-9, 14; an image processing method, false goods identification method and electronic device; because the target splicing feature is spliced by multiple independent features, in order to improve the identification rate, the plurality of independent features can be fused, so that the correlation between each independent feature in the feature after fusion is represented; therefore, before identifying the object to be identified, firstly the target splicing characteristic fusion, namely the target splicing characteristic in the mutually independent existing processing characteristic and processing characteristic to be combined into a characteristic form, so as to identify whether the object to be identified is a target object based on the fusion characteristic; the target splicing feature fusion step; can pre-train a fusion layer model, the fusion layer model may include one or more fusion layer and a full connection layer, each fusion layer can use linear rectification function ReLu as the activation function).
It would have been obvious to one of ordinary skill in the art before the filing date of the claimed invention to include the machine learning-based brand authentication techniques of Chase with the machine learning-based brand authentication system of Venkatraman because the combination merely applies a known technique to a known device/method/product ready for improvement to yield predictable results (see KSR Int’l Co. v. Teleflex, Inc., 550 U.S. 398, 415-421 (2007) and MPEP 2143). The known techniques of Chase are applicable to the base device (Venkatraman), the technical ability existed to improve the base device in the same way, and the results of the combination are predictable because the function of each piece (as well as the problems in the art which they address) are unchanged when combined. It would further have been obvious to one of ordinary skill in the art before the filing date of the claimed invention to include the machine learning-based brand authentication techniques of Chen with the machine learning-based brand authentication system of Venkatraman and Chase because the combination merely applies a known technique to a known device/method/product ready for improvement to yield predictable results (see KSR Int’l Co. v. Teleflex, Inc., 550 U.S. 398, 415-421 (2007) and MPEP 2143). The known techniques of Chen are applicable to the base device (Venkatraman and Chase), the technical ability existed to improve the base device in the same way, and the results of the combination are predictable because the function of each piece (as well as the problems in the art which they address) are unchanged when combined.
Regarding Claims 2 and 9, Venkatraman in view of Chase and Chen discloses the limitations of Claims 1 and 8. Venkatraman additionally discloses wherein the brand-focused machine learning model is integrated into a brand-focused listing management tool of the item listing system to support brand-focused listing management (¶ 0014, 0025-0026, 0032, 0035, 0038-0039, 0041, 0060; Figs. 1, 3-4; training/learning module; the entity identification module establishes an entity resolution layer and includes natural language processing, machine learning and image matching techniques to identify fields and patterns in collected data and categorize them using an ontology of entities relevant to brand protection; entity resolution layer—using data fusion, machine learning and image matching techniques, identifies fields and patterns in the data and classify and categorize them using an ontology of entities relevant to brand protection; access device, such as a client device, includes interface tools to provide user access to the infringement detection engine interface as in Fig. 1).
Regarding Claims 3, 10, and 16, Venkatraman in view of Chase and Chen discloses the limitations of Claims 1, 8, and 15. Venkatraman does not explicitly disclose but Chase does disclose wherein the multi-dimensional authenticity analysis dataset comprises intellectual property rights enforcement data and authenticated memorabilia data, that provide ground truth labeling and annotations associated with model training (¶ 0006, 0029, 0048, 0056-0057, 0080, 0094, 0110; an identification algorithm of the one or more identification algorithms may be trained using a dataset of a plurality of images of golfballs, the dataset including one or more tags for each of the plurality of images, the one or more tags including at least one of a brand, a model, and a logo; one or more models may be trained using — or may otherwise retrieve this information for a condition analysis — one or more of: a plurality of images and/or scans of golfballs, which may be of different types, and which may include different conditions for comparison for identification by a model; data retrieved from manufacturer specifications and the like (such as data that indicates coloring, markings, shape, features, dimple properties, size, tolerances, material properties, and/or the like); and the like; one or more scans (and/or images) of a golf ball may be processed to identify (e.g., via machine learning models trained on voluminous images of similar objects having various attributes and features) attributes and features thereof, such as manufacturer, brand, model, damage, condition, cleanliness, advantages, disadvantages, third-party information, and so on; supervised machine learning may find an association between data (e.g., feature vectors X) and a corresponding label (y, which can be categorical or continuous) so that the computer can learn an algorithm that maps the input to the output). Venkatraman additionally discloses wherein the multi-dimensional authenticity analysis dataset comprises intellectual property rights enforcement data and item listing system data (¶ 0014, 0025-0026, 0038-0039, 0041, 0055-0056, 0060; Figs. 1, 3-4; training/learning module; the entity identification module establishes an entity resolution layer and includes natural language processing, machine learning and image matching techniques to identify fields and patterns in collected data and categorize them using an ontology of entities relevant to brand protection; entity resolution layer—using data fusion, machine learning and image matching techniques, identifies fields and patterns in the data and classify and categorize them using an ontology of entities relevant to brand protection; the entity identification module establishes an entity resolution layer and includes natural language processing, machine learning and image matching techniques to identify fields and patterns in collected data and categorize them using an ontology of entities relevant to brand protection; collected data includes marketplace listings and intellectual property rights enforcement data).
The rationale to combine remains the same as for Claim 1.
Regarding Claims 4, 11, and 17, Venkatraman in view of Chase and Chen discloses the limitations of Claims 1, 8, and 15. Venkatraman additionally discloses wherein training the brand-focused machine learning model is based on machine learning techniques that fuse the collected data that corresponds with intellectual property rights enforcement data and item listing system data (¶ 0014, 0025-0026, 0038-0039, 0041, 0055-0056, 0060; Figs. 1, 3-4; training/learning module; the entity identification module establishes an entity resolution layer and includes natural language processing, machine learning and image matching techniques to identify fields and patterns in collected data and categorize them using an ontology of entities relevant to brand protection; entity resolution layer—using data fusion, machine learning and image matching techniques, identifies fields and patterns in the data and classify and categorize them using an ontology of entities relevant to brand protection; the entity identification module establishes an entity resolution layer and includes natural language processing, machine learning and image matching techniques to identify fields and patterns in collected data and categorize them using an ontology of entities relevant to brand protection; collected data includes marketplace listings and intellectual property rights enforcement data). Venkatraman does not explicitly disclose but Chase does disclose wherein training the brand-focused machine learning model is based on machine learning techniques that fuse the brand multi-dimensional authenticity features that correspond to the collected data which includes intellectual property rights enforcement data and the authenticated memorabilia data (¶ 0006, 0029, 0048, 0056-0057, 0080, 0094, 0110; an identification algorithm of the one or more identification algorithms may be trained using a dataset of a plurality of images of golfballs, the dataset including one or more tags for each of the plurality of images, the one or more tags including at least one of a brand, a model, and a logo; one or more models may be trained using — or may otherwise retrieve this information for a condition analysis — one or more of: a plurality of images and/or scans of golfballs, which may be of different types, and which may include different conditions for comparison for identification by a model; data retrieved from manufacturer specifications and the like (such as data that indicates coloring, markings, shape, features, dimple properties, size, tolerances, material properties, and/or the like); and the like; one or more scans (and/or images) of a golf ball may be processed to identify (e.g., via machine learning models trained on voluminous images of similar objects having various attributes and features) attributes and features thereof, such as manufacturer, brand, model, damage, condition, cleanliness, advantages, disadvantages, third-party information, and so on; supervised machine learning may find an association between data (e.g., feature vectors X) and a corresponding label (y, which can be categorical or continuous) so that the computer can learn an algorithm that maps the input to the output).
The rationale to combine remains the same as for Claim 1.
Regarding Claims 5, 12, and 18, Venkatraman in view of Chase and Chen discloses the limitations of Claims 1, 8, and 15. Venkatraman does not explicitly disclose but Chase does disclose wherein generating the brand-focused security notification is based on identifying a predicted brand for the item information, wherein the item information does not include a brand input (¶ 0104, 0129-0131; Figs. 5, 8; a user scans a golfball, e.g., without any identification; a first screen may be displayed once a scan of a golfball has been processed, including the identified brand name/manufacturer thereof). Venkatraman additionally discloses wherein the item information comes from the item listing (¶ 0014, 0040, 0061, 0063; Figs. 3-4; data collection and brand abuse reporting system (DBAR) includes an input section comprising one or more collection engines adapted to harvest and crawl to obtain information available over internet websites and potentially other sources; collected detection data sets include marketplace listings, ecommerce websites, etc.).
The rationale to combine remains the same as for Claim 1.
Regarding Claims 6, 13, and 19, Venkatraman in view of Chase and Chen discloses the limitations of Claims 1, 8, and 15. Venkatraman additionally discloses:
based on the brand-focused security notification, executing a listing quality analysis based in part on brand-focused security data associated with the brand-focused security notification (Abstract; ¶ 0011-0012, 0014, 0029, 0037-0039, 0060, 0065; a scoring module adapted to generate a set of score data based on the set of comparison data; the BAMS extracts and identifies relevant to brand protection from structured and unstructured data sources; BAMS, directly or in combination with other services, 1) finds relations between those entities, 2) finds, identifies and links online and physical entities in order to expose and collect evidence from criminal/infringing networks responsible of the abuse our clients are subject to, 3) provides a search functionality ranked by a relevance metric adapted to brand protection, 4) provides smart group-by capability allowing bulk action in brand protection, 5) classifies entities by abuse category, 6) classifies entities by targeted brand, 7) provides ways to measure the impact of criminal actors on our clients and compare this impact with other brands in the same industry, etc.; a Natural Language Processor module is used to identify text strings or other data from the collected data to determine potential brand abuse, e.g., unauthorized use of Nike, Gucci, Rolex, and other well known marks; the NLP may be used to identify text strings known to be protected in some form, e.g., copyright, for potential notice and enforcement action; a score is generated for bad actors based on the volume and breadth of incidents to indicate the level of online brand abuse they have conducted; bad actors that exceed a configurable level of abuse are identified as “High Value Targets”); and
based on the listing quality analysis, flagging the item listing for remedial action (Abstract; ¶ 0011-0012, 0014, 0029, 0037-0039, 0060, 0065; a scoring module adapted to generate a set of score data based on the set of comparison data; the BAMS extracts and identifies relevant to brand protection from structured and unstructured data sources; BAMS, directly or in combination with other services, 1) finds relations between those entities, 2) finds, identifies and links online and physical entities in order to expose and collect evidence from criminal/infringing networks responsible of the abuse our clients are subject to, 3) provides a search functionality ranked by a relevance metric adapted to brand protection, 4) provides smart group-by capability allowing bulk action in brand protection, 5) classifies entities by abuse category, 6) classifies entities by targeted brand, 7) provides ways to measure the impact of criminal actors on our clients and compare this impact with other brands in the same industry, etc.; a Natural Language Processor module is used to identify text strings or other data from the collected data to determine potential brand abuse, e.g., unauthorized use of Nike, Gucci, Rolex, and other well known marks; the NLP may be used to identify text strings known to be protected in some form, e.g., copyright, for potential notice and enforcement action; a score is generated for bad actors based on the volume and breadth of incidents to indicate the level of online brand abuse they have conducted; bad actors that exceed a configurable level of abuse are identified as “High Value Targets”).
Claims 7, 14, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Venkatraman in view of Chase, Chen, and Silverman et al (WO 2022108886) (hereafter, “Silverman”).
Regarding Claims 7, 14, and 20, Venkatraman in view of Chase and Chen discloses the limitations of Claims 1, 8, and 15. Venkatraman does not explicitly disclose but Silverman does disclose wherein communicating the brand-focused security notification comprises communicating the brand-focused security notification to a seller associated with the item listing and a security administrator of the item listing system (¶ 0049, 0094-0095, 0116, 0134; Figs. 4, 8; a system and authentication technique for integrating product authenticity verifications into product listings in real (offline) and/or virtual (online) marketplaces; this authenticity checking service may provide a standard and a secure process that allows a seller of a product and an interested buyer in an online (or offline) marketplace to verify the product authenticity through a trusted third party; once the authentication modules have completed their work, the results may be directed to marketplace and client).
The rationale to combine Venkatraman and Chase remains the same as for Claim 1. One of ordinary skill in the art would further have been motivated to include the marketplace product authentication communication techniques of Silverman with the machine learning-based brand authentication system of Venkatraman, Chase, and Chen to make participants and relevant parties aware of whether a product is genuine or counterfeit, and to facilitate improved supply-chain management (e.g., by reducing confusion, errors and/or malicious actions, as well as the associated expenses) (see at least Paragraphs 0002-0003, 0053, 0049, and 0094-0095 of Silverman).
Discussion of Prior Art Cited but Not Applied
For additional information on the state of the art regarding the claims of the present application, please see the following documents not applied in this Office Action (all of which are prior art to the present application):
PGPub 20170206574 – “Method of, and System for, Preventing Unauthorized Products from Being Sold on Online Sites,” Sharma et al, disclosing a system for determining authenticity of products listed for sale in online marketplaces
PGPub 20240185079 – “Semi-Supervised Similarity-Based Clustering in Resource Evaluation,” Singh et al, disclosing a system for utilizing machine learning to determine authenticity of products for sale
PGPub 20210248624 – “System, Device, And Method of Protecting Brand Names,” Keren et al, disclosing a system for utilizing machine learning to determine potential fraud in relation to branding in online marketplaces
Gunawardhana et al, Effectiveness of Machine Learning Algorithms on Battling Counterfeit Items in E-commerce Marketplaces, IEEE 2023 Int’l Research Conference on Smart Computing and Systems Engineering (SCSE), Vol. 6, pgs. 1-7 (published 8/18/2023)
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 MARK C CLARE whose telephone number is (571)272-8748. The examiner can normally be reached Monday-Friday 6:30am-2:30pm EST.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Jeffrey Zimmerman can be reached at (571) 272-4602. 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.
/MARK C CLARE/Examiner, Art Unit 3628
/MICHAEL P HARRINGTON/Primary Examiner, Art Unit 3628