CTFR 18/779,474 CTFR 86955 DETAILED ACTION 07-03-aia AIA 15-10-aia The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA. This Office Action is in response to the claim amendment filed on April 1, 2026 and wherein claims 1-3, 6, 8-10, 13, 15-17, 20 amended. In virtue of this communication, claims 1-20 are currently pending in this Office Action. With respect to the objection of claims 1-20 due to formality issues, as set forth in the previous Office Action, the claim amendment, and argument, see title “IV Response to Claim Objections” of page 7 in Remarks filed on April 1, 2026, have been fully considered and the argument is persuasive. Therefore, the objection of claims 1-20 due to the formality issues, as set forth in the previous Office Action, has been withdrawn . With respect to the rejection of claims 15-20 under 35 USC §101 , as set forth in the previous Office Action, the Applicant’s amendment, and argument, see paragraphs 1-2 of page 8 in Remarks filed on April 1, 2026, have been fully considered and the argument is persuasive and wherein the application specification read “computer readable storage medium, …, is not to be construed as … transitory signals per se …”, etc. i.e., the claimed “one or more computer-readable storage media” comprised in “a computer program product” has been construed not to include “transitory signals per se” and therefore, the rejection of claims 15-20 under 35 USC §101 , as set forth in the previous Office Action, has been withdrawn . The Office appreciates the explanation of the amendment and analyses of the prior arts, and however, although the claims are interpreted in light of the specification, limitations from the specification are not read into the claims. See In re Van Geuns, 988 F.2d 1181, 26 USPQ2d 1057 (Fed. Cir. 1993) and MPEP 2145. Claim Rejections - 35 USC § 103 07-20-aia AIA 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 of this title, 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. 07-21-aia AIA Claim s 1-3, 5-6, 8-10, 12-13, 15-17, 19-20 are rejected under 35 U.S.C. 103 as being unpatentable over Devesa (US 20200097814 A1) and in view of reference Arras et al. (Explaining Recurrent Neural Network Predictions in Sentiment Analysis”, hereinafter Arras, arXiv:1706.07206v2[cs.CL], https://arxiv.org/pdf/1706.07206, pp.1-10, August 4, 2017), and Lin et al. (US 11537616 B1, hereinafter Lin) . Claim 1 : Devesa teaches a processor-implemented method (title and abstract, ln 1-17, method steps in figs. 5, 7) , the method comprising: identifying data from one or more conversations (extracting a plurality of dialogue conversation from the corpus of the medical-training dataset upon created word embedding, para 24, 33) , wherein the data includes at least one request and at least one response (question-answer pairs in the corpus of the medical-training dataset, para 32 and question/answer presented in the extracted dialogue conversation, para 34) ; training a machine learning model (training a virtual medical assistant system in fig. 5, para 125-126) on the identified data using long short-term memory (the tokenized data as vectors representation manipulated by using LSTM applied in RNN as units, a cell, an input gate, an output gate and a forget gate, para 107 and tokenizing the context and utterances present in the word embedding of words created from the corpus of medical-training dataset, para 16) ; and assisting a user in responding with a new response to a new request using the trained model (fig. 2, a user 102 in interactive dialogue session to the virtual medical assistant system 204, para 69, interactive with the computing device 104 with the virtual medical assistant system 204, including a first question from the user 102, para 82-83 and return an appropriate utterance as response to the enquired question in the bi-directional conversation, para 87) . However , Devesa does not explicitly teach with layer-wise relevance propagation in using long short-term memory LSTM and does not explicitly teach and wherein the disclosed assisting comprises generating the new response as a priority-based response comprising a timing and a prioritization of the new request, deprioritizes one or more other requests under the new request, and estimates a time to completion of the new request. Arras teaches an analogous field of endeavor by disclosing a processor-implemented method (title and abstract, ln 1-19, implementing algorithms LRP applied on LSTMs of RNN architectures, session 2.2 Layer-wise Relevance Propagation LRP, p.2) and wherein using a long short-term memory LSTM is disclosed (used in RNN architectures, session 2.2, p.2) with layer-wise relevance propagation LRP (using the trained bi-LSTM with application of extended LRP method, session 1 Introduction, col 2, para 5-6, p.1, e.g., relevance decomposed with LRP, session 3 Recurrent Model and Data, p.3) for benefits of improving a performance of machine learning model (by producing reliable explanation of words that are responsible for attributing sentiment in individual texts, session 1, Introduction, col 2, para 5-6, p.1 and col 1, para 1, p.2 and by emphasizing extremal LRP relevance corresponding to positive sentiment values of the words for target classification, session 4.2 Representative Words for a Sentiment, col 1, p.6) . Therefore , it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have applied the layer-wise relevance propagation used in the long short-term memory, as taught by Arras, to using the long short-term memory in the processor-implemented method, as taught by Devesa, for the benefits discussed above. However , the combination of Devesa and Arras does not explicitly teach wherein the disclosed assisting comprises generating the new response as a priority-based response comprising a timing and a prioritization of the new request, deprioritizes one or more other requests under the new request, and estimates a time to completion of the new request. Lin teaches an analogous field of endeavor by disclosing a method (title and abstract, ln 1-8 and method steps in figs. 6-8 and executed in processing cluster 300) and wherein assisting a user in responding with a new response to a new request (a query or request 302 initiated by a user or client via an interface to interact with the processing cluster 220 in fig. 2 or 300 in fig. 3, col 6, ln 50-58 and for database services 210 in fig. 2) using a trained model (a trained machine learning model applied for predicting a performance measure, e.g., execution time and/or resource usage, col 2, ln 49-56) is disclosed and wherein the assisting comprises generating the new response as a priority-based response (final result for a request achieved via prediction-based prioritization 315, etc., in fig. 3, col 9, ln 21-24) comprising a timing (e.g., the shorter the execution time of a query is, the higher the priority is and then earlier performed, col 2, ln 15-25, i.e., timing for the response) and a prioritization of the new request (assigned by dynamic prioritization 440 based on query performance measurement prediction 430 and actual query performance 406 in figs. 4, and dynamic updating priority based on actual query performance in fig. 7, col 12, ln 33-41) , down-prioritizes one or more other requests under the new request (e.g., a query having two seconds execution time down-prioritizes the query having 5 minute execution time) , and estimates a time to completion of the new request (predicted execution time, col 2, ln 49-52) for benefits of improving operation effectiveness and efficiency at system level (by allowing efficient utilization of system resources, col 2, ln 25-32). Therefore , it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have applied the assisting the user in responding to the new request using the trained model and wherein the assisting comprises generating the new response as the priority-based response comprising the timing and the prioritization of the new request, down-prioritizes one or more other requests under the new request, and estimates the time to completion of the new request, as taught by Lin, to the assisting the user in responding to the new request using the trained model in the method, as taught by the combination of Devesa and Arras, for the benefit discussed above. However , the combination of Devesa, Arras, and Lin does not explicitly teach wherein down-prioritizing is de-prioritizing. It has been a recognized problem and need in the art, which may include a design need to solve the problem for prioritizing a task to be completed such as response from a query or a request in a designed priority level or order for better utilization of system resources and there had been a finite number of identified, predictable potential solutions to the priority level or degree of performing the task: down-prioritizing one or more other requests if they are insignificant while compared to others at the moment, prioritizing a request if it is significant at the moment compared to others, de-prioritizing one or more others if there is no chance to be significant while compared to others at the moment, it would have been obvious for one having ordinary skill in the art before the effective filing date of the claimed invention to have pursued the known potential solutions with a reasonable expectation of success or obvious to try, see MPEP 2141, III. Therefore , it would have been obvious for one having ordinary skill in the art before the effective filing date of the claimed invention to have applied deprioritizing one or more other requests, as taught in the obvious to try above, to down-prioritizing one or more other request, as taught by the combination of Devesa, Arras, and Lin, for the benefits discussed above. Claim 8 has been analyzed and rejected according to claim 1 above and the combination of Devesa, Arras, and Lin further teaches, a computer system (Devesa, a computing device 104 in fig. 1, e.g., laptop, smartphone, tablet, PDA, etc., para 58) , comprising: a processor set (Devesa, one or more processors, para 23 and Lin, computer system, col 3, ln 62-67, col 4, ln 1-4 and GPUs of at least one processor 1010 in fig. 9, col 14, ln 47-51) , one or more computer-readable storage media (Devesa, a variety of computer-readable media, para 138 and Lin, memory 1020, including RAM, etc., col 15, ln 1-7) , and program instructions stored on the one or more computer-readable storage media (Devesa, memory to store instructions, para 23 and Lin, program and instructions stored via the computer-readable medium, col 15, ln 20-26) to cause the processor set to perform operation of claim 1 (discussed in claim 1 above) . Claim 15 has been analyzed and rejected according to claims 1, 8 above. Claim 2 : the combination of Devesa, Arras, and Lin further teaches, according to claim 1 above, wherein training the machine learning model is further performed using a bidirectional long short-term memory with layer-wise relevance propagation (Devesa, LSTM and discussed in claim 1 above, and Arras, word-based bi-directional LSTM used with the Layer-wise Relevance Propagation LRP, abstract, and session 3 Recurrent Model and Data, p.3 and Session 5 Conclusion, p.7) . Claim 3 : the combination of Devesa, Arras, and Lin further teaches, according to claim 2 above, wherein training the machine learning model is further performed using a bidirectional long short-term memory with an epsilon layer-wise relevance propagation (Devesa, the LSTM and discussed in claim 1 above, and Arras, epsilon layer-wise relevance propagation in equation R i ←i PNG media_image1.png 59 310 media_image1.png Greyscale as extended relevance with further weighted by epsilon ε and delta δ, session 2.2 Layer-wise Relevance Propagation LRP, p.2, col 2, and p.3, col 1) . Claim 5 : the combination of Devesa, Arras, and Lin further teaches, according to claim 1 above, the method further comprising: collecting feedback about the assisting (Devesa, feedback for manual annotation of questions, para 75) ; further training the model based on the collected feedback (Devesa, following the manual annotation, a binary annotation of the dialog conversations is performed and building automatic classifiers for prioritizing the plurality of dialogue conversations, para 75 and the prioritization is done to train the virtual medical assistance system 204, para 78) . Claim 6 : the combination of Devesa, Arras, and Lin further teaches, according to claim 1 above, wherein identifying data further includes analyzing the data using a sentiment analysis (Devesa, extracting a plurality of dialogue conversation from the corpus of the medical-training dataset upon created word embedding, para 24, 33, and the word embedding represented as retrieved specific concept, para 29, and the specific concept is retrieved based on a determination of whether the specific concept match with the one or more medical protocols, para 29, i.e., through data analyses, and Arras, sentiment analysis in long-range interactions in texts for extending LRP and then, applying the extended LRP to bi-directional LSTM to produce reliable explanations of which words are responsible for distributing sentiment in individual texts, session 1 Introduction, col 2, the last two paragraphs and also, decomposing the words according to sentimental positive and negative for a target class, session 4.1 Decomposing Sentiment onto Words, p.4) . Claim 9 has been analyzed and rejected according to claims 8, 2 above. Claim 10 has been analyzed and rejected according to claims 9, 3 above. Claim 12 has been analyzed and rejected according to claims 8, 5 above. Claim 13 has been analyzed and rejected according to claims 8, 6 above. Claim 16 has been analyzed and rejected according to claims 15, 2 above. Claim 17 has been analyzed and rejected according to claims 16, 3 above. Claim 19 has been analyzed and rejected according to claims 15, 5 above. Claim 20 has been analyzed and rejected according to claims 15, 6 above . 07-21-aia AIA Claim s 4, 7, 11, 14, 18 are rejected under 35 U.S.C. 103 as being unpatentable over Devesa (above) and in view of references Arras (above), Lin (above), and Wang et al. (CN 117438047 A, hereinafter Wang, original and its translation attached herein, and the translation version is referred for paragraph citation under Wang below) . Claim 4 : the combination of Devesa , Arras , and Lin further teaches, according to claim 1 above, wherein assisting the user includes word explanations (the user 102 in interactive dialogue with the virtual medical assistant system 204, para 69, the user requested a first question, para 82-83, and return an appropriate utterance as response to the enquired question, para 87 and Arras, providing reliable explanation of which words are responsible for attributing sentiment in individual texts, session 1 Introduction, last two paragraph of col 2, p.1 and col 1, para 1, p.2) , except providing reasoning about the new response. Wang teaches an analogous field of endeavor by disclosing a processor-implemented method (title and abstract, ln 1-15 and method steps in figs. 1-2 and implemented by a processor 910 in fig. 9, para 2, p.30) and wherein assisting a user is disclosed (the trained psychological consultation model can be assistant of psychological counselor, teacher at school to achieve a real-time psychological counseling and proposal to answer question, para 1, p.26) to include providing reasoning about the new response (about providing pertinent suggestion and solution, the ability of event cause reasoning and emotion analysis identification are provided to not just determine the psychological intention of the user, but also infer the potential reason of the problem, and understand the emotional state of the user in the dialogue process, para 1, p.30 and by using a large language model LLM, para 1, p.14) for benefits of improving the efficiency and effects of machine learning model (by constructing reasoning chain, last paragraph of p.23, by pre-training on the model by using a large amount of domain knowledge, para 5 of p.24) and improving the performance of the model (by pre-processed training data and optimizing the parameter and loss function of the model, para 3, p.11 and by improving the accuracy of the recognition of the psychological intention of the user and realizing the credible and interpretable recognition process, para 2, p.15) . Therefore , it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have applied wherein assisting the user includes providing reasoning about the new response, as taught by Wang, to assisting the user including word explanations in the processor-implemented method, as taught by the combination of Devesa and Arras, for the benefits discussed above. Claim 7 : the combination of Devesa, Arras, Lin, and Wang further teaches, according to claim 1 above, wherein assisting the user includes generating the new response using a large language model (Devesa, the user 102 in interactive dialogue session to the virtual medical assistant system 204, para 69, including a first question from the user 102, para 82-83 and return an appropriate utterance as response to the enquired question in the bi-directional conversation, para 87 and by using the trained virtual medical assistance system 204 above, and Arras, using the model having extended LRP with the recurrent neural network RNN having bidirectional LSTM model, session 4.1 Decomposing Sentiment onto Words, col 2, p.4, and Wang, using the large language model LLM, para 1, p.14) . Claim 11 has been analyzed and rejected according to claims 8, 4 above. Claim 14 has been analyzed and rejected according to claims 8, 7 above. Claim 18 has been analyzed and rejected according to claim 15, 4 above. The prior art (US 20210263663 A1) by Bansal et al. and (US 20180034763 A1) by Rincon et al. made of record and not relied upon is considered pertinent to applicant's disclosure because Bansal and Rincon above disclosed prioritizing requests and deprioritizing other requests, which is part of the disclosures disclosed by the instant application. Response to Arguments Applicant's arguments filed on April 1, 2026 have been fully considered and but are moot in view of the new ground(s) of rejection necessitated by the applicant amendment. The Office has thoroughly reviewed Applicants' arguments but firmly believes that the cited references to reasonably and properly meet the claimed limitations. In the response to this office action, the Office respectfully requests that support be shown for language added to any original claims on amendment and any new claims. That is, indicate support for newly added claim language by specifically pointing to page(s) and line numbers in the specification and/or drawing figure(s). This will assist the Office in prosecuting this application. Conclusion 07-40 AIA Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL . See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a). A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action. Any inquiry concerning this communication or earlier communications from the examiner should be directed to LESHUI ZHANG whose telephone number is (571)270-5589. The examiner can normally be reached Monday-Friday 6:30amp-4:00pm 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, Vivian Chin can be reached at 571-272-7848. 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. /LESHUI ZHANG/ Primary Examiner, Art Unit 2695 Application/Control Number: 18/779,474 Page 2 Art Unit: 2695 Application/Control Number: 18/779,474 Page 3 Art Unit: 2695 Application/Control Number: 18/779,474 Page 4 Art Unit: 2695 Application/Control Number: 18/779,474 Page 5 Art Unit: 2695 Application/Control Number: 18/779,474 Page 6 Art Unit: 2695 Application/Control Number: 18/779,474 Page 7 Art Unit: 2695 Application/Control Number: 18/779,474 Page 8 Art Unit: 2695 Application/Control Number: 18/779,474 Page 9 Art Unit: 2695 Application/Control Number: 18/779,474 Page 10 Art Unit: 2695 Application/Control Number: 18/779,474 Page 11 Art Unit: 2695 Application/Control Number: 18/779,474 Page 12 Art Unit: 2695 Application/Control Number: 18/779,474 Page 13 Art Unit: 2695
Prosecution Insights