DETAILED ACTION
This communication is in response to the Applicant’s Arguments filed on 10/16/2025. Claims 1-20 are pending and have been examined. Hence, this action has been made FINAL.
Any objections/rejections not mentioned in this Office Action has been withdrawn by the Examiner.
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 .
Applicant’s Arguments
The Applicant responds to the Specification Objections made by citing to paragraph [0021], specifically for the “transformer-based language model” and “small or tiny”. The Applicant’s arguments are persuasive and the objection has been withdrawn by the examiner.
With respect to the 35 USC 101 abstract rejections, the Applicant asserts the claims are not directed towards an abstract idea and are directed towards eligible subject matter.
More specifically, the Applicant asserts on the bottom of page 3 into page 4:
“The present claims can easily be deemed as directed to eligible subject matter because they are not directed to any of the ineligible concepts ("judicial exceptions"). The claims are directed to processing multiple URLs to detect malicious behavior collectively exhibited by API calls. This is not even slightly similar to any of the subject matter the Court identified as directed to an ineligible concept. For the abstract idea ineligible category, the Court identifies a fundamental truth, motive, an original cause, an idea itself; and then the Court provides case based examples including a mathematical formula, an algorithm itself, the basic concept of hedging, and a fundamental economic practice. A machine-learning ensemble that classifies a file as malware or benign cannot reasonably be considered a fundamental truth, idea itself, a motive, or an original cause.“
The Examiner disagrees with this assertion. The Applicant notes that “the claims directed to processing multiple URLs to detect malicious behavior collectively exhibited by API calls”. The Examiner notes that the claims only use API calls in the following manner per the claims “wherein each of the first plurality of URLs corresponds to one or more application programming interfaces (APIs) as recited in claim 1 and similarly in claim 3. Claims 10 and 17 uses language of determining URLs indicated in traffic logs “that corresponds to application programming interface calls (API)”. The examiner notes that although the claims mention API calls, the claims broadly indicate that they “correspond to”. Also, the Applicant does not provide any reasoning as to how the following cannot be done mentally/via pen paper by a human and appears to make a general allegation of subject matter eligibility. It is reasonable for a person to process URLs determined from network traffic logs as identified by a network router. These URLs can be provided from printouts that the user already has from network traffic and therefore further processes based on these URLs. The fact that an API call is mentioned in the claims without specifics would not take the claims away from an abstract idea. Furthermore, the Applicant appears to note that “a machine-learning ensemble that classifies a file as malware or benign cannot reasonably be considered a fundamental truth, idea itself, a motive, or an original cause”. This is not currently claimed in any of the present claims.
More specifically, the Applicant asserts on page 4:
The Office concludes that the independent claims are directed to an abstract idea because the first word of each claim element are mental activities. This is not a proper "directed to" analysis. Instead of an accurate characterization for the "directed to" analysis, the Office has done specifically what has been warned against by the Supreme Court and the Federal Circuit. "[W]e have reiterated the Supreme Court's caution against 'overgeneralizing claims' in the § 101 analysis, explaining that characterizing the claims at 'a high level of abstraction' that is 'untethered from the language of the claims all but ensures that the exceptions to § 101 swallow the rule.' Enfish, 822 F.3d at 1337; see Solutran, 931 F.3d at 1167-68; McRO, Inc. v. Bandai Namco Games America, Inc., 837 F.3d 1299, 1313 (Fed. Cir. 2016) (explaining that courts 'must be careful to avoid oversimplifying the claims' by looking at them generally and failing to account for the specific requirements of the claims')."2"The Step 1 'directed to' analysis called for by our cases depends on an accurate characterization of what the claims require and of what the patent asserts to be the claimed advance. The accuracy of those characterizations is crucial to the sound conduct of the inquiries into the problem being addressed and whether the line of specificity of solution has been crossed."3 Concluding ineligibility based on nothing more than the first word of each claim element is unquestionably overgeneralizing and characterization that is untethered from the language of the claims. For the "directed to" inquiry, the Office should look at the "'focus of the claimed advance over the prior art' to determine if the claim's 'character as a whole' is directed to excluded subject matter."4"In conducting that inquiry, we 'must focus on the language of the Asserted Claims themselves,' Synopsys, Inc. v. Mentor Graphics Corp., 839 F.3d 1138, 1149 (Fed. Cir. 2016), 'considered in light of the specification,' Enfsh, LLC v. Microsoft Corp., 822 F.3d 1327, 1335 (Fed. Cir. 2016)."5 The Office has not focused on the language of the claims and has not made an accurate characterization of the claims.
The Examiner respectfully disagrees. The Examiner on page 4 has clearly noted how each of the limitation are directed towards an abstract idea by providing how each limitation can be performed by a human. The Applicant appears to again make general allegations of subject matter eligibility by not clearly indicating how the office “has not focused on the language of the claims and has not made an accurate characterization of the claims”. In other words, what language has the Office not focused on and what was not indicated accurately. Instead, the applicant provides several case laws and does not provide a showing as to what the 35 USC 101 abstract rejection made by the Examiner fails to address or why the claims are not directed towards mental activity.
More specifically, the Applicant asserts on page 5:
Appellant's Specification clearly explains the problem of determining malicious behavior exhibited by multiple API calls in URLs. And the claims clearly recite the processing of URLs, intent classification, and summarization of intents to address the problem. Claim 1 specifies that sentences are generated from URLs and that intents of the URLs based on the sentences are determined with a first language model. Claim 1 specifies a particular ordering of the intents and specifies summarizing the ordered intents with a second language model. Claims 10 and 17 each specify instructions that determine URLs from network traffic logs that corresponds to API calls and extracting words from the URLs in network traffic logs that correspond to API calls. Claims 10 and 17 each specify instructions determine intent of each URL based on the extracted words with a first language model and instructions to summarize the intents based on metadata of the intents, order of the intents, and the intents. It is grossly inaccurate to characterize the claims as directed to mental activities. There should be no reason to turn to Alice/Mayo Step 2 since it is abundantly clear that the claims are directed to patent eligible subject matter.
The Examiner respectfully disagrees. The Applicant asserts that the Specification clearly explains the problem “of determining malicious behavior by multiple API calls in URLs” and the claims recite “processing of URLs , intent classification, and summarization of intents to address the problem”. The Examiner first would like to make clear that there is nothing in the claim language that clearly establishes or notes that the sequence of steps are related or meant for determining malicious behavior exhibited by multiple calls in URLs. This is not claimed. The Examiner directs the Applicant to MPEP 2106.05(a): “After the examiner has consulted the specification and determined that the disclosed invention improves technology, the claim must be evaluated to ensure the claim itself reflects the disclosed improvement in technology. Intellectual Ventures I LLC v. Symantec Corp., 838 F.3d 1307, 1316, 120 USPQ2d 1353, 1359 (Fed. Cir. 2016) (patent owner argued that the claimed email filtering system improved technology by shrinking the protection gap and mooting the volume problem, but the court disagreed because the claims themselves did not have any limitations that addressed these issues). That is, the claim must include the components or steps of the invention that provide the improvement described in the specification.”
Next, the Applicant summarizes claims 1, 10, and 17 and makes another general allegation of subject matter eligibility by noting “It is grossly inaccurate to characterize the claims as directed to mental activities. There should be no reason to turn to Alice/Mayo Step 2 since it is abundantly clear that the claims are directed to patent eligible subject matter”. However, again it is unclear how it is “grossly inaccurate” to indicate the claims as directed to mental activities. Why is the reasoning provided by the Examiner “grossly inaccurate”. As noted in the non-final rejection mailed starting on page 4, each of the limitations in each of the claims are addressed on how each could be mentally performed or via pen and paper. The applicant has not commented specifically on any of these steps and even if an argument were to be made it would be reasonable such as for claim 1 with URL data on hand to generate sentences, determine intents, ordering the URLs based on timing, and then generating a summary and further making use of two language models for the steps of determining an intent and for generating a summary. Language models have been known for many years even prior to the usage of LLMs which were traditional based on handwritten rules or stored information surrounding specific word ordering/semantics and knowledge of words including n-grams and their associated frequencies. Likewise similar reasoning can be applied to claims 10 and 17.
Hence, the Applicant’s arguments are not persuasive.
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. The independent claims 1, 10, and 17, recites a method, CRM and apparatus thus relating to a statutory category. Claim 1 recites “generating sentences based, at least in part, on a first plurality of uniform resource locators (URLs) of a first time window, wherein each of the first plurality of URLs corresponds to one or more application programming interfaces (APIs); determining, with a first language model, a first plurality of intents of the first plurality of URLs based, at least in part, on the sentences of the URLs; forming a first input according to temporal order of the first plurality of URLs, wherein the first input is formed with the first plurality of intents and metadata associated with the first plurality of URLs; and generating a summary of the first plurality of intents with a second language model based, at least in part, on the first input.” Claims 10 and 17 recite “generating sentences based, at least in part, on a first plurality of uniform resource locators (URLs) of a first time window, wherein each of the first plurality of URLs corresponds to one or more application programming interfaces (APIs); determining, with a first language model, a first plurality of intents of the first plurality of URLs based, at least in part, on the sentences of the URLs; forming a first input according to temporal order of the first plurality of URLs, wherein the first input is formed with the first plurality of intents and metadata associated with the first plurality of URLs; and generating a summary of the first plurality of intents with a second language model based, at least in part, on the first input.”
The limitation of claim 1 of “generating…”, “determining…”, “forming…” and “generating…”, as drafted covers mental activities. More specifically, a human generating sentences from URLs that are provided he/she based on APIs. Then, determining an intent of the sentences followed by creating a first input to be used in determining a summary where the first input is the URLS, intents, and metadata. Then providing the summary based on these pieces of information using a mapping relationship and a model known beforehand. The limitation of claims 10 and 17 of “determine…”, “ preprocess…”, “determine…”, “form…”, and “generate…” as drafted covers mental activities. More specifically, a human receiving a list of URLs visited based on network tragic logs. Then, based on the URLs determining intent from words followed by determining various data such as metadata and order of URLs to then summarize the URLs and when each was accessed using the intents and metadata.
This judicial exception is not integrated into a practical application. In particular, claims 10 and 17 and recites the additional elements of “processor” (claim 10 and 17), non-transitory computer/machine readable medium (claim 10 and 17) as well as (two tower deep neural network) in the independent claims. For example, paragraph [0045] of the as filed specification, there is description of general purpose computers. Such, devices amount to general purpose computing devices. Claim 1 does not recite any additional limitations. Accordingly, these additional elements does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea. The claim is directed to an abstract idea.
The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element of using a computer is noted as a general computer as noted. Mere instructions to apply an exception using a generic computer component cannot provide an inventive concept. Further, the additional limitation in the claims noted above are directed towards insignificant solution activity. The claims are not patent eligible.
With respect to claim 2,, the claim relates to “extracting a second plurality of URLs and corresponding metadata from network traffic data and determining which of the second plurality of URLs corresponds to one or more APIs to obtain the first plurality of URLs.” This reads on a human determining another set of URLs and associated metadata from a printout from network traffic and determining which of the URLs corresponds to an API that was used to obtain the first set of URLs. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claim 3, the claims relate to “wherein determining which of the second plurality of URLs corresponds to an API comprises classifying each URL of the second plurality of URLs as corresponding to an API or not corresponding to an API with a regression model based, at least in part, on features of the URL.” This reads on a human classifying each URL as related to an API or not using known relationships. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claim 4, the claim relates to “wherein generating the sentences comprises, for each of the first plurality of URLs, determining at least one of a subject and a verb based, at least in part, on the URLs.” This relates to a human creating sentences based for each URL based on a determined subject and verb of the URL. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claim 5, the claim relates to “wherein generating the sentences further comprises, for each of the first plurality of URLs, determining at least one of a subject and a verb from metadata associated with the URL.” This relates to a human creating sentences based for each URL based on a determined subject and verb of the URL. No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claim 6 and 13, the claim relates to “wherein generating the sentences comprises, for each of the first plurality of URLs, tokenizing the URL based on camel case detection, tokenizing a blended word with the soft version of the Viterbi algorithm, detecting an abbreviation and expanding the abbreviation, and removing punctuation.” This relates to a human generating sentences by breaking the URL based on camel case, determining blended words and using a soft version, expanding abbreviations and removing punctuations. No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claim 7, the claim relates to “selecting the first plurality of intents from a second plurality of intents based on a common attribute of the first plurality of URLs, wherein the second plurality of intents corresponds to a second plurality of URLs, wherein the first plurality of URLs is a subset of the second plurality of URLs.” This relates to a human selecting a set of intents from another group of intents based on a common feature (domain) where the intents are related to a first set of URLs and a second set of URLs. No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claim 8, the claim relates to “for each of the first plurality of URLs, extracting at least one of an application name, a path parameter, and a query parameter from metadata of the URL.” This relates human for each URL extracting pieces of information such name of the application, path, query and metadata. No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claim 9, the claim relates to “wherein the first language model is smaller than a large language model.” This relates human using a less intensive relationship model than a LLM is. No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claim 11 and 18, the claim relates to “determine URLs in network traffic logs that correspond to API calls comprise instructions to generate feature vectors for URLs indicated in the network traffic logs based on the URLs and metadata associated with the URLs in the traffic logs and classify the URLs with a classifier based on the feature vectors.” This relates to human determining URLs in network traffic logs and generating features for each URL in the traffic logs based on metadata and classify each URL based on the features determined to a specific domain or topic. No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claim 12 and 19, the claim relates to “wherein the program code further has stored thereon instructions to generate a dataset to train the first language model, wherein the instructions to generate the dataset comprise instructions to crawl one or more API specifications to extract words and intent classifications.” This relates to human crawling and locating information from APIs related to words and intent classifications and determining relationships to then be applied to the URL data. No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claim 14 and 20, the claim relates to “wherein the first and the second language models are lightweight transformer-based language models.” This is specifically defining the language model to be lightweight transformer based models. It is interpreted that these models are additional limitations. However, the specification does not provide any information or details regarding how this language model is implemented and what it comprises. Rather paragraph [0015] and [0021] broadly denotes these as a lightweight language model with less than 1billion parameters). Therefore, the Examiner interprets this to mean any transformer LM that makes use of less parameters such as the well-known BERT. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claim 15, the claim relates to “wherein the network traffic logs indicate network traffic occurring within a specified time window at a set of one or more security appliances.” This relates to a human receiving a printout of network traffic for a specific day based on a specific day. The “one or more security appliances” is only being used to determine and gather data. No additional elements are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claim 16, the claim relates to “to select the subset of the URLs based on a common attribute.” This relates to a human selecting only certain URLs related to an attribute. No additional elements are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
These claims further do not remedy the judicial exception being integrated into a practical application and further fail to include additional elements that are sufficient to amount to significantly more than the judicial exception.
Allowable Subject Matter
Claim 1-20 would be allowable if rewritten or amended to overcome the rejection(s) under 35 U.S.C. 101, set forth in this Office action.
The following is a statement of reasons for the indication of allowable subject matter:
The closest prior art of record, with respect to claim 1, Arora (US 2025/0124620) is cited to disclose generating sentences based, at least in part, on a first plurality of uniform resource locators (URLs) of a first time window, wherein each of the first plurality of URLs corresponds to one or more application programming interfaces (APIs) generating a summary of the first plurality of intents with a second language model based (see [0025] conversion of tabular information into one or more language sentences which is then inputted into a natural language model to generate a summary.) Although the reference does not speak of URLs associated with APIs it does teach the concept of generating sentences so that a language model can make use and process the data from the input.
With respect to claim 1: 2nd limitation, Al-Kabra (US 2019/0130036) is cited to disclose determination of user intent from browsing activity such as determining a URL and converting to a host name. (see [0055], [0060]).
However, none of the prior art either alone or in combination thereof teaches or makes obvious the combination of limitations as recited in claim 1 of “determining, with a first language model, a first plurality of intents of the first plurality of URLs based, at least in part, on the sentences of the URLs; forming a first input according to temporal order of the first plurality of URLs, wherein the first input is formed with the first plurality of intents and metadata associated with the first plurality of URLs and generating a summary of the first plurality of intents with a second language model based, at least in part, on the first input”.
The closest prior art of record, with respect to claim 10 and 17, Al-Kabra (cited above) is cited to disclose a non-transitory, machine-readable medium having program code stored thereon, the program code comprising instructions to (see [0020], where CRM and computer and instructions described): determine uniform resource locators (URLs) indicated in network traffic logs that correspond to application programming interface (API) calls (see [0055], [0060], where URL is determined from logs); preprocess the URLs to extract words (see [0051], where terms are identified from the URL); for each of the URLs, determine an intent with [[a first language model]] based on the words of the URLs (see [0050], [0060], where intent is determined from the URL). Chen (CN110912861 A) is also cited to disclose accessing URLS of a network access log, parsing the words within the URL to determine topics related to the user IP.
Reddy (US 2025/0086394) is cited to disclose receipt of API documents and use of LLMs to extract intents, where these intents are aggregated and then submitted to a LLM for further processing. (see Figure 4)
However, none of the cited art either alone or in combination teaches or makes obvious the combination of limitations as recited in claims 10 and 17. More specifically, the limitation of “form a first input for a second language model based, at least in part, on the intents, metadata, and temporal order of at least a subset of the URLs; and generate a summary of the intents with the second language model based on the first input” in conjunction with the other limitations are not taught by the prior art of record.
Another closest prior art Narang (US 2024/0202440) is cited to disclose receiving a URL of a webpage, identifying key points for the webpage in response to a threshold and presenting these keypoints to a user (see Fig 5). However, Narang only uses the URL for retrieving the webpage and the key points are based on the information on the webpage itself not within the URL link.
NPL references by Su, Ming-Yang (“Bert-Based Approaches to Identifying Malicious URLs”) is cited to disclose input of URLs from a dataset and passing the URLs to a LLM (see Figure 2), where figure 4 shows the tokenization process for URLs. Chang (“Research in Malicious URL Detection Technology Based on Bert Model”) provides similar technology as in Su (see sect. III, C, Input and Output). Yang (“Detection of Malicious URL Based on Bert-CNN) is cited to disclose processing of a URL link into segments and passing it through a BERT model to determine if a URL is malicious (see sect. III, section A-B).
Conclusion
THIS ACTION IS MADE FINAL. 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 PARAS D SHAH whose telephone number is (571)270-1650. The examiner can normally be reached Monday-Thursday 7:30AM-2:30PM, 5PM-7PM (EST), Friday 8AM-noon (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, PARAS D SHAH can be reached at 571-270-1650. 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.
/Paras D Shah/ Supervisory Patent Examiner, Art Unit 2653
01/11/2026