DETAILED ACTION
This communication is a Final Office Action on the merits in response to communications received on 11/25/2025. Claims 1, 12, and 18 have been amended. Therefore, claims 1-20 are pending and have been addressed below. The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim Rejections - 35 USC § 101
1. 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.
2. Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract without significantly more.
3. Under Step 1 of the two-part analysis from Alice Corp, claim 1 recites a process (i.e., a series of acts or steps), claim 12 recites a machine (i.e., a concrete thing, consisting of parts, or of certain devices and combination of devices.), and claim 18 recites a manufacture (i.e., an article that is given a new form, quality, property, or combination through man-made or artificial means.). Thus, each of the claims fall within one of the four statutory categories.
4. Under Step 2A – Prong One of the two-part analysis from Alice Corp, the claimed invention is directed to an abstract idea.
5. Claim(s) 1, 12, and 18 recite:
“classify whether a post is a noncomparative or comparative post…a plurality of training samples, each training sample generated from the corpus of historical posts;”, “inputting
different portions…and comparing predictions of customer actions with target values of the training samples to adjust weights;”, “in response to receiving a product analysis request…, retrieving…a plurality of posts from the corpus of historical posts;”, “filtering…the plurality of posts, wherein the filtering identifies one or more posts which discuss products which are of interest for the product analysis request;”, “generating…a feature vector representing the one or more posts;”, “inputting…the feature vector;”, “classifying…individual posts of the filtered posts as a comparative post or a noncomparative post, wherein the posts classified as comparative posts compare two or more products and are used for the product analysis;”, “extracting…information from the comparative posts, wherein the information is to be used for the product analysis, wherein extracting the information comprises, for each comparative post, identifying product identifiers and attributes of the two or more compared products and determining which of the compared products is recommended and a reason for the recommendation;” ,“generating…a product analysis report based on the information extracted from the comparative posts;”
The limitations above recite an abstract idea of identifying and classifying information from postings of products to generate a comparative product analysis report which encompasses commercial interactions (i.e., advertising, marketing or sales activities, business relations), mental processes (i.e., observations, evaluations, judgments, and opinions) which falls within the certain methods of organizing human activity and mental processes groupings of abstract ideas. See MPEP 2106.04 II
The Applicant’s Specification in at least [0019] Product reviews can be very helpful to users in making the decision to purchase a specific product. Product reviews can also bring new opportunities for product feature and configuration analysis to organizations, such as companies that manufacture products. For example, customer product reviews posted on social media platforms are a potentially valuable information that can be used for understanding the strengths and weaknesses of an organization’s product and competing products as well as the product features that are important to users in making a purchase decision. Unfortunately, product reviews, and user generated content (UGC) in general, typically include extremely broad and freeform textual content. Accordingly, UGC is traditionally difficult to search to discover posts which compare a product which is of interest (or “focal product”) and competing products, and to extract information from such posts.
As such, the limitations when viewed as a whole recite sales or marketing activities and business relations that product managers or researchers typically perform to gain knowledge regarding their products or brands and for supporting future decision making. The series of steps carry out a user request for a comparative product analysis which allows the user to compare products and understand insights from various user-generated postings. Also, the limitations recite “filtering”, “generating”, “inputting” and “extracting” which are mental processes for parsing and comparing known data associated with the postings and “classifying” which is a mental process for recognizing information from collected information, i.e., determining whether the evaluated information from the postings compares two or more products for inclusion in the product analysis report. As such, the steps recite mental comparisons that can be practically performed in the human mind with or without the use of pen and paper. Therefore, the limitations may be reasonably characterized as subject matter falling within the certain methods of organizing human activity and mental processes groupings of abstract ideas. The recited limitations in the claim(s) is/are an abstract idea.
6. Under Step 2A – Prong Two of the two-part analysis from Alice Corp, this judicial exception is not integrated into a practical application because the additional elements of: “by a computing device”, “one or more first interfaces”, “one or more online forums, social media platforms, or other websites”, “by the computing device”, “a second interface”, “a client device”, “continuously training, by the computing device”, “a machine learning (ML) model implemented as a multi-layer neural network having a plurality of trainable parameters”, “using a modeling dataset comprising”, “wherein training the ML model includes”, “in the ML model”, “from the client device over the second interface”, “to the ML model”, “by the computing device” – see claim 1, “a computing device comprising:”, “one or more non-transitory machine-readable mediums configured to store instructions;”, “one or more processors configured to execute the instructions stored on the one or more non-transitory machine-readable mediums, wherein execution of the instructions causes the one or more processors to carry out a process comprising:”, “a non-transitory machine-readable medium encoding instructions that when executed by one or more processors cause a process to be carried out, the process including:” – see claims 12 and 18 is/are recited at a high-level of generality in light of the specification, [i.e., See Figs. 1-2]. The specification discusses the additional elements generically, without describing any of the particulars, such that the additional elements may be broad but reasonably construed as generic computer components being used to perform the judicial exception. As such, the additional elements merely add the words “apply it” with the judicial exception or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea as discussed in MPEP 2106.05 (f).
The other additional elements of: “continuously scanning…for posts related to one or more products of interest to generate a corpus of historical posts;”, “updating…the corpus of historical posts with the extracted information”, “transmitting…the product analysis report …with instructions to display the information extracted from the comparative posts.” add insignificant extra-solution activity to the judicial exception, i.e., mere data gathering/storage/output, but do not add any meaningful limitations to the claimed invention, as discussed in MPEP 2106.05 (g)
7. Thus, the additional claim elements are not indicative of integration into a practical application, because the claims do not involve improvements to the functioning of a computer, or to any other technology or technical field (MPEP 2106.05(a)), the claims do not apply or use the abstract idea to effect a particular treatment or prophylaxis for a disease or medical condition (Vanda Memo), the claims do not apply the abstract idea with, or by use of, a particular machine (MPEP 2106.05(b)), the claims do not effect a transformation or reduction of a particular article to a different state or thing (MPEP 2106.05(c)), and the claims do not apply or use the abstract idea in some other meaningful way beyond generally linking the use of the abstract idea to a particular technological environment, such that the claim as a whole is more than a drafting effort designed to monopolize the exception (MPEP 2106.05(e) and Vanda Memo). Therefore, the claims do not, for example, purport to improve the functioning of a computer. Nor do they effect an improvement in any other technology or technical field. Accordingly, the additional elements do not impose any meaningful limits on practicing the abstract idea and the claims are directed to an abstract idea.
8. The claim(s) does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception because, as discussed above with respect to integration of the abstract idea into a practical application, the additional element(s) of: “by a computing device”, “one or more first interfaces”, “one or more online forums, social media platforms, or other websites”, “by the computing device”, “a second interface”, “a client device”, “continuously training, by the computing device”, “a machine learning (ML) model implemented as a multi-layer neural network having a plurality of trainable parameters”, “using a modeling dataset comprising”, “wherein training the ML model includes”, “in the ML model”, “from the client device over the second interface”, “to the ML model”, “by the computing device” – see claim 1, “a computing device comprising:”, “one or more non-transitory machine-readable mediums configured to store instructions;”, “one or more processors configured to execute the instructions stored on the one or more non-transitory machine-readable mediums, wherein execution of the instructions causes the one or more processors to carry out a process comprising:”, “a non-transitory machine-readable medium encoding instructions that when executed by one or more processors cause a process to be carried out, the process including:” – see claims 12 and 18 amount to no more than mere instructions in which to apply the judicial exception and does not provide an inventive concept at Step 2B.
The other additional elements of “continuously scanning…for posts related to one or more products of interest to generate a corpus of historical posts;”, “updating…the corpus of historical posts with the extracted information”, “transmitting…the product analysis report …with instructions to display the information extracted from the comparative posts.” were considered to be insignificant extra-solution activity in Step 2A, and thus re-evaluated in Step 2B to determine if it is more than well-understood, routine, conventional activity in the field.
The Symantec, TLI, OIP Techs, Alice, Ultramerical, Versata Dev. Group court decisions cited in MPEP 2106.05(d)(II) indicate: “receiving or transmitting data over a network”, “electronic recordkeeping”, “storing and retrieving information in memory” is/are well-understood, routine, conventional activity. Thus, at Step 2B the claim(s) are ineligible.
9. Claims 2-11, 13-17, 19-20 are dependents of claims 1, 12 and 18.
Claims 2, 9, 13, 17 and 19 recite “wherein the filtering includes filtering content of the plurality of posts using keywords indicative of comparison and names of products which are of interest for the product analysis; wherein the filtering includes filtering content of the plurality of posts using keywords indicative of comparison and names of products which are of interest for the product analysis and lexico-syntactic pattern matching.” which further narrows how the abstract idea may be performed but does not make the claim any less abstract, Claim 3 recites “wherein at least one keyword of the keywords is derived from historical posts.” which further narrows how the abstract idea may be performed but does not make the claim any less abstract, Claims 4 and 14 recite “wherein the one or more posts which discuss products which are of interest for a product analysis are identified based on a threshold match percentage of the keywords indicative of comparison and the names of products which are of interest for the product analysis.” which further narrows how the abstract idea may be performed but does not make the claim any less abstract, Claims 5 and 20 recite “wherein the filtering includes filtering content of the plurality of posts using lexico-syntactic pattern matching.” which further narrows how the abstract idea may be performed but does not make the claim any less abstract, Claims 6, 7, 8 and 16 recite “wherein the lexico-syntactic pattern matching is of titles of the plurality of posts; wherein the lexico-syntactic pattern matching includes applying one or syntactic rules to titles of the plurality of posts; wherein at least one of the one or more syntactic rules is derived from historical comparative posts all serve to further narrow how the judicial exception may be performed and/or describe the information recited in the judicial exception. Claim 10 recites “wherein the ML model is trained using instances of historical posts labeled as either comparative or noncomparative” is recited at a high-level of generality and merely adds generic computer components and functionality to the abstract idea. Claim 11 recites “retrieving…the information extracted from the comparative posts; and generating…a product analysis report using the information retrieved from the data repository.” which further narrows how the abstract idea may be performed but does not make the claim any less abstract. Accordingly, when viewed individually and in combination, none of the dependent claims provide an inventive concept or integrate the abstract idea into a practical application.
Claim Rejections - 35 USC § 103
10. 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.
11. 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 non-obviousness.
12. Claim(s) 1-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Sureshkumar (US 2022/0092651 A1) in view of Alexandrov (US 11,699,177 B1) in further view of Garrow (US 2016/0092793 A1)
With respect to claims 1, 12, and 18, Sureshkumar discloses
a method (¶ 0108: discloses a method), a computing device (abstract, ¶ 0027, 0107: discloses provides a system which effectively and efficiently extracts insights on user perception regarding product or other items.),
one or more non-transitory machine-readable mediums configured to store instructions (¶ 0113-0114); and one or more processors (¶ 0107, 0115: discloses a processor) configured to execute the instructions stored on the one or more non-transitory machine-readable mediums,
wherein execution of the instructions causes the one or more processors to carry out a process comprising:
providing, by the computing device, a second interface to a client device (¶ 0050-0051, 0053: discloses during operation user 112 can determine via display 114 and device 102 to obtain insights relating to a specific product or products.);
continuously training, by the computing device, a machine learning (ML) model to classify whether a post is a noncomparative or comparative post (¶ 0086-0087, 0098: discloses the system can train the model to determine whether a review is relevant or not. Not all reviews/opinions in an opinion set may be relevant to a particular analysis. For example, in a set of reviews for specific cell phone, some of the reviews may not address the cell phone itself, but instead may address issues with customer service, shipping, delivery, etc. The model of the system can therefore be retrained by fine-tuning the existing model based on corrections.)
using a modeling dataset comprising a plurality of training samples (¶ 0086-0087: discloses the system can use a training or test dataset), each training sample generated from the corpus of historical posts (¶ 0086-0087: discloses the training dataset includes reviews over diverse sets of domains, product, platforms.);
wherein training the ML model includes inputting different portions of the modeling dataset and comparing predictions of customer actions with target values of the training samples to adjust weights in the ML model (¶ 0098: discloses the model can learn from the users to adjust the analysis and provide updated qualitative insights. The model can be retrained based on corrections to the model predictions by fine-tuning the existing model based on corrected labels.);
in response to receiving a product analysis request from the client device over the second interface (¶ 0051-0052, 0099: discloses the system receives a request for insights based on reviews for a product via display 114.),
retrieving, by the computing device, a plurality of posts from the corpus of historical posts (¶ 0032, 0033, 0050: discloses using readily available unstructured data such as reviews of products or other items. Device 104 can obtain from other networked entities and store on storage device 106 reviews from various websites of companies, brands, products, and other items.),;
filtering, by the computing device, the plurality of posts (¶ 0037, 0086: discloses the system narrows down a sets of reviews by performing context filtering.),
wherein the filtering identifies one or more posts which discuss products which are of interest for the product analysis request (¶ 0004, 0036, 0039, 0059, 0079, 0086: discloses the system filters the reviews based on a context associated with each reviews and automatically identifies desired feature information as terms of interest and dominant terms or aspects in the reviews.);
classifying, by the computing device using the ML model, individual posts of the filtered posts as a comparative post or a noncomparative post (¶ 0086-0087: discloses the system uses a binary classification to determine whether a review is relevant or not. The task of the model can be treated as a binary classification task, i.e., relevant or not relevant.),
wherein the posts classified as comparative posts compare two or more products and are used for the product analysis (¶ 0051, 0087: discloses the device 102 can display a select products for insight evaluation which includes various companies and products for selection. The system can filter out reviews which are out of context and use reviews from diverse sets of products which can be labeled as relevant.);
The Sureshkumar reference does not explicitly disclose the following limitations. In the same field of endeavor, the Alexandrov reference is related to systems and methods for automated modeling of quality for a product and for products of competitors (col. 2:1-11) and teaches:
providing, by a computing device, one or more first interfaces to one or more online forums, social media platforms, or other websites (col. 3:30-38, col.5:8-50, col. 6:1-19: discloses crawlers 201 may issue HTTP Get requests or other requests to access each of the review sites and source sites.);
continuously scanning, by the computing device, the one or more online forums, social media platforms, or other websites over the one or more first interfaces for posts related to one or more products of interest to generate a corpus of historical posts (col. 3:30-38, col.5:8-50, col. 6:1-19: discloses crawlers 201 may continuously run and may periodically scrape review and source sites for new or updated reviews and/or descriptive characteristics for a set of products or services.);
implemented as a multi-layer neural network having a plurality of trainable parameters (col. 6:20-63: discloses contextual relevance neural network 205 which represents a multi-layer neural network may use long short term memory networks “LSTM” or convolutional neural networks “CNN” to determine relevance of features, functionality, qualities, attributes, and/or other descriptive words to different classes of products. The contextual neural network may be trained based on descriptive characteristics from products that compete with or are similar to the particular product. The contextual neural network may detect patterns and may determine that the first feature is relevant for the particular class of products, whereas the second feature is not relevant feature for the particular class of products.)
extracting, by the computing device, information from the comparative posts, wherein the information is to be used for the product analysis (col. 6:1-19: discloses identifies one or more competing or similar products and descriptive characteristics of the competing or similar products for comparative purposes.),
wherein extracting the information comprises, for each comparative post, identifying product identifiers (col. 3:27-29: discloses the particular product may be identified by its name and/or other product identifiers, i.e., model number, serial number, etc.) and attributes of the two or more compared products (col. 6:1-19) and
determining which of the compared products is recommended (col. 13:1-10: discloses selecting the closet competitors for a product.) and a reason for the recommendation (col. 13:1-33: discloses may notify an entity of new competition, competition that is gaining in popularity, and/or competition that was previously unknown.);
updating, by the computing device, the corpus of historical posts with the extracted information (col. 6:14-19: discloses store the downloaded or retrieved reviews and descriptive characteristics to database 203. Database 203 includes different records for different products or services and may populate the record of the particular product with reviews and descriptive characteristics retrieved for the particular product.);
generating, by the computing device, a product analysis report based on the information extracted from the comparative posts (col. 4:55-63, col. 13:11-33: discloses quality assessment system 100 may generate and provide comparative assessments of the particular product to competing products or competitors.); and
transmitting, by the computing device, the product analysis report over the second interface to the client device with instructions to display the information extracted from the comparative posts. (col. 4:55-63, col. 13:11-33: discloses generating a visualization or interface with which to present the closet competitors. For instance, a user may view a first set of accounting products as related and identify a second set of accounting products that have a greater number of features in common in the aggregated reviews.)
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Sureshkumar system and methods for generating insights from reviews, to include steps for providing, by a computing device, one or more first interfaces to one or more online forums, social media platforms, or other websites; continuously scanning, by the computing device, the one or more online forums, social media platforms, or other websites over the one or more first interfaces for posts related to one or more products of interest to generate a corpus of historical posts; implemented as a multi-layer neural network having a plurality of trainable parameters; extracting, by the computing device, information from the comparative posts, wherein the information is to be used for the product analysis, wherein extracting the information comprises, for each comparative post, identifying product identifiers and attributes of the two or more compared products and determining which of the compared products is recommended and a reason for the recommendation; updating, by the computing device, the corpus of historical posts with the extracted information; generating, by the computing device, a product analysis report based on the information extracted from the comparative posts; and transmitting, by the computing device, the product analysis report over the second interface to the client device with instructions to display the information extracted from the comparative posts, as disclosed by Alexandrov to achieve the claimed invention. As disclosed by Alexandrov, the motivation for the combination would have been to provide meaningful insight to users regarding experiences with a product as well as providing more accurate neural network model by gaining a greater knowledge of descriptive characteristics from user reviews of competing products. (col. 1:8-20 and col. 7:20-50)
The combination of Sureshkumar and Alexandrov does not explicitly disclose the following limitations. In the same field of endeavor, the Garrow reference is related to filtering and classifying social media data that can be used to identify trends and relationships in connection with pharmaceuticals, professional and patient feedback on drugs, medical devices, etc. (¶ 0006, Fig. 7) and teaches:
generating, by the computing device, a feature vector representing the one or more posts (¶ 0041-0042, 0048-0049, 0052-0054: discloses the feature extractor 120 extracts feature vector 125 from message 115);
inputting, by the computing device, the feature vector to the ML model (¶ 0054, 0058: discloses the filtered message 115 and its associated feature vector 125 are provided to the classifier 130. Classifier 130 is made up of one or more machine learning models 140 each of which has been trained to recognize feature vectors 125 that belong to a particular class of messages. The feature vectors 125 of filtered messages 115 are first input into first machine learning model 140a.);
As can be seen from the Garrow reference, techniques for generating and inputting feature vectors into machine learning model were known in the state of the art and previously in the industry used to categorize social media posts containing information about pharmaceuticals and medical devices into groups or classes based on their features.
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the combined system and methods of Sureshkumar and Alexandrov, to include steps for generating, by the computing device, a feature vector representing the one or more posts; inputting, by the computing device, the feature vector to the ML model, as disclosed by Garrow to achieve the claimed invention. As disclosed by Garrow, the motivation for the combination would have been to provide the ability to recognize social media posts with features relevant to particular products of interest and allow users to identify trends and relationships in the related information, as expressly suggested by Garrow. (¶ 0005-0006, 0012, 0054)
With respect to claims 2, 13, and 19, the combination of Sureshkumar, Alexandrov, and Garrow discloses the method, the computing device, and the machine-readable medium,
wherein the filtering includes filtering content of the plurality of posts using keywords indicative of comparison and names of products which are of interest for the product analysis. (¶ 0036, 0075: Sureshkumar discloses using a set of predetermined phrases or statements and/or desired feature information as a list of terms of interest which appear in the reviews.)
With respect to claim 3, the combination of Sureshkumar, Alexandrov, and Garrow discloses the method of claim 2,
wherein at least one keyword of the keywords is derived from historical posts. (¶ 0077: Sureshkumar discloses the system can identify and provide a suggested aspect name.)
With respect to claims 4 and 14, the combination of Sureshkumar, Alexandrov, Garrow discloses the method and the computing device,
wherein the one or more posts which discuss products which are of interest for a product analysis are identified based on a threshold match percentage of the keywords indicative of comparison and the names of products which are of interest for the product analysis. (cols. 12-13:41-67 and 1-10: Alexandrov discloses determining the number or percentage of descriptive characteristics for the different products have in common.)
With respect to claims 5, 15, and 20, the combination of Sureshkumar, Alexandrov, and Garrow discloses the method, the computing device, and the machine-readable medium,
wherein the filtering includes filtering content of the plurality of posts (¶ 0033, 0065: Sureshkumar discloses the system takes reviews of a product as a stream of data and parses the reviews into attributes such as the title.) using lexico-syntactic pattern matching. (¶ 0032: Sureshkumar discloses the system uses natural language processing to extract insights from reviews across various features or aspects of a product.)
With respect to claim 6, the combination of Sureshkumar, Alexandrov, and Garrow disclose the method of claim 5,
wherein the lexico-syntactic pattern matching is of titles of the plurality of posts. (¶ 0033, 0065: Sureshkumar discloses the system takes reviews of a product as a stream of data and parses the reviews into attributes such as the title.)
With respect to claims 7, 8, and 16, t the combination of Sureshkumar, Alexandrov, and Garrow discloses the method, the computing device, and the machine-readable medium,
wherein the lexico-syntactic pattern matching includes applying one or syntactic rules to titles of the plurality of posts. (¶ 0081: Sureshkumar discloses assigns tags identifying mentions or occurrences of terms.), wherein at least one of the one or more syntactic rules is derived from historical comparative posts. (¶ 0081 - Sureshkumar)
With respect to claims 9 and 17, the combination of Sureshkumar, Alexandrov, and Garrow discloses the method and the computing device,
wherein the filtering includes filtering content of the plurality of posts using keywords indicative of comparison and names of products which are of interest for the product analysis (¶ 0036-0037, 0039, 0086: Sureshkumar discloses a set of reviews for a specific cell phone and some reviews may not address the cell phone itself…not all review/opinions in an opinion set may be relevant to a particular analysis.) and lexico-syntactic pattern matching. (¶ 0032: Sureshkumar discloses the system uses natural language processing to extract insights from reviews across various features or aspects of a product.)
With respect to claim 10, the combination of Sureshkumar, Alexandrov, and Garrow the method of claim 1,
wherein the ML model is trained using instances of historical posts labeled as either comparative or noncomparative. (col. 6:20-63: Alexandrov discloses contextual relevance neural network 205 which represents the ML model may use long short term memory networks “LSTM” or convolutional neural networks “CNN” to determine relevance of features, functionality, qualities, attributes, and/or other descriptive words to different classes of products. The contextual neural network may be trained based on descriptive characteristics from products that compete with or are similar to the particular product. The contextual neural network may detect patterns and may determine that the first feature is relevant for the particular class of products, whereas the second feature is not relevant feature for the particular class of products.)
With respect to claim 11, the combination of Sureshkumar, Alexandrov, and Garrow discloses the method of claim 1, further comprising:
retrieving, by the computing device from the data repository, the information extracted from the comparative posts (cols. 5-6:65-20: Alexandrov discloses retrieve the reviews and descriptive characteristics for a particular product in response to a user request for an assessment of the product. Database 203 stores downloaded or retrieved reviews and descriptive characteristics.); and
generating, by the computing device, a product analysis report using the information retrieved from the data repository. (col. 4:55-63, col. 13:11-33: Alexandrov discloses quality assessment system 100 may generate and provide comparative assessments of the particular product to competing products or competitors.)
Response to Arguments
Applicant's arguments filed 25 November 2025 have been fully considered but they are not persuasive.
With Respect to Rejections Under 35 USC 101
Applicant argues “Applicant respectfully submits that, in view of claim 1 and the USPTO Director's August 4, 2025 guidance on subject-matter eligibility, the § 101 rejection should be withdrawn. As amended, claim 1 now expressly recites "continuously training, by the computing device, a machine learning (ML) model implemented as a multi-layer neural network having a plurality of trainable parameters" using a modeling dataset of training samples derived from the corpus of historical posts, and further recites generating feature vectors and inputting them to that neural- network text-classification model to classify posts as comparative or non-comparative. These limitations define concrete AI/ML processing that cannot practically be performed in the human mind and therefore should not be characterized as a "mental process" under § 101 (and as emphasized under the Director's memo). Nor do the claims recite any specific mathematical formula or algorithm. Further, even if the claims are interpreted to involve mathematical operations in the course of training and applying the neural-network model, the Director's guidance makes clear such involvement does not, by itself, place the claim in the "mathematical concepts" grouping.” The Examiner respectfully disagrees.
Contrary to the remarks, the claims remain ineligible under Step 2A Prong One of the two-part analysis. In the instant case, the Applicant’s reply does not challenge the abstract ideas identified in the previous rejection. Here, the remarks point to the Director’s Memo issued August 2025 and features of ML processing from the claim that cannot be performed mentally, however, simply reciting continuously training a machine learning model, i.e., a particular technological module or piece of equipment, in a claim does not confer eligibility. Also, it is important for Applicant to note that claims can recite a mental process even if they are claimed as being performed on a computer. The focus of the claims is for collecting postings, analyzing/classifying information from postings, and displaying certain results, i.e., product analysis report. Thus, the claim recites subject matter that falls within the certain methods of organizing human activity and mental processes groupings. As for, the continuously training (updating) step is merely being used to facilitate the tasks of the abstract idea, which provides nothing more than a results oriented solution that lacks detail of the mechanism for accomplishing the result and is equivalent to the words “apply it”, per MPEP 2106.05(f). At best, the claim merely implements a ML model, i.e., multi-layer neural network, to aid in performing the product analysis. For these reasons, the rejections under 101 are being maintained.
Applicant further argues “Moreover, when viewed as a whole, claim 1 recites a specific, integrated technical solution for automated text classification in noisy, large-scale user-generated content streams: continuously scanning multiple online sources to build a corpus of historical posts, continuously training the multi-layer neural-network model on labeled training samples from that corpus, generating feature vectors for posts responsive to a product analysis request, classifying those posts with the trained model, and updating the corpus and generating a product-analysis report based on the extracted comparative information. This is not a mere instruction to "apply" an abstract idea on a generic computer, but a particular AI/MIL-based processing pipeline that improves the way computer systems automatically identify and exploit comparative information from unstructured text. The eligibility of such AI/MVL text-classification claims is, under the Director's guidance, should be resolved in favor of eligibility.
Claims 12 and 18 recite similar features. Accordingly, Applicant respectfully requests withdrawal of the § 101 rejection.” The Examiner respectfully disagrees.
Contrary to the remarks, the claims remain ineligible under Step 2A Prong Two of the two-part analysis. The specificity of the presently recited techniques recited in the claims does not automatically make the claimed invention eligible. In the instant case, the Applicant purports in a conclusory manner that the claims provide a technical solution without pointing to any support from their disclosure identifying technical improvements or technological problem to be solved. The Applicant’s own Specification discusses the use of AI/ML processing at a high-level of generality and various types machine learning models that may be implemented. Thus, the combination of limitations do not recite any technical improvements to computer systems or machine learning technology as disputed by Applicant. See Elec. Power Grp, 830 F.3d at 1356 (using result-focused, functional claim language is a frequent feature of ineligible claims, especially those that claim used of generic computer components and network technology to carry out business transactions). For these reasons, the rejections under 101 are being maintained.
With Respect to Rejections Under 35 USC 103
Applicant’s arguments with respect to claim(s) 1-20 have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
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 EHRIN PRATT whose telephone number is (571)270-3184. The examiner can normally be reached 8-5 EST Monday-Friday.
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, Lynda Jasmin can be reached at 571-272-6782. 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.
/EHRIN L PRATT/Examiner, Art Unit 3629
/LYNDA JASMIN/Supervisory Patent Examiner, Art Unit 3629