DETAILED ACTION
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim Rejections - 35 USC § 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 1 to 7, 11 to 16, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Strader et al. (U.S. Patent Publication 2019/0122766) in view of Mathur et al. (U.S. Patent Publication 2024/0412226).
Concerning independent claims 1, 5, and 14, Strader et al. discloses a method, system, and computer program product for patient-provider conversation and auto-generation of a note or summary, comprising:
“receive an input text for performing [a text analysis task or] a request to perform a summarization task on a natural language text” – a note summarization of a conversation between a patient and a healthcare provider is performed (“a summarization task”) (Abstract); audio input 102 is provided to a machine learning model 104 which generates a transcript (“a natural language text” or “receive an input text”) of a recording (¶[0034]: Figure 1); implicitly, there is a control for activating an interface to perform auto-generation of a summary (“receive a request to perform a summarization task”); here, a summarization task is a specific “text analysis task”; Compare Specification, ¶[0013], which states that text analysis tasks include tasks to generate medical summaries;
“extract one or more domain entities from the natural language text using a machine learning model trained to recognize entities of a domain in a given text” – words or phrases in the transcript related to medical topics (“entities of a domain”) are extracted with the aid of a trained machine learning model (“using a machine learning model”) (Abstract); a named entity recognition model 112 processes the text to recognize medically relevant words or phrases; a named entity recognition (NER) model can be trained from annotated medical encounter transcripts (“trained to recognize entities of a domain in a given text”) (¶[0035]: Figure 1); words or phrases in the transcript are extracted relating to medical topics with the aid of a trained machine learning model (NER model 112) (¶[0042]: Figure 1);
“insert the one or more domain entities into instructions as part of generating the instructions to perform the summarization task [or text analysis task] [using a pre-trained large language model fine-tuned to the domain]” – a note region simultaneously displays elements of a note summarizing the conversation; extracted words or phrases are highlighted in the transcript and displayed in the note region (Abstract); a note region summarizes the conversation in a note, the note including automatically extracted words or phrases in the transcript related to medical topics (¶[0005]); a method links or maps between extracted/highlighted words or phrases in note region 314 and portions of the transcript from which the extracted words or phrases originated (¶[0043]: Figure 3); highlighted phrases ‘leg hurts’ 402 and ‘feeling feverish’ 403 are extracted from the transcript (¶[0046]: Figure 4); here, words or phrases extracted from a transcript are “one or more domain entities”, and “generating instructions” link these words or phrases from a transcript of text input into a note summary;
“input the generated instructions to the pre-trained [large language] model [fine-tuned to the domain to prompt the pre-trained large language model fine-tuned to the domain] to perform the summarization task [or text analysis task] on the natural language text using the generated instructions” – a method automatically generates a note summarizing a conversation between a patient and a healthcare provider (Abstract); medically relevant words or phrases in the transcript are extracted from the conversation and automatically populated into a note; extracted words or phrases in the note are linked to the corresponding portions of the transcript from which they originated (¶[0006]); a service provider includes a machine learning system 218 that implements machine learning models 104; a service provider implementing the machine learning models generates the transcript and note (¶[0038]: Figure 2); machine learning models are used in extracting highlighted text in the transcript; machine learning models are trained from labeled training data (”the pre-trained . . . model”) (¶[0054]: Figure 12);
“provide the result of the summarization task [or the text analysis task] performed on the natural language text” – a note region simultaneously displays elements of a note summarizing the conversation; extracted words or phrases are highlighted in the transcript and displayed in the note region (Abstract); a user selects a note tab 306 on workstation 210, and a transcript is displayed along with a note in region 314 (¶[0050]: Figure 8); Figure 20 illustrates a transcript 2000 and a summary 2008 on display region 314.
Concerning independent claims 1, 5, and 14, Strader et al. discloses all of the limitations with the exception of performing a summarization task using “a pre-trained large language model fine-tuned to the domain” and inputting the generated instructions “to the pre-trained large language model fine-tuned to the domain to prompt the pre-trained large language model pre-trained to the domain” to perform the summarization task on the natural language text. Generally, Strader et al. discloses training a machine learning model to a medical domain, which arguably provides “a pre-trained . . . model fine-tuned to the domain.” Still, Strader et al. does not disclose “a large language model” or “instructions . . . to prompt the pre-trained large language model”. Applicants’ independent claims set forth limitations of “perform a summarization task” and “a text analysis task”. Here, Strader et al. discloses performing “a summarization task” of summarizing a conversation between a patient and a healthcare provider, and “a text analysis task” implicitly is a task of broader scope that includes a summarization task. Compare Specification, ¶[0013], which states that text analysis tasks include tasks to generate medical summaries.
Concerning independent claims 1, 5, and 14, Mathur et al. teaches a system and method for providing a response to a product help inquiry that includes receiving a product help inquiry, classifying the product help inquiry as being associated with a topic, and generating a prompt based on the product help inquiry. (Abstract) Large language models (LLMs) are used to provide responses relating to specific domains including specific products or topics, so it is important to finetune an LLM when it is being used for a specific domain. (¶[0001]) Artificial intelligence (AI) models can recognize, summarize, translate, and generate text based on large language models (LLMs). (¶[0015]) A technical solution evaluates LLM responses based on multiple factors, and a metric is then used to correctly create prompts for the LLM, e.g., in prompt engineering. (¶[0018]) LLM 130 is a large language model which may be a deep learning algorithm that can recognize, summarize, translate and generate text. Model evaluation system 110 utilizes a prompt generation engine to generate a prompt that is likely to result in accurate and/or relevant responses from LLM 130. (¶[0024] - ¶[0025]: Figure 1) User query 210 may be in the form of text or voice, and pre-processing engine 220 performs pre-processing operations on user query 210. (¶[0029]: Figure 2) Contextual embeddings are capable of achieving state-of-the-art performance in various natural language processing (NLP) tasks including text classification, named entity recognition, machine translation, and question answering. (¶[0037]: Figure 2) A copilot assistant, e.g., language model, provides summarized instructions on how to create a calendar meeting using an application by displaying the instructions in a UI element 414. Instructions displayed in UI element 420 summarize the response. (¶[0044] - ¶[0045]: Figures 4A to 4C) Mathur et al., then, teaches providing a prompt to a pre-training a large language model (LLM) that is it fine-tuned to a domain to perform a summarization task or a text analysis task. An objective is to obtain an improved system and method with accurate and relevant responses from LLMs. (¶[0001] - ¶[0002]) It would have been obvious to one having ordinary skill in the art to perform a summarization of notes of a conversation between a patient and a healthcare provider in Strader et al. by prompting a pre-trained large language model that is fine-tuned to a domain as taught by Mathur et al. for a purpose of providing improved accuracy and relevance of responses from an LLM.
Concerning claims 2, 7, and 16, Strader et al. discloses that medically relevant words or phrases are recognized (“specify that the one or more domain entities are to be included”) and highlighted words or phrases are extracted as text data for note generation (“to be included in the result of the summarization task”) (¶[0035] - ¶[0036]: Figure 1); words or phrases in the transcript related to medical topics and a health condition of a patient are extracted and displayed in a second region (¶[0042]: Figure 3); highlighted phrases ‘leg hurts’ 402 and ‘feeling feverish’ 403 are extracted from a transcript and placed in a note (¶[0046]: Figure 4).
Concerning claims 3, 6, and 15, Strader et al. discloses that audio input is provided to speech to text conversation model 110, and text is generated by speech-to-text conversion model 110 from audio input (¶[0034] - ¶[0035]: Figure 1); a transcript is generated from an audio recording using a speech-to-text engine 110 (“generate natural language text as a transcript from obtained audio data using an automatic speech recognition system” or “generate the input text as a transcript from obtained audio data using an automatic speech recognition system”) (¶[0042]: Figure 1).
Concerning claims 4, 13, and 20, Strader et al. discloses automatically generating a note summarizing a conversation between a patient and healthcare provider from an audio recording (“a medical audio summarization service”) (Abstract); a computing environment of a clinic, hospital, or medical office 200 is a location of a visit between a healthcare provider 202 and patient 204; a recording device 206 at workstation 210 captures and records speech, and workstation 210 is used by provider 202 to view the transcript and note; audio input may be sent over a network 208 to a service provider which implements machine learning models (“offered as part of a provider network”); an entity or service provider generates the transcript and note and transmits them over an application programming interface to software resident on workstation 210 (“and wherein the request is received via an interface of the medical audio summarization service”) (¶[0037] - ¶[0038]: Figure 2).
Concerning claim 11, Strader et al. discloses a computing environment of a clinic, hospital, or medical office 200 is a location of a visit between a healthcare provider 202 and patient 204; audio input may be sent over a network 208 to a service provider which implements machine learning models (“sending one or more requests to a remote host for the machine learning model to perform recognition on the input text”); an entity or service provider generates the transcript and note and transmits them over an application programming interface to software resident on workstation 210 (¶[0037] - ¶[0038]: Figure 2).
Concerning claim 12, Strader et al. discloses automatically generating a note summarizing a conversation between a patient and healthcare provider from an audio recording (“wherein the text analysis task is a summarization task”) (Abstract). Compare Specification, ¶[0013], which states that text analysis tasks include tasks to generate medical summaries.
Claims 8 to 10 and 17 to 19 are rejected under 35 U.S.C. 103 as being unpatentable over Strader et al. (U.S. Patent Publication 2019/0122766) in view of Mathur et al. (U.S. Patent Publication 2024/0412226) as applied to claims 5 and 14 above, and further in view of Mahapatra et al. (U.S. Patent Publication 2023/0153533).
Concerning claims 8 and 17, Mathur et al. teaches pre-training of a large language model so that it is fine-tuned to a domain, but does not provides details for training by “a selection of the domain of a plurality of domains supported by the text analysis system, wherein the machine learning model and the pre-trained large language model correspond to the selected domain and are respectively selected for performing the text analysis task out of respective pluralities of machine learning models that recognize entities out of different ones of the plurality of domains and pre-trained large language models fine-tuned to the different ones of the plurality of domains.” That is, Mathur et al. does not teach these details of training machine learning models for a plurality of domains.
Concerning claims 8 and 17, Mahapatra et al. teaches pre-training techniques for entity extraction models to facilitate domain adaptation. One or more selection models are used to select a subset of an available source domain corpus. (¶[0002]) One or more selection models are used to select a subset of available source domain training data for an entity extraction model based on similarity to a particular target domain. (¶[0016]) Application 110 and entity extraction tool 150 coordinate to train or adapt entity extraction model 160 to an appropriate domain. Application 110 is designed specifically for documents in a particular domain, e.g., contracts or financial documents. Document dashboard 115 may prompt a user to identify an applicable domain for documents 185, e.g., contracts, financial documents, biomedical documents, artificial intelligence research papers, scientific publications, political publications, musical publications, literary publications. Dashboard 115 prompts and/or accepts inputs from the user identifying ground truth entities (text and corresponding classifications) from a subset of target corpus 192. (¶[0028]: Figure 1) Predicted classes are compared with ground truth and updated to update entity extraction model 160. (¶[0035]) Mahapatra et al., then, teaches “a selection of the domain out of a plurality of domains supported by the text analysis system” by accepting user input to identify a domain for an application being trained to perform entity extraction. An objective is to perform entity extraction in a circumstance that models do not exist for every domain and do not provide sufficient accuracy in certain domains that diverge from training datasets. (¶[0012]) It would have been obvious to one having ordinary skill in the art to select a domain out of a plurality of domains supported by a text analysis system to recognize entities as taught by Mahapatra et al. to generate a note summarizing a conversation between a patient and a healthcare provider in Strader et al. for a purpose of performing entity extraction with sufficient accuracy for models that do not exist for every domain and domains that diverge from training datasets.
Concerning claims 9 and 18, Mathur et al. teaches pre-training of a large language model (LLM), but does not expressly provide for “receiving a request to fine-tune the pre-trained large language models for one or more additional domain entities, wherein the request identifies further training data for fine-tuning that includes one or more additional domain entities in ground truth data”, and “performing further fine-tuning on the pre-trained large language models for the domain using the further training data annotated with the one or more additional domain entities extracted from the ground truth data.” However, Mahapatra et al. teaches that one or more selection models are used to select a subset of available source domain training data for an entity extraction model. (¶[0016]) After pre-training a sentence selection model on an unlabeled target domain corpus, the sentence selection model is fine-tuned on the labeled target domain corpus and used to select sentences from a source domain corpus. (¶[0018]) Document dashboard 115 prompts input from a user identifying ground truth entities (text and corresponding classifications) from a subset of target domain corpus 192. Entity extraction model 160 is pre-trained on training sentences 198 selected from document domain corpus 196 and/or fine-tuned on labeled target domain corpus 194. (¶[0028] - ¶[0029]: Figure 1) Predicted classes are compared with ground truth and updated to update entity extraction model 160. (¶[0035]) Mahapatra et al., then, receives a request from a user to fine-tune a language model to a domain for one or more additional domain ground truth entities.
Concerning claims 10 and 19, Mathur et al. teaches pre-training of a large language model (LLM) that is fine-tuned to a domain, but does not expressly provide for “using domain entities from ground truth data included in the training data set.” However, Mahapatra et al. teaches that document dashboard 115 prompts input from a user identifying ground truth entities (text and corresponding classifications) from a subset of target domain corpus 192. Entity extraction model 160 is pre-trained on training sentences 198 selected from document domain corpus 196 and/or fine-tuned on labeled target domain corpus 194. (¶[0028] - ¶[0029]: Figure 1) Predicted classes are compared with ground truth and updated to update entity extraction model 160. (¶[0035]) Mahapatra et al., then, teaches fine-tuning a language model for a domain with ground truth entities.
Response to Arguments
Applicants’ arguments filed 08 January 2026 have been considered but are moot in view of new grounds of rejection necessitated by amendment.
Applicants amend independent claims 1, 5, and 14 to set forth a new limitation of “inputting the generated instructions to the pre-trained large language model fine-tuned to the domain to prompt” the pre-trained large language model. Then Applicants present arguments directed against the prior rejection of the independent claims as being obvious under 35 U.S.C. §103 over Strader et al. (U.S. Patent Publication 2019/0122766) in view of Leidner et al. (U.S. Patent Publication 2021/0043211). Specifically, Applicants argue that Strader et al. does not disclose the new limitations of “instructions . . . to prompt the pre-trained large language model”, and that the prior rejection relies upon a broad construction of “instructions” that does not address the new limitations of “to prompt the pre-trained large language model”.
Applicants’ amendments necessitate new grounds of rejection as directed to independent claims 1, 5, and 14 being obvious under 35 U.S.C. §103 over Strader et al. (U.S. Patent Publication 2019/0122766) in view of Mathur et al. (U.S. Patent Publication 2024/0412226). Specifically, Mathur et al. is maintained to teach a pre-trained large language model that is fine-tuned to a specific domain, and that generates a prompt for the large language model for text analysis tasks that include summarization based on a product help inquiry. The rejection of certain dependent claims continues to rely upon Mahapatra et al. (U.S. Patent Publication 2023/0153533).
Mainly, Strader et al. represents prior art directed to an earlier technology for summarizing a note that does not include more recent developments in machine learning with large language models that receive prompts as instructions to perform a task. Figure 1 of Strader et al. illustrates “instructions” including medically relevant entities 112 and data for note generation 114 as being input to machine learning model 104. Strader et al. is a prior art method of generating a note summarizing a conversation between a patient and a healthcare provider that would benefit from large language models fine-tuned to a domain that receive prompts to perform tasks including summarization of responses based on product help inquiries as taught by Mathur et al. A rejection based on a combination of Strader et al. and Mathur et al., then, is proper under a rationale of KSR International Co. v. Teleflex Inc. (KSR), 550 U.S. 398, 82 USPQ2d 1385 (2007): (D) Applying a known technique to a known device (method, or product) ready for improvement to yield predictable results. Here, Strader et al. represents a known device for summarizing a conversation between a patient and a healthcare provider that is ready for improvement according to recent technological developments in large language models that receive prompts to perform natural language processing tasks as taught by Mathur et al. Moreover, Strader et al. and Mathur et al., in combination, would yield predictable results because large language models are now being applied to a wide variety of tasks to obtain improvements according to techniques of machine learning,
Applicants’ arguments are moot in view of these new grounds of rejection. This Office Action is NON-FINAL.
Conclusion
The prior art made of record and not relied upon is considered pertinent to Applicants’ disclosure.
Yan et al., Pauli et al., Hirshberg et al., and Pedersen et al. disclose related prior art.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MARTIN LERNER whose telephone number is (571) 272-7608. The examiner can normally be reached Monday-Thursday 8:30 AM-6:00 PM.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Richemond Dorvil can be reached at (571) 272-7602. 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.
/MARTIN LERNER/Primary Examiner
Art Unit 2658
February 23, 2026