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 .
This communication is responsive to the application filed on 10/24/2024.
Status of claims:
Claims 1-46 were canceled.
Claims 46-67 are pending for examination.
Information Disclosure Statement
The information disclosure statement (IDS) filed on 02/26/2025 complies with the provisions of M.P.E.P 609. The information referred to therein has been considered as to the merits.
Abstract
The Abstract filed on 10/24/2024 has been considered as to the merits.
Drawings
The Drawings filed on 10/24/2024 have been considered as to the merits.
Claim Objections
Claims 57 and 62-66 are objected because of the following informalities:
- Claim 57 recites “prompting the LLM Al to perform extraction;” should be rephrased as “prompting the LLM Al to perform extraction; or”.
- Claim 62 recites “identifying at least one ML prediction error from the predictions;” should be rephrased as “identifying at least one ML prediction error from the prediction; and”
- Claim 63 recites “applying… a web page context, and” should be rephrased as “applying … a web page context[[,]]; and”
- Claim 64 recites “generating a HTML … the Rols; and passing each HTML …”
should be rephrased as “generating a HTML … the Rols; passing each HTML …”
- Claim 65 recites “processing the 2D tensor … can be chosen; and generating …”
should be rephrased as “processing the 2D tensor … can be chosen; generating …”
Corrections are suggested.
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 61 and 65 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 applicant regards as the invention.
Claims 61 and 65 are rejected because the claims recite the term "can” is not an absolute indicated process. There is no indication that limitation(s) following the term “can” is necessarily a required part of the claimed invention. Appropriate correction is required.
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 54 and 64 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract without significantly more.
With respect to subject matter eligibility under 35 USC 101, it is determined whether the claims are directed to one of the four statutory categories of invention, i.e., process, machine, manufacture, or composition of matter.
Claim 54 recites the limitation “determining if a feed item in the extracted targeted data meets the probability threshold” as drafted, is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind, but for the recitation of generic computer components. That is, other than reciting “when executed by the processor and a navigation depth limit middleware”, nothing in the claim element precludes the steps from practically being performed in a human mind. For example, the “determining” step, in the context of the claim encompasses one can mentally, or manually with the aid of pen and paper determines whether a feed item meets the probability threshold.
- “extracting or generating the feed item from the extracted targeted data”; “extracting a web page and extracting target data from the web page with at least one of the plurality of spiders”, is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind, but for the recitation of generic computer components. That is, other than reciting “instructions that”, nothing in the claim element precludes the steps from practically being performed in a human mind. For example, the “extracting” step, in the context of the claim encompasses one can mentally, or manually with the aid of pen and paper extract target data from the web page.
If a claim limitation, under its broadest reasonable interpretation, covers mental processes but for the recitation of generic computer components, then it falls within the "Mental Processes" grouping of abstract ideas (concepts performed in the human mind including an observation, evaluation, judgement, and opinion). Accordingly, the claim recites an abstract idea.
Claim 64 recites generating a plurality of Regions of Interest (Rol) from the 2D feature maps is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind, but for the recitation of generic computer components. That is nothing in the claim element precludes the steps from practically being performed in a human mind. For example, the “generating” step, in the context of the claim encompasses one can mentally, or manually with the aid of pen and paper generate a plurality of Regions of Interest.
- generating a HTML feature vector for each of the Rols as drafted is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind, but for the recitation of generic computer components. That is nothing in the claim element precludes the steps from practically being performed in a human mind. For example, the “generating” step, in the context of the claim encompasses one can mentally, or manually with the aid of pen and paper generate a HTML feature vector for each of the Rols.
- generating a CSS class and ID feature vector for each Rol by computing an average of the CSS classes and IDs is a process that, under its broadest reasonable interpretation, covers a mathematically calculation, but for the recitation of generic computer components. That is nothing in the claim element precludes the steps from practically being also performed in a human mind. For example, the “generating” step, in the context of the claim encompasses one can mathematically generate a CSS class and ID feature vector for each Rol.
- passing each HTML Rol feature vector to the classification layer” and “passing the CSS class and ID Rol feature vectors to the classification layer is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind, but for the recitation of generic computer components. That is nothing in the claim element precludes the steps from practically being performed in a human mind.
If a claim limitation, under its broadest reasonable interpretation, covers mental processes but for the recitation of generic computer components, then it falls within the "Mental Processes" grouping of abstract ideas (concepts performed in the human mind including an observation, evaluation, judgement, and opinion). Accordingly, the claim recites an abstract idea.
The claim recites additional the limitations:
- obtaining a raw HTML snapshot comprising HTML tags and CSS attributes;
- processing the raw HTML snapshot to obtain a plurality of 2D feature maps using a neural network
- processing the classification layer with a deep neural network (DNN);
- processing the DNN output with the BiLSTM; and
- outputting final classification of N classes including a probability score that a part of a web page belongs to a particular class
The above additional the limitations represent an extra-solution activity because it is a mere nominal or tangential addition to the claim, a mere generic transmission and presentation of collected and analyzed data. (See MPEP 2106.05 (g)).
identified as insignificant extra-solution activity above when re-evaluated this element is well-understood, routine, and conventional as evidenced by the court cases in MPEP 2106.05(d)(II), "i. Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information); … OIP Techs., Inc., v. Amazon.com, Inc., 788 F.3d 1359, 1363, 115 USPQ2d 1090, 1093 (Fed. Cir. 2015) (sending messages over a network); buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014) (computer receives and sends information over a network);" and thus remains insignificant extra-solution activity that does not provide significantly more.
Looking at the claim as a whole does not change this conclusion and the claim is ineligible.
Claim Rejections - 35 USC § 103
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.
Claims 47-60 and 62 are rejected under 35 U.S.C. 103 as being unpatentable over of Kuksta et al., (US 20230018387), hereinafter “Kuksta”, in view of Cooper et al., (US 20040210589), hereinafter “Cooper”, and further in view of Korobov et al., (US 20200210511), hereinafter “Korobov”.
As per claim 47, Kuksta discloses the a system for a computer comprising an input and a memory including non-transitory program memory for storing at least instructions and a processor that is operative to execute instructions for scraping and processing a web page (par. [0028], computer system that comprises one or more processors and a computer-readable storage medium encoded with instructions executable by at least one of the processors and operatively coupled to at least one of the processors for scraping and processing a web page), comprising:
an Application Programming Interface (API) including a smart crawling selector and further comprising (par. [0070], API constructs a datapoint for each data collection response item);
a Machine Learning (ML) module configured for extraction, including an ML rendering extraction module, an ML HTML extraction module for extracting raw HTML (par. [0035], extracting specific classification attributes from the URL, textual content, HTML elements and meta elements to determine the category of the target web page based on the Machine Learning predictive analytics algorithm); and
- an interface module configured to allow a user to select a fix for at least one ML prediction error (par. [0079] and [0121], the input for training of the webpage classifier model contains data that promotes correct prediction behaviour).
However, Kuksta fails to disclose “custom spiders and template spiders configured to implement a crawling strategy”.
On the other hand, Cooper discloses template spiders configured to implement a crawling strategy (par. [0023], [0024], template is coupled to the spider engine that provides useful information to the spider engine related to a search).
Therefore, it would have been obvious to one having ordinary skill in the art before the effective date of the claimed invention to have modified the system of Kuksta to have a template spiders configured to implement a crawling strategy, in order to prevent a spider engine from overloading a web site with web page requests.
However, neither Kuksta nor Cooper discloses a custom spider.
Meanwhile, Korobov discloses a custom spider (par. [0002], creating custom crawlers (“spiders”) for each website being crawled using manually specified rules).
Therefore, it would have been obvious to one having ordinary skill in the art before the effective date of the claimed invention to have modified the system of Kuksta to have a custom spider, in order to efficiently decrease extraction time.
As per claim 48, the combination of Kuksta, Cooper and Korobov discloses the invention as claimed. In addition, Cooper discloses the plurality of spiders including template spiders (par. [0023]-[0024], template is coupled to the spider engine that provides useful information to the spider engine related to a search);
Kuksta discloses extracting a web page and extracting target data from the web page with at least one of the plurality of spiders (par. [0035], extracting specific classification attributes from the URL, textual content, HTML elements and meta elements to determine the category of the target web page based on the Machine Learning predictive analytics algorithm); and
Korobov discloses custom spiders (par. [0002], creating custom crawlers (“spiders”) for each website being crawled using manually specified rules).
As per claim 49, the combination of Kuksta, Cooper and Korobov discloses the invention as claimed. In addition, Kuksta discloses generating the custom spider from a template spider; and extracting the target data from the web page with the custom spider (par. [0002], creating custom crawlers (“spiders”) for each website being crawled using manually specified rules).
As per claim 50, the combination of Kuksta, Cooper and Korobov discloses the invention as claimed. In addition, Kuksta discloses wherein the template spider includes an article template and/or an e-commerce template (par. [0083], e-commerce product page, an e-commerce search page, or hotel listing page, etc.).
As per claim 51, the combination of Kuksta, Cooper and Korobov discloses the invention as claimed. In addition, Kuksta discloses wherein the API includes a strategy selection interface for selecting a full strategy, an incremental strategy, or a navigation strategy (par. [0068]-[0070], API sends the response data to metadata parser, which extracts the necessary metadata information from the response data, wherein after which, metadata parser returns the extracted information, i.e., the metadata information, to API).
As per claim 52, the combination of Kuksta, Cooper and Korobov discloses the invention as claimed. In addition, Korobov discloses at least one a navigation depth limit middleware configured to enforce a limit on the number of hops the spider scrapes (par. [0003], obtain structured data, including text, image and other kinds of data, from web pages, and process them for, among other advantages, more efficient web crawling, website analysis, creating knowledge databases and graphs, and providing more useful web page representation for other automated web page processing components, wherein scraping module includes a program or algorithm for, when executed by the processor, performing some or all of these actions: rendering a web page in a web browser, downloading of the related resources like images or CSS files, executing JavaScript and other instructions, obtaining screenshots of web pages and its parts, measuring on-screen coordinates, colors, visibility, and other properties of HTML elements and other parts of web pages, obtaining final HTML representation of the web pages, and automatically executing commands to control the browser);
- a seed domain restriction middleware configured to restrict extraction requests to an original seed URL domain;
- a feed link filtering middleware configured to include filter requests to feed links;
- a seed request limitation middleware configured to limit the number of requests from a seed request;
- a duplicate item extraction prevention middleware configured to skip items extracted from a prior crawl;
- an allow offsite middleware configured to ensure that category links that do not belong to a same domain are crawled;
- a crawling logs middleware; or
- any combination thereof.
As per claim 53, the combination of Kuksta, Cooper and Korobov discloses the invention as claimed. In addition, Kuksta discloses extract all links from a from a web page (par. [0035], sending the received HTML content and the URL of the target web page to a classifier unit, parsing the HTML content to extract the entirety of the textual content);
- split the links into a plurality of groups (par. [0035], extracting specific classification attributes from the URL, textual content, HTML elements and meta elements to determine the category of the target web page based on the Machine Learning predictive analytics algorithm);
- filter target feed links into one of the groups (par. [0052], evaluation of said data comprises pre-processing the data contained therein, extracting relevant datapoints aligned with the original data collection request, classifying and labelling the resultant content, and ultimately returning the classified and labeled data to the Scraping Agent, providing the probability percentile for the classification identified), and
- extract only the targeted data from the target feed links (par. [0035], determine the category of web pages, receiving the HTML content from a target web server, sending the received HTML content and the URL of the target web page to a classifier unit, parsing the HTML content to extract the entirety of the textual content, HTML element names).
As per claim 54, the combination of Kuksta, Cooper and Korobov discloses the invention as claimed. In addition, Kuksta discloses wherein the feed link filtering comprises a feed item probability threshold, and the feed link filtering middleware is configured to, when executed by the processor, enable actions comprising:
- determining if a feed item in the extracted targeted data meets the probability threshold, and if so, extracting or generating the feed item from the extracted targeted data (par. [0035], determine the category of web pages and receiving the HTML content from a target web server, sending the received HTML content and the URL of the target web page to a classifier unit, parsing the HTML content to extract the entirety of the textual content, HTML element names).
As per claim 55, the combination of Kuksta, Cooper and Korobov discloses the invention as claimed. In addition, Kuksta discloses the ML module (par. [0035], extracting specific classification attributes from the URL, textual content, HTML elements and meta elements to determine the category of the target web page based on the Machine Learning predictive analytics algorithm).
As per claim 56, the combination of Kuksta, Cooper and Korobov discloses the invention as claimed. In addition, Kuksta discloses wherein the ML module is configured to filter or discard non-targeted web content (par. [0035], extracting specific classification attributes from the URL, textual content, HTML elements and meta elements to determine the category of the target web page based on the Machine Learning predictive analytics algorithm).
As per claim 57, the combination of Kuksta, Cooper and Korobov discloses the invention as claimed. In addition, Kuksta discloses wherein the system comprises the non- transitory program memory for storing the instructions that, when executed by the processor enable actions comprising at least one of:
- accepting inputs for custom attribute data from a user;
- accessing a web page from a website;
- extracting HTML text from the web page (par. [0035], extracting specific classification attributes from the URL, textual content, HTML elements and meta elements to determine the category of the target web page based on the Machine Learning predictive analytics algorithm);
- processing the HTML text from the web page with a lightweight machine learning (ML) artificial intelligence (Al) model;
- inputting data from the ML AI model and the custom attribute data from the user to a Large Language Model (LLM) Al;
- prompting the LLM Al to perform extraction;
- obtaining LLM AI extracted attributes from the LLM Al.
As per claim 58, the combination of Kuksta, Cooper and Korobov discloses the invention as claimed. In addition, Cooper discloses wherein the non-transitory program memory for storing at least instructions and the processor that is operative to execute instructions that enable actions further comprises: a template or code configured to allow the user to describe and input custom attributes (par. [0023]-[0024], template can be written in a description language, wherein the template determines for the spider engine: what data to search for, where the data resides (location information), the nature of the data, and what to do with the data. For instance, the location information may include the location of data within a particular web page, and the location of data in a particular web site, or the like).
As per claim 59, the combination of Kuksta, Cooper and Korobov discloses the invention as claimed. In addition, Kuksta discloses wherein the scraping module is configured to execute data type-specific extraction using the ML models, which also identify the main text of the extracted item (par. [0035], extracting specific classification attributes from the URL, textual content, HTML elements and meta elements to determine the category of the target web page based on the Machine Learning predictive analytics algorithm).
As per claim 60, the combination of Kuksta, Cooper and Korobov discloses the invention as claimed. In addition, Kuksta discloses wherein the non-transitory program memory for storing at least instructions and the processor that is operative to execute instructions that enable actions further comprise:
- extracting the HTML text from the web page (par. [0035], parsing the HTML content to extract the entirety of the textual content); and
- using the ML AI model to identify and exclude portions of the extracted webpage
from the extraction of the HTML text (par. [0035], extracting specific classification attributes from the URL, textual content, HTML elements and meta elements to determine the category of the target web page based on the Machine Learning predictive analytics algorithm).
As per claim 62, the combination of Kuksta, Cooper and Korobov discloses the invention as claimed. In addition, Korobov discloses comprising non-transitory program memory for storing at least instructions and a processor that is operative to execute instructions that enable actions comprising:
- accessing a web page from a website (par. [0028], obtaining an image of a web page; processing the web page with an object detection neural net to obtain feature vectors for one or more image regions of interest on the image of the web page);
extracting data from the web page (par. [0040], extract HTML text from the web page);
- processing the data from the web page with a machine learning (ML) model to generate predictions for web page elements, wherein the ML model assigns a probability to the web page elements that are higher than an ML model threshold (par. [0016]-[0017] and [0038], processing the HTML text from the web page with the neural net to obtain feature vectors for one or more HTML regions of interest using a machine learning components and outputting a probability score that a part of a web page belongs to a defined class).
Kuksta also discloses identifying at least one ML prediction error from the predictions (par. [0035], determine the category of the target web page based on the Machine Learning predictive analytics algorithm); and
- providing the interface to an operator including a fix for the at least one ML prediction error (par. [0071] and [0083], return a prediction of the category the specific set of attributes corresponds to, together with the probability score for each category).
Claim 61 is rejected under 35 U.S.C. 103 as being unpatentable over of Kuksta et al., in view of Cooper and Korobov, and further in view of Wang et al., (US 20230133392), hereinafter “Wang”.
As per claim 61, the combination of Kuksta, Cooper and Korobov discloses the invention as claimed, except for tokenizing a source text as N-question tokens and a context concatenating the N question tokens and the context, wherein N is a hyper-parameter indicating the number of questions the model can answer;
However, Wang discloses tokenizing a source text as N-question tokens and a context concatenating the N question tokens and the context, wherein N is a hyper-parameter indicating the number of questions the model can answer (par. [0011], select a predetermined number of highest ranked question-answer pairs, and return the predetermined number of highest ranked question-answer pairs to a user).
Therefore, it would have been obvious to one having ordinary skill in the art before the effective date of the claimed invention to have modified the system of Kuksta to indicate the number of questions the model can answer in order to provide real-time insight and predictive analysis fast enough to meet user service level expectations, and providing visibility into performance data and dependencies across all environments.
Claim 63-67 are rejected under 35 U.S.C. 103 as being unpatentable over of Kuksta, in view of Cooper and Korobov, and further in view of Bahrami et al., (US 20210064453), hereinafter “Bahrami”.
As per claim 63, the combination of Kuksta, Cooper and Korobov discloses the invention as claimed. In addition, Korobov discloses comprising memory for storing at least instructions and a processor that is operative to execute instructions that enable actions for a method comprising:
- auto extracting HTML text from web page based on an HTTP request (par. [0018] and [0028], extracting HTML text from the web page);
- operating a Machine Learning model on a browser request to use an image modality to capture the page context (par. [0015], [0094], obtain structured data, including text, image and other kinds of data, from web pages, and process them for, among other advantages, more efficient web crawling, website analysis, creating knowledge databases and graphs and Machine Learning-based image, text and HTML structure processing on web pages).
The combination of Kuksta, Cooper and Korobov to discloses the claimed “applying a stacked bidirectional long short term memory network (BiLSTM) in a head configured to operate on a sequence of DOM elements to capture a web page context”.
Meanwhile Bahrami discloses applying a stacked bidirectional long short term memory network (BiLSTM) in a head configured to operate on a sequence of DOM elements to capture a web page context (par. [0034]-[0036], train a machine learning model to create a method of auto-extracting the data that make up the API feature list. In some embodiments, the machine learning model may receive the extracted content (whether it has been extracted automatically or through manual interaction). The content may be fed into a Bi-direction long short-term memory (Bi-LSTM) model, for example, which may encode the extracted content. The model may operate to understand the variety of HTML tags and the parse tree of the DOM which corresponds to each API object. For example, the machine learning model may learn that the HR tags in “<HR>Verb Endpoint</HR> corresponds to “Paths->Endpoint->verb in the OAS file, wherein the API feature decoder may then feed into an API object encoder (e.g., BiLSTM) model, another dense layer, and an object decoder, resulting in a trained model).
Therefore, it would have been obvious to one having ordinary skill in the art before the effective date of the claimed invention to have modified the system of Kuksta to apply a stacked bidirectional long short term memory network (BiLSTM) in a head configured to operate on a sequence of DOM elements to capture a web page context, in order to increase efficiency and communication time.
As per claim 64, the combination of Kuksta, Cooper, Korobov and Bahrami discloses the invention as claimed. In addition, Korobov discloses wherein the method further comprises:
- obtaining a raw HTML snapshot comprising HTML tags and CSS attributes (par. [0016] and [0073]-[0078]);
- processing the raw HTML snapshot to obtain a plurality of 2D feature maps using a neural network (par. [0073]-[0078]);
- generating a plurality of Regions of Interest (Rol) from the 2D feature maps (par. [0073]-[0078], generating candidate regions of interest (RoIs), wherein each RoI on a feature map is then resized to a fixed size, keeping the same depth of the feature map, using a process called RoI pooling);
- generating a HTML feature vector for each of the Rols (par. [0074]-[0078], use of HTML element and other part of web page measurements in a process of generating candidate regions of interest (RoIs)); and
- passing each HTML Rol feature vector to the classification layer (par. [0073]-[0078]);
- generating a CSS class and ID feature vector for each Rol by computing an average of the CSS classes and IDs (par. [0073]-[0078]);
- passing the CSS class and ID Rol feature vectors to the classification layer (par. [0016] and [0076]-[0078);
- processing the classification layer with a deep neural network (DNN) (par. [0016] and [0076]-[0078);
-- processing the DNN output (par. [0025], output a probability score that a part of a web page belongs to a defined class);
- outputting final classification of N classes including a probability score that a part of a web page belongs to a particular class (par. [0025], output a probability score that a part of a web page belongs to a defined class).
- processing the DNN output (par. [0025], output a probability score that a part of a web page belongs to a defined class).
Bahrami discloses processing the DNN output with the BiLSTM (par. [0034]-[0036], train a machine learning model to create a method of auto-extracting the data that make up the API feature list. In some embodiments, the machine learning model may receive the extracted content (whether it has been extracted automatically or through manual interaction). The content may be fed into a Bi-direction long short-term memory (Bi-LSTM) model, for example, which may encode the extracted content. The model may operate to understand the variety of HTML tags and the parse tree of the DOM which corresponds to each API object. For example, the machine learning model may learn that the HR tags in “<HR>Verb Endpoint</HR> corresponds to “Paths->Endpoint->verb in the OAS file, wherein the API feature decoder may then feed into an API object encoder (e.g., BiLSTM) model, another dense layer, and an object decoder, resulting in a trained model).
As per claim 65, the combination of Kuksta, Cooper, Korobov and Bahrami discloses the invention as claimed. In addition, Korobov discloses removing information other than the HTML tags and CSS attributes from the HTML snapshot (par. [0073]-[0078]);
- encoding the HTML snapshot to a byte string (par. [0073]-[0078]);
- converting the byte string to a 2D tensor using a character embedding layer (par. [0073]-[0078]);
- processing the 2D tensor to obtain a plurality of 2D feature maps using a neural network, wherein the lengths of the 2D feature maps are the same lengths of the input byte string, and wherein the depth parameter can be chosen (par. [0073]-[0078]); and
- generating the plurality of Regions of Interest (Rol) from the 2D feature maps, wherein each Rol corresponds to a respective Rol on the byte string and respective 2D feature maps (par. [0073]-[0078]);
- resizing each Rol to a fixed length, wherein the resizing keeps a depth of the feature map and uses ROI pooling (par. [0073]-[0078]); and
- passing the resized Rols to a classification layer (par. [0073]-[0078]).
As per claim 66, the combination of Kuksta, Cooper, Korobov and Bahrami discloses the invention as claimed. In addition, Korobov discloses wherein the method further comprises: for each possible HTML tag (par. [0073]-[0078]);
- computing an embedding during training, the embedding being a vector of a fixed size; and for each of the Rols, extracting an HTML tag (par. [0073]-[0078]);
looking up the HTML tag in an embeddings table (par. [0073]-[0078]);
generating the HTML feature vector for the Rol (par. [0073]-[0078]); and
passing each HTML Rol feature vector to the classification layer (par. [0073]-[0078]).
As per claim 67, the combination of Kuksta, Cooper, Korobov and Bahrami discloses the invention as claimed. In addition, Korobov discloses computing an embedding for popular CSS classes or ids during training, the embedding being a vector of fixed size (par. [0073]-[0078]);
- for each of the Rols, extracting all CSS classes and IDs (par. [0073]-[0078]);
- looking up the CSS classes and IDs in a table (par. [0073]-[0078]);
- generating the CSS class and ID feature vector for each of the Rols by computing an average of the CSS classes and IDs (par. [0073]-[0078]); and
- passing the CSS class and ID Rol feature vector the classification layer (par. [0073]-[0078]).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure (see PTO-892).
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Loan T. Nguyen whose telephone number is (571) 270-3103. The examiner can normally be reached on Monday from 10:00 am - 6:00 pm, Thursday-Friday from 10:00 am - 2:00 pm. If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Aleksandr Kerzhner can be reached on (571) 270-1760. The fax phone number for the organization where this application or proceeding is assigned is 571-270-4103. Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/LOAN T NGUYEN/Examiner, Art Unit 2165
10/28/2025