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 .
Response to Arguments
Applicant’s arguments, see page 11, filed 5/11/2026, with respect to objections have been fully considered and are persuasive. The objections of claims 7,15 has been withdrawn.
Applicant's arguments filed 5/11/2026 have been fully considered but they are not persuasive.
The applicant contends
Claims 1-8 are rejected under 35 U.S.C. 101 as allegedly directed to an abstract idea in the form of a mental process without significantly more. According to MPEP 2106.04(d)(1), a claim reciting a judicial exception is not directed to the judicial exception if it also recites additional elements demonstrating that the claim as a whole integrates the exception into a practical application. One way to demonstrate such integration is when the claimed invention improves the functioning of a computer or improves another technology or technical field.
In order to make this determination, first the specification should be evaluated to determine if the disclosure provides sufficient details such that one of ordinary skill in the art would recognize the claimed invention as providing an improvement. The specification need not explicitly set forth the improvement, but it must describe the invention such that the improvement would be apparent to one of ordinary skill in the art. According to Applicant's specification artificial intelligence (AI) systems have been developed to perform many different tasks, including generating summaries of text. Such summary generation is referred to as "summarization." A difficulty arises when text to be summarized is greater than the context window of a language model (LM) that has been trained, using one or more machine learning techniques, to summarize text. The context window refers to the maximum length of input to an LM in a single call to the LM. If the size of text to be summarized is greater than the context window (e.g., four kilobytes or four thousand tokens), then the LM must be called multiple times, where each call includes a different portion of the input text. There are numerous challenges for summarizing large sets of data (e.g., a large corpus of documents) using a generative (Gen) Al system, including: (1) how to efficiently process high volume of documents in an effective way with low latency at inference time; (2) how to handle diversity of the input documents in summarization where different documents use different terminology, have conflicting information, and/or pertain to different topics; and (3) how to extract and combine information in an iterative fashion with low latency (at runtime) while at the same time guaranteeing the quality of summarization. Embodiments improve computer-related technology related to generative Al, particularly to generating summaries. For example, embodiments reduce the latency of generating summaries because some portions of text data do not need to be summarized due to their duplicative nature. As another example, embodiments ensure that each topic in text data is reflected in the output.
The examiner disagrees. Although the specification discloses an improvement, the recited claimed language broadly recites language directed towards an abstract idea in the form of mental process. The applicant’s remarks include aspects of the applicant’s invention that is not recited in the claimed language. MPEP 2106 requires the claimed invention to qualify as patent-eligible subject matter. This indicates the claimed invention must include positively recited language integrating the abstract idea into practical application and/or indicating significantly more than the judicial exception. Unfortunately, the recited claimed language is merely directed towards the judicial exception without positively recited language indicating significantly more and/or integrating the abstract idea into practical application.
Second, according to MPEP 2106.04(d)(1), if the specification sets forth an improvement in technology, the claim must be evaluated to ensure that the claim itself reflects the disclosed improvement. That is, the claim includes the components or steps of the invention that provide the improvement described in the specification. The claim itself does not need to explicitly recite the improvement described in the specification. According to Claim 1, multiple portions of text data are identified and an embedding is generated for each portion of the multiple portions. Given the generated embeddings, multiple clusters of embeddings are generated. Each cluster may correspond to a different topic. For each cluster of embeddings, a first language model generates a cluster summary based on portions that correspond to embeddings associated with that cluster. A second language model generates a final summary based on the generated cluster summaries. Thus, according to Claim 1, instead of inputting each text portion into a language model and simply aggregating the respective outputs, embeddings representing the different text portions are first generated and then clustered, which allows for duplicative text to be dealt with at the same time, even though the duplicative text may be spread throughout the text data.
The examiner disagrees. Although the MEPE does not require the claimed language to explicitly recite the improvement described in the specification, the MPEP 2106 does specify the claimed invention must include positively recited language integrating the abstract idea into practical application and/or indicating significantly more than the judicial exception. The claimed language, as recited and mentioned in the applicant’s remarks, are merely directed towards the abstract idea without positively recited language indicating significantly more than the judicial exception and/or integrating the abstract idea into practical application. Recitation of language such as identifying a plurality of portions of text data, generating embedding based on each portion of text data is merely organizing data and mapping data such as words of the text to numbers. This can be performed mentally by a human using pen and paper. Further language of generating clusters or groups based on the embeddings, generating cluster summary of each cluster of embeddings using a first language model, gathering or adding cluster summary of a set of cluster summaries and generating a final summary based on the set of cluster summaries are written documentation (summaries) of clustered or gathered organized data (embeddings) that can be performed by a human using pen and paper. The language does not include positively recited language integrating the abstract idea into practical application and/or indicating significantly more than the judicial exception. Hence, as per the office action (below as a well as non-final rejection mailed 2/10/2026), the claims are ineligible.
Based on the foregoing, the claims improve computer-related technology related to Al summarization. Thus, the claims are not directed to an abstract idea. Reconsideration and withdrawal of the rejection under 35 U.S.C. 101 is therefore respectfully requested.
The examiner disagrees. As explained above, the claims are ineligible. Please see the remarks above.
III. THE ISSUE RLEATED TO THE CITED ART
Claims 1, 2, 9, 10, 17, and 18 are rejected under 35 U.S.C. 103 as allegedly unpatentable over Mattivi in view of Somech.
A. CLAIM 1
Claim 1 recites:
A method comprising:
identifying a plurality of portions of text data;
for each portion of the plurality of portions, generating an embedding based on said each portion;
based on a plurality of embeddings that are generated for the plurality of portions, generating a plurality of clusters of embeddings; for each cluster of embeddings of the plurality of clusters of embeddings: generating, by a first language model, a cluster summary based on portions, of the plurality of portions, that correspond to embeddings associated with said each cluster of embeddings; adding the cluster summary to a set of cluster summaries; generating, using a second language model, a final summary based on the set of cluster summaries;
wherein the method is performed by one or more computing devices.
(Emphasis added.) The Office Action cites label 535 of Fig. 5, labels 725 and 730 of Fig. 7, and paragraph 90 of Mattivi for allegedly disclosing the bolded portion of Claim 1. According to paragraph 81 of Mattivi, a document representation module generates an inferred document representation for a document at step 535. The inferred document representation includes each of the inferred per-type section cluster identifiers for each of the sections found in the document. Specifically, the inferred document representation is a vector of the inferred per- type section cluster identifiers. Thus, the Office Action equates Mattivi's inferred document representation with the recited final summary. However, Mattivi's inferred document representation is not generated using a model, is not a summary in any way, and is not generated based on cluster summaries that were generated by a language model.
The examiner disagrees. The applicant’s remarks indicate Mattivi’s inferred document representation with the recited final summary is not a summary and is not generated based on cluster summaries that were generated by a language model. The applicant’s remarks, as well as the recited limitation, does not specify what would be constituted as a final summary. Is it a final inferred prediction made based on the set of cluster summaries? Is it a final drafted essay summarizing topics of the set of cluster summaries?
Due to the breath of the recited claimed language, the limitation, as highlighted in the applicant’s remarks, is interpreted as a final summary, generated using a second language model, based on the set of cluster summaries. Mattivi’s prediction, Fig. 7, label 735, is the final summary. This final summary is generated based on inferred document representation and using a machine learning model, 730. Paragraph 90 discloses “The inferred document representation has numerical features identifying the content found in the various sections of the contract that makes the representation ideal as input to the machine learning model 730. Paragraph 79 discloses the document representation model may assign a cluster to the section of the multi-section document in Operation 525. “In some embodiments, the cluster of a section in a multi-section document is determined based at least in part on the per-type section cluster for the section as determined in accordance with aspects of the process 400 of Fig. 4.” This indicates the document representation module outputting the inferred per-type section cluster identifier for the per-type section cluster assigned to the section of the multi-section document, such as the inferred document representation of Fig. 7, label 725, is determined in accordance to Fig. 4. This indicates the prediction output from label 735 (final summary) is generated based on the set of cluster summaries (process performed in Fig. 4) using a machine learning model (Fig. 7, label 730). For these reasons, the examiner believes Mattivi discloses the recited limitation as indicated in the office action (below as well as previous).
The Office Action asserts that "Somech et al discloses natural language model generating summarization versions of the relevant portions of content using one or more natural language models (paragraph 9)." However, this cited paragraph fails to teach or suggest inputting auto-generated summaries into a language model to generate a summary. According to paragraph 111 of Somech, summarized versions of content portions are simply aggregated to yield an aggregated summary. Like Mattivi, no language model in Somech is involved in this recited final summary generation step.
The examiner disagrees. As explained above, Mattivi discloses the recited limitation. Somech is presented to demonstrate the machine learning model of Mattivi (Fig. 7, label 730) can be a language model generating a summary as disclosed by Somech. The office action includes a motivation for the combination of Mattivi in view of Somech. Please see the office action.
Based on the foregoing, Mattivi and Somech fail to teach or suggest all the features of Claim 1. Therefore, Claim 1 is patentable over Mattivi and Somech. Reconsideration and withdrawal of the rejection of Claim 1 under 35 U.S.C. § 103(a) is therefore respectfully requested.
The examiner disagrees. Please see the rebuttal above.
B. CLAIM 9 AND 17
Claim 9 is an independent claim that recites the features of Claim 1 discussed above that render Claim 1 patentable over the art of record. Therefore, Claim 9 is patentable over the art of record for at least the same reasons given above for Claim 1.
Claim 17 is an independent claim that recites the features of Claim 1 discussed above that render Claim 1 patentable over the art of record. Therefore, Claim 17 is patentable over the art of record for at least the same reasons given above for Claim 1.
The examiner disagrees. Please see the rebuttal above for claim 1.
C. DEPENDENT CLAIMS
Claims 2-8, 10-16, and 18-20 are dependent claims, each of which depends (directly or indirectly) on one of the claims discussed above. Each of Claims 2-8, 10-16, and 18-20 is therefore allowable at least for the same reasons given above for the claim on which it depends. In addition, each of Claims 2-8, 10-16, and 18-20 introduces one or more additional limitations that may independently render it patentable over the cited art. However, due to the fundamental differences already identified and to expedite the positive resolution of this case, a separate discussion of each of those limitations is not included at this time. The Applicants reserve the right to further point out the differences between the cited art and the novel features recited in Claims 2-8, 10-16, and 18-20.
The examiner disagrees. Please see the rebuttal for respective independent claims.
For the reasons indicated above, the previous office action stands as previously stated. A copy is found below.
Drawings
The drawings were received on 5/28/2024. These drawings are accepted.
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 1-8 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea in the form of a mental process without significantly more. The claim(s) recite(s) clustering embeddings of portions of text, cluster summaries of the clusters and a final summary based on the cluster summaries. Such is directed towards actions performed by a human being mentally using pen and paper to take text portions and generate embeddings such as a vector or list of characteristics of the text portions, clustering or grouping such embeddings, summarizing the clusters and generating a summary based on the cluster summaries. The limitations additionally recites using language models, such as a first language model, a second language model and one or more computing devices. Such recitation is merely directed towards generic device performing the abstract idea. This judicial exception is not integrated into a practical application because the recited claimed language is merely directed towards the judicial exception without positively recited language integrating the abstract idea into practical application. The claim(s) does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception because the recited claimed language is merely directed towards the judicial exception without positively recited language indicating significantly more than the abstract idea.
Claims 2-8 recite language adding to the judicial exception, but does not include positively recited language indicating significantly more than the abstract idea and/or integrating the abstract idea into practical application.
Claims 9-16 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea in the form of a mental process without significantly more. The claim(s) recite(s) clustering embeddings of portions of text, cluster summaries of the clusters and a final summary based on the cluster summaries. Such is directed towards actions performed by a human being mentally using pen and paper to take text portions and generate embeddings such as a vector or list of characteristics of the text portions, clustering or grouping such embeddings, summarizing the clusters and generating a summary based on the cluster summaries. The limitations additionally recites using language models, such as a first language model, a second language model and “one or more non-transitory storage media storing instructions which, when executed by one or more computing devices”. Such recitation is merely directed towards generic device performing the abstract idea. This judicial exception is not integrated into a practical application because the recited claimed language is merely directed towards the judicial exception without positively recited language integrating the abstract idea into practical application. The claim(s) does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception because the recited claimed language is merely directed towards the judicial exception without positively recited language indicating significantly more than the abstract idea.
Claims 10-16 recite language adding to the judicial exception, but does not include positively recited language indicating significantly more than the abstract idea and/or integrating the abstract idea into practical application.
Claims 17-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea in the form of a mental process without significantly more. The claim(s) recite(s) clustering embeddings of portions of text, cluster summaries of the clusters and a final summary based on the cluster summaries. Such is directed towards actions performed by a human being mentally using pen and paper to take text portions and generate embeddings such as a vector or list of characteristics of the text portions, clustering or grouping such embeddings, summarizing the clusters and generating a summary based on the cluster summaries. The limitations additionally recites using language models, such as a first language model, a second language model, one or more computing devices and “one or more non-transitory storage media storing instructions which, when executed by the one or more computing devices”. Such recitation is merely directed towards generic device performing the abstract idea. This judicial exception is not integrated into a practical application because the recited claimed language is merely directed towards the judicial exception without positively recited language integrating the abstract idea into practical application. The claim(s) does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception because the recited claimed language is merely directed towards the judicial exception without positively recited language indicating significantly more than the abstract idea.
Claims 18-20 recite language adding to the judicial exception, but does not include positively recited language indicating significantly more than the abstract idea and/or integrating the abstract idea into practical application.
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.
Claim 2 recites the limitation "the embeddings" in claim 1. There is insufficient antecedent basis for this limitation in the claim.
Claim 10 recites the limitation "the embeddings" in claim 9. There is insufficient antecedent basis for this limitation in the claim.
Claim 18 recites the limitation "the embeddings" in claim 17. There is insufficient antecedent basis for this limitation in the claim.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 1,2,9,10,17,18 is/are rejected under 35 U.S.C. 103 as being unpatentable over Mattivi et al (US Publication No.: 20220215274) in view of Somech et al (US Publication No.: 20190005024).
Claim 1, Mattivi et al discloses
Identifying a plurality of portions of text data (Fig. 5, label 510. Paragraph 77 discloses receiving a multi-section document with content of the document organized into different sections. Fig. 5, label 515 selects a section of the document.);
For each portion of the plurality of portions, generating an embedding based on said each portion (Fig. 5, label 520. Paragraph 78 discloses generate an embedding representation of the section selected at 515.);
Based on a plurality of embeddings that are generated for the plurality of portions, generating a plurality of clusters of embeddings (Fig. 5, label 525. Paragraph 79 discloses assign a cluster to the section of the multi-section document at 525. Cluster assignment is determined based on per-type section cluster for the section as per process shown in Fig. 4. Fig. 4, label 430. Paragraph 69 discloses clustering of the embedding representations having similar semantic meaning together.);
For each cluster of embeddings of the plurality of clusters of embeddings:
Generating, by a first language model, a cluster summary based on portions, of the plurality of portions, that correspond to embeddings associated with said each cluster of embeddings (Fig. 4, label 440 generates a cluster summary for each cluster. Paragraph 72,73 discloses a document summarization machine learning model or first language model to generate a per-type section cluster summary.);
Adding the cluster summary to a set of cluster summaries (Fig. 4, label 435-445 shows a loop for continuous cluster section and cluster summarization which indicates an adding of cluster summary to the set of cluster summaries when a cluster summary for a selected section subset for cluster is performed.);
Generating, using a second model, a final summary based on the set of cluster summaries (Fig. 5, label 535 generates an inferred document where Fig. 7 shows an example of using the inferred document to generate a prediction or summary of relevant features with respect to the claim being likely subject to overpayment. Fig. 7, label 725, as the inferred document generated via Fig. 4,5. Fig. 7, label 730 and paragraph 90 discloses a machine learning model is used to generate the prediction or summary as indicated in paragraph 90.);
Wherein the method is performed by one or more computing devices (Fig. 2 shows one or more computing devices).
Mattivi et al discloses machine learning model as the second model in paragraph 90 and discloses the output or final summary or prediction includes relevant contract features along with a summary of each relevant section of the contract explaining its importance. (paragraph 90) Although Mattivi et al discloses a summary of relevant sections of the contract as part of the output from the machine learning model as per paragraph 90, Mattivi et al fails to disclose the machine learning model is a natural language model.
Somech et al discloses natural language model generating summarization versions of the relevant portions of content using one or more natural language models (paragraph 9).It would be obvious to one skilled in the art before the effective filing date of the application to substitute one well known element of Mattivi et al’s machine learning model that generates summary of relevant sections of the contract with predictions with another well-known element of a natural language model to generate summary of relevant content as disclosed by Somech et al so to yield predictable results of outputting a summary.
Claim 2, Mattivi et al discloses wherein the embeddings, associated with a first cluster of embeddings in the plurality of clusters of embeddings, upon which a first cluster summary is based is less than all embeddings that are associated with the first cluster embeddings (Paragraph 69 discloses clustering embeddings of similar semantic meaning together. Fig. 4, label 415,420,425 continuously generates embeddings for sections in the contract or text. Label 430 performs clustering. Since clustering is performed based on semantic similarity, this indicates embeddings generated for each section can be grouped together. As the embeddings are grouped or clustered together, the embeddings yet to be clustered can be less than embeddings already clustered.).
Claim 9 recites similar limitations as claim 1 and is rejected on the same grounds as claim 1.
Claim 10 recites similar limitations as claim 2 and is rejected on the same grounds as claim 2.
Claim 17 recites similar limitations as claim 1 and is rejected on the same grounds as claim 1. In addition, Mattivi et al discloses one or more computing devices (Fig. 2); one or more non-transitory storage media storing instructions which, when executed by the one or more computing devices (Paragraph 55 discloses processing element as microprocessors executing instructions stored in media such as memory such as label 220, paragraph 56.) cause the performance of limitations similarly rejected in claim 1 (please see claim 1).
Claim 18 recites similar limitations as claim 2 and is rejected on the same grounds as claim 2.
Allowable Subject Matter
Claims 3-8,11-16,19-20 are objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims.
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 LINDA WONG whose telephone number is (571)272-6044. The examiner can normally be reached 9-5.
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, Andrew C Flanders can be reached at 571-272-7516. 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.
/LINDA WONG/Primary Examiner, Art Unit 2655
Prosecution Insights