DETAILED ACTION
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 .
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 11/11/2025 has been entered. Claims 1,4-5,9-10,13-14,18-19,21-24 are pending. Claims 2-3,6-8,11-12,15-17 and 20 have been cancelled.
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 1,4-5,9-10,13-14,18-19,21-24 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (i.e., a law of nature, a natural phenomenon, or an abstract idea) without significantly more.
Step 1: Is the claimed invention a statutory category of invention?
Claims 1, 10 and 19 are directed to a drug knowledge question answering method / electronic device / computer instructions (Step 1, Yes).
Step 2A, Prong 1: Does the claim recite an abstract idea?
The limitation of steps: … obtaining a drug question text input by a user through an interactive apparatus of the electronic device; performing retrieval in predefined question-answer text pairs in a drug knowledge base based on the drug question text, to obtain answer text corresponding to a drug question text of a corresponding question-answer text pair that matches the drug question text; performing retrieval in drug instructions in the drug knowledge base based on the drug question text, to obtain instruction text corresponding to the drug question text, comprising: obtaining a drug instruction with the highest similarity by calculating a cosine similarity between a vector of the drug question text and a vector of the drug instructions in the drug knowledge base; performing intent recognition on the drug question text through a trained intent classification model, to obtain an intent recognition result; and extracting the instruction text from the drug instruction with the highest similarity based on the intent recognition result; and generating response text corresponding to the drug question text based on the answer text and the instruction text, comprising: combining the answer text and the instruction text to obtain evidence text; constructing prompt text based on the drug question text, the evidence text, and a text form requirement; and generating the response text corresponding to the drug question text based on an output of the large language model by inputting the prompt text into the large language modelas drafted, is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components. This type of mental process can be practically performed in the human mind, for instance by a pharmacist or a physician. This akin to the abstract idea of performing mental observations of the drug question, evaluation question – answer text in a drug knowledge base, and providing response text). The mere nominal recitation of at least one processor performing these steps does not take the claim limitation outside of the mental processes grouping. Thus, the claim recites a mental process (Step 2A, Prong 1: yes).
Step 2A, Prong 2: Does the claim recite additional elements that integrate the judicial exception into a practical application?
Per the 2019 Revised Patent Subject Matter Eligibility Guidance, if a claim as a whole integrates the recited judicial exception into a practical application of that exception, a claim is not "directed to" a judicial exception. Alternatively, a claim that does not integrate a recited judicial exception into a practical application is directed to the exception. Evaluating whether a claim integrates an abstract idea into a practical application is performed by a) identifying whether there are any additional elements recited in the claim beyond the abstract idea, and b) evaluating those additional elements individual and in combination to determine whether they integrate the abstract idea into a practical application, using one or more of the considerations laid out by the Supreme Court and the Federal Circuit. Exemplary considerations indicative that an additional element (or combination of elements) may have or has not been integrated into a practical application are set forth in the 2019 PEG.
With respect to the instant claims, Claim 1 requires “a computer”. Claim 10 recites “An electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor”. Claim 19 recites “A non-transitory computer-readable storage medium” and “computer”. It is particularly noted that the use of a computing device "as a tool" to perform an abstract method and steps that only amount to extra solution activity are indicated in the 2019 PEG as examples that an additional element has not been integrated into a practical application. Even in combination, the recited additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits, such as an improvement to a computing system, on practicing the abstract idea (STEP 2A, Prong 2: NO).
Step 2B: Does the claim recite additional elements that amount to significantly more than the judicial exception?
Claim 1 requires “a computer”. Claim 10 recites “An electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor”. Claim 19 recites “A non-transitory computer-readable storage medium” and “computer” set forth above for Step 2A, Prong 2. Regarding these limitations: Applicant's specification only describes these features in a highly generic manner by stating that "The computing unit 1201 may be various general purpose and/or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 1201 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units that run machine learning objective function algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc." in the Applicant’s published application, para. [0142]. Examiner notes the lack of any description of specific algorithms of programming used by the processing components to perform the claimed steps. There is no indication in the Specification that Applicants have achieved an advancement or improvement in computer, machine learning and language processing technology. Dependent claims 4-5,9,13-14,18,21-24 inherit the deficiencies of their respective parent claims through their dependencies and do not recite additional limitations sufficient to direct the claims to more than the claimed abstract idea, and are thus rejected for the same reasons.
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1, 4, 9 – 10, 13 and 18 – 19, 21 – 24 are rejected under 35 U.S.C. 103 as being unpatentable over Alkhalifa et al. (US 2024/0419908 A1) in view of Nguyen et al. (US 2025/0139377 A1) and Iwamasa (US 2025/0094730 A1).
Re claims 1, 10, 19:
1. A drug knowledge question answering method (Alkhalifa, Abstract; [0004]), the method being implemented by a computer and (Alkhalifa, fig. 1, 102) comprising:
obtaining a drug question text input by a user through an interactive apparatus of the computer (Alkhalifa, [0004]; [0039], “The regulatory question may be a question that was entered by a user in a user interface (e.g., via display device 164 and user input device 166”);
performing retrieval predefined question-answer text pairs in a drug knowledge base (Alkhalifa, [0025], “the database 126 contains data that may be used to train machine learning models (e.g., the NLP models 130 discussed below), as well as an archive of past regulatory questions and their answers”; [0032], “extracting portions of the similar documents (e.g., portions of actual answers to past regulatory questions identified by similarity unit 142B)”; [0010], “answer generation unit may identify relevant historical answers by first identifying similar questions”) based on the drug question text, to obtain an answer text corresponding to a question text of a corresponding question-answer text pair that matches the drug question text (Alkhalifa, [0004]; fig. 7B; [0006]; fig. 10; [0021], “method for generating potential answers to a regulatory question”; [0032], “The answer generation unit”);
performing retrieval in drug instructions in the drug knowledge base based on the drug question text, to obtain instruction text corresponding to the drug question text (Alkhalifa, [0004]; fig. 7B; [0006]; fig. 10; [0021], “method for generating potential answers to a regulatory question”; [0032], “The answer generation unit”), comprising:
obtaining a drug instruction with the highest similarity by calculating a cosine similarity between a vector of the drug question text and a vector of the drug instructions in the drug knowledge base (Alkhalifa, [0004]; fig. 7B; [0010]; [0031], “The similarity unit 142B … most similar”);
performing intent recognition on the drug question text through a trained intent classification model, to obtain an intent recognition result (Alkhalifa, [0006], “use an NLP model to classify each question into a category that helps users identify who is best suited to provide an answer”; [0034], “the NLP models 130 includes multiple NLP classification models each specialized to determine whether textual data corresponding to a particular question should, or should not, be classified as belonging to a single, respective category (e.g., with one of NLP models 130 determining whether to classify as “Safety,””); and
extracting the instruction text from the drug instruction with the highest similarity based on the intent recognition result (Alkhalifa, [0004]; fig. 7B; [0010]; [0031], “The similarity unit 142B … most similar”).
Alkahlifa does not explicitly disclose generating response text corresponding to the drug question text based on the answer text and the instruction text.
Iwamasa (US 2025/0094730 A1) teaches an information processing apparatus comprising a processing circuitry configured to: select, based on a content of an input question, a database corresponding to the question among a plurality of databases respectively storing different kinds of data; generate a prompt to be input to a language model based on the selected database and the question; and generate an answer based on the prompt and the language model.
Iwamasa teaches the limitation:
obtaining a question text input by a user through an interactive apparatus of the computer (Iwamasa, fig. 1, 100);
generating response text corresponding to the drug question text based on the answer text and the instruction text (Iwamasa, [0015], “a prompt processor”; [0018], “The prompt processor 130 includes a prompt generator 131 and a DB selector 132”), comprising:
combining the answer text and the instruction text to obtain evidence text (Iwamasa, fig. 7, fig. 3; [0029], “When processing of the first partial question is completed (an answer to the first partial question is obtained) through processing at the prompt processor 130 or later, the next partial question is taken out of the partial question list and sent to the prompt processor 130”);
constructing prompt text based on the drug question text, the evidence text, and a text form requirement (Iwamasa, fig. 7, S104 - “Generate Prompt”; [0019], “The data to be sent to the question decomposer 110 may be text data, encrypted data, or data in any other format”); and
generating the response text corresponding to the drug question text based on an output of the large language model by inputting the prompt text into the large language model (Iwamasa, [0046], “The prompt generator 131 generates a prompt to be provided to the LLM executer 160 based on the DB (referred to as a selected DB) selected by the DB selector 132”; [0054], “in a case where the content of the answer is not limited to cooling, the prompt corrector 180 instructs the prompt generator 131 to add, to the prompt text, a description clearly indicating that the answer is to be narrowed down to cooling”).
Therefore, in view of Iwamasa, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the method/system described in Alkhalifa, by generating prompt for llm as taught by Iwamasa, since Iwasmasa states that “when the prompt is processed by the LLM executer, the range of processed data can be narrowed down to an appropriate range and thus a highly accurate answer is highly likely to be obtained” (Iwasmasa, [0046]).
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the method/system described in Alkhalifa, by providing cosine similiarity as taught by Nguyen, since the search engine may calculate the cosine similarity between the query vector and the feature vector in the feature database as a similarity degree, select the feature vector having the largest similarity degree as a similar feature vector, and acquire the knowledge information corresponding to the selected feature vector. A feature vector is extracted for each document that is knowledge information (Nguyen, [0061]).
Alkahlifa does not explicitly disclose cosine similarity. Nguyen teaches the missing feature (Nguyen, [0041]). It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the method/system described in Alkhalifa, by providing cosine similiarity as taught by Nguyen, since the search engine may calculate the cosine similarity between the query vector and the feature vector in the feature database as a similarity degree, select the feature vector having the largest similarity degree as a similar feature vector, and acquire the knowledge information corresponding to the selected feature vector. A feature vector is extracted for each document that is knowledge information (Nguyen, [0061]).
10. An electronic device, comprising:
at least one processor; and
a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions, when executed by the at least one processor, cause the at least one processor to perform the following steps:
obtaining a drug question text input by a user through an interactive apparatus of the electronic device;
performing retrieval in predefined question-answer text pairs in a drug knowledge base based on the drug question text, to obtain answer text corresponding to a drug question text of a corresponding question-answer text pair that matches the drug question text;
performing retrieval in drug instructions in the drug knowledge base based on the drug question text, to obtain instruction text corresponding to the drug question text, comprising:
obtaining a drug instruction with the highest similarity by calculating a cosine similarity between a vector of the drug question text and a vector of the drug instructions in the drug knowledge base;
performing intent recognition on the drug question text through a trained intent classification model, to obtain an intent recognition result; and
extracting the instruction text from the drug instruction with the highest similarity based on the intent recognition result; and
generating response text corresponding to the drug question text based on the answer text and the instruction text, comprising:
combining the answer text and the instruction text to obtain evidence text;
constructing prompt text based on the drug question text, the evidence text, and a text form requirement; and
generating the response text corresponding to the drug question text based on an output of the large language model by inputting the prompt text into the large language model (See claim 1 rejection above).
19. A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to perform the following steps:
obtaining a drug question text input by a user through an interactive apparatus of the computer;
performing retrieval in predefined question-answer text pairs in a drug knowledge base based on the drug question text, to obtain an answer text corresponding to a question text of a corresponding question-answer text pair that matches the drug question text;
performing retrieval in drug instructions in the drug knowledge base based on the drug question text, to obtain instruction text corresponding to the drug question text, comprising:
obtaining a drug instruction with the highest similarity by calculating a cosine similarity between a vector of the drug question text and a vector of the drug instructions in the drug knowledge base;
performing intent recognition on the drug question text through a trained intent classification model, to obtain an intent recognition result; and
extracting the instruction text from the drug instruction with the highest similarity based on the intent recognition result: and
generating response text corresponding to the drug question text based on the answer text and the instruction text, comprising:
combining the answer text and the instruction text to obtain evidence text;
constructing prompt text based on the drug question text, the evidence text, and a text form requirement; and
generating the response text corresponding to the drug question text based on an output of the large language model by inputting the prompt text into the large language model (See claim 1 rejection above).
Re claims 4, 13:
4. The method according to claim 1, wherein the performing retrieval in drug instructions in the drug knowledge base based on the drug question text, to obtain instruction text corresponding to the drug question text comprises:
performing entity recognition on the drug question text, to obtain a drug name corresponding to the drug question text; performing retrieval in the drug knowledge base based on the drug name, to obtain a drug instruction corresponding to the drug question text; and extracting, from the drug instruction, the instruction text corresponding to the drug question text (Alkhalifa, [0028], “identify other documents (e.g., other regulatory questions) that are similar to the questions under consideration, generate answers to the questions under consideration, and/or summarize the questions under consideration”; [0004], “may refer to either an explicit question (e.g., "What is the maximum dosage of Drug X?") or an implicit question or prompt ( e.g., describing a potential problem with the administration of Drug X”; pg. 8, Table 1, “The statement with correct amount of sterile water for reconstitution should be displayed on the side display panel. After reconstitution with 5 mL of Sterile Water for Injection, the concentration of molecule is 2%. The correct amount of Sterile Water for Injection is indicated in the Prescribing Information”; fig. 7B, “How many minutes does the UPLC dissolution method take?”).
13. The electronic device according to claim 10, wherein the performing retrieval in drug instructions in the drug knowledge base based on the drug question text, to obtain instruction text corresponding to the drug question text comprises:
performing entity recognition on the drug question text, to obtain a drug name corresponding to the drug question text; performing retrieval in the drug knowledge base based on the drug name, to obtain a drug instruction corresponding to the drug question text; and extracting, from the drug instruction, the instruction text corresponding to the drug question text (Alkhalifa, [0028], “identify other documents (e.g., other regulatory questions) that are similar to the questions under consideration, generate answers to the questions under consideration, and/or summarize the questions under consideration”; [0004], “may refer to either an explicit question (e.g., "What is the maximum dosage of Drug X?") or an implicit question or prompt ( e.g., describing a potential problem with the administration of Drug X”; pg. 8, Table 1, “The statement with correct amount of sterile water for reconstitution should be displayed on the side display panel. After reconstitution with 5 mL of Sterile Water for Injection, the concentration of molecule is 2%. The correct amount of Sterile Water for Injection is indicated in the Prescribing Information”; fig. 7B, “How many minutes does the UPLC dissolution method take?”).
Re claims 6, 15:
6. The method according to claim 4, wherein the extracting, from the drug instruction, the instruction text corresponding to the drug question text comprises: performing intent recognition on the drug question text, to obtain an intent recognition result; and extracting the instruction text from the drug instruction based on the intent recognition result (Alkhalifa, [0004], “maximum dosage of Drug X … Condition Z … “; [0056], “the regulatory questions pertain to pharmaceuticals ( e.g., usage, risks, etc.)”; intent recognition – dosage, condition, usage, risk..etc.).
15. The electronic device according to claim 13, wherein the extracting, from the drug instruction, the instruction text corresponding to the drug question text comprises: performing intent recognition on the drug question text, to obtain an intent recognition result; and extracting the instruction text from the drug instruction based on the intent recognition result (Alkhalifa, [0004], “maximum dosage of Drug X … Condition Z … “; [0056], “the regulatory questions pertain to pharmaceuticals ( e.g., usage, risks, etc.)”; intent recognition – dosage, condition, usage, risk..etc.).
Re claims 7, 16:
7. The method according to claim 6, wherein
the performing retrieval in the drug knowledge base based on the drug name, to obtain a drug instruction corresponding to the drug question text (Alkhalifa, [0004]) comprises: calculating similarity based on the drug name corresponding to the drug question text and each of the drug instructions in the drug knowledge base, to obtain the drug instruction with the highest similarity (Alkhalifa, [0004]; fig. 7B; [0010]; [0031], “The similarity unit 142B … most similar”); and
the extracting, from the drug instruction, the instruction text corresponding to the drug question text comprises: extracting the instruction text from the drug instruction with the highest similarity based on the intent recognition result (Alkhalifa, [0031] – [0032]; [0004]; [0056]).
16. The electronic device according to claim 15, wherein
the performing retrieval in the drug knowledge base based on the drug name, to obtain a drug instruction corresponding to the drug question text Alkhalifa, [0004]) comprises: calculating similarity based on the drug name corresponding to the drug question text and each of the drug instructions in the drug knowledge base, to obtain the drug instruction with the highest similarity (Alkhalifa, [0004]; fig. 7B; [0010]; [0031], “The similarity unit 142B … most similar”); and
the extracting, from the drug instruction, the instruction text corresponding to the drug question text comprises: extracting the instruction text from the drug instruction with the highest similarity based on the intent recognition result (Alkhalifa, [0004]; fig. 7B; [0010]; [0031], “The similarity unit 142B”).
Re claims 9, 18:
9. The method according to claim 1, wherein before the performing retrieval in question-answer text in a drug knowledge base based on the drug question text, the method further comprises: parsing question answering data into the question-answer text, and storing the question-answer text in the drug knowledge base. 18. The electronic device according to claim 10, wherein before the performing retrieval in question-answer text in a drug knowledge base based on the drug question text, the following step is further comprised: parsing question answering data into the question-answer text, and storing the question-answer text in the drug knowledge base (Alkhalifa, [0044], “pre-processing unit 140 parses the text into its constituent questions. The pre-processing unit 140 may parse the text into questions using known delimiters or fields in data files that contain the text”; [0046], “the pre-processing unit 140 tokenizes the text of the questions (e.g., parses each question into individual words or other linguistic units)”; [0048]; [0032], “The answer generation unit 142C generally applies one or more of the NLP models 130 to the textual data ( or to pre-processed textual data) in order to generate one or more potential answers to a particular regulatory question”).
Re claims 21 – 24:
21. The method according to claim 1, wherein before the performing retrieval in question-answer text in a drug knowledge base based on the drug question text, the method further comprises: rephrasing the drug question text. 22. The method according to claim 21, wherein the rephrasing the drug question text comprises: obtaining historical dialog information related to the drug question text; and rephrasing the drug question text based on drug information in the historical dialog information. 23. The electronic device according to claim 10, wherein before the performing retrieval in question-answer text in a drug knowledge base based on the drug question text, the following step is further comprised: rephrasing the drug question text. 24. The electronic device according to claim 23, wherein the rephrasing the drug question text comprises: obtaining historical dialog information related to the drug question text; and rephrasing the drug question text based on drug information in the historical dialog information (Alkhalifa, fig. 7A, “Similarly Ask Questions”; [0045], “cleans the text of the questions by removing words and/or characters that are irrelevant”; [0007] – [0008]; [0010], “answer generation unit may identify relevant historical answers by first identifying similar questions”; fig. 7B; fig. 7C; [0004]; [0042]).
Claims 5 and 14 are rejected under 35 U.S.C. 103 as being unpatentable over Alkhalifa, Nguyen and Iwamasa as applied to claims 5 and 14 above, and further in view of Arat et al. (US 2020/0320365 A1).
Re claim 5, 14:
Alkhalifa does not explicitly disclose scanning a two-dimensional barcode corresponding to a drug package, or performing text recognition based on an obtained image of the drug package, to obtain the drug name corresponding to the drug question text.
Arat teaches an invention related to facilitate interactions between two individuals (e.g., a customer support agent and a consumer) (Arat, Abstract). Arat further teaches 5. The method according to claim 4, wherein the performing entity recognition on the drug question text, to obtain a drug name corresponding to the drug question text comprises: scanning a two-dimensional barcode corresponding to a drug package, or performing text recognition based on an obtained image of the drug package, to obtain the drug name corresponding to the drug question text. 14. The electronic device according to claim 13, wherein the performing entity recognition on the drug question text, to obtain a drug name corresponding to the drug question text comprises: scanning a two-dimensional barcode corresponding to a drug package, or performing text recognition based on an obtained image of the drug package, to obtain the drug name corresponding to the drug question text (Arat, [0232] – [0235], “received data identifying prescription bottle 1100 that has been detected via computing device 1102, back-end platform 102 may determine prescription information associated with prescription bottle 1100, which may involve extracting the prescription information on the prescription label of prescription bottle 1100 using various image processing technique … prescription information on a given prescription label may comprise drug specific information. Drug specific information may generally include information about the prescription drug, such as a name of the drug, a dosage or strength of the drug, a manufacturer of the drug, a description of the drug, directions on how and when to take the drug, a quantity of the drug prescribed, an expiration date, and/or a date in which the drug was filled, a number of times the drug can be refilled, as some non-limiting examples”; [0240]).
Therefore, in view of Arat, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the method / electronic device described in Alkhalifa, by providing image processing technique as taught by Arat, since prescription information on a given prescription label includes comprehensive information such as: a name of the drug, a dosage or strength of the drug, a manufacturer of the drug, a description of the drug, directions on how and when to take the drug, a quantity of the drug prescribed, an expiration date, and/or a date in which the drug was filled, a number of times the drug can be refilled. The image processing technique allows a user quickly extract information in an efficient metho without manual entry.
Response to Arguments
Applicant's arguments filed 11/11/2025 have been fully considered but they are not persuasive.
Regarding the rejection under 35 U.S.C. §101
Applicant argues … calculate the cosine similarity between a vector representing a drug question text and a vector representing drug instructions, the vectors themselves must first be obtained. This requires processing the respective textual content using a natural language model capable of converting textual data into numerical vector representations. Such conversion cannot be performed mentally by a human being. A person cannot transform textual information into high-dimensional numerical vectors without a computational model
The examiner submits that the method of calculating a cosine similarity is merely a mathematical equation capable of being performed by a human mind. Second, the step for generating a rephrase drug question text based on the algorithm (LLM). Such methods can and have been performed by the mind of a pharmacist. There is no requirement to the size of the large language model (LLM) and drug knowledge base. None of the steps require any type of complexity; the amount of data; nor require any precision on the calculation of the cosine similarity and rephrasing a drug question (summarizing drug question) in the claims; a human can perform the steps and analysis in his/her mind by using pen and paper. The mere citation of calculating a cosine similarity between vectors and generating a rephrase drug question text do not preclude the steps from being performed by a human.
Applicant further disagrees with the Examiner's statement that "a human can perform the steps and analysis in his/her mind by using pen and paper." (Office Action, p. 22-23) As understood by those skilled in the art, the vectors extracted from textual data are typically multi-dimensional and often comprise tens or even hundreds of components. Calculating the cosine similarity between such high-dimensional vectors involves numerous multiplications, additions, and square root operations, which cannot be reasonably performed manually. Therefore, these operations necessarily require the use of a computer or other processing device to perform the underlying mathematical computations.
The examiner respectfully submits that none of the limitation in the claims require any type of “high-dimensional vectors involves numerous multiplications, additions, and square root operations”.
Applicant argues: The USPTO guidance has explained that "claims do not recite a mental process when they contain limitations that cannot practically be performed in the human mind, for instance when the human mind is not equipped to perform the claim limitations." MPEP 2106.04(a)(2), subsection III.A (citing SRI Int'l, Inc. v. Cisco Sys., Inc., 930 F.3d 1295, 1304 (Fed. Cir. 2019) (declining to identify the claimed collection and analysis of network data as abstract because "the human mind is not equipped to detect suspicious activity by using network monitors and analyzing network packets as recited by the claims"); "
The Examiner respectfully disagrees. MPEP2106.04(a)(2) Ill MENTAL PROCESSES, with regard to computers clearly states the following: "Nor do the courts distinguish between claims that recite mental processes performed by humans and claims that recite mental processes performed on a computer. As the Federal Circuit has explained, "[c]ourts have examined claims that required the use of a computer and still found that the underlying, patent-ineligible invention could be performed via pen and paper or in a person's mind." Versata Dev. Group v. SAP Am., Inc., 793 F.3d 1306, 1335, 115 USPQ2d 1681, 1702 (Fed. Cir. 2015). See also Intellectual Ventures I LLC v. Symantec Corp., 838 F.3d 1307, 1318, 120 USPQ2d 1353, 1360 (Fed. Cir. 2016) ("[W]ith the exception of generic computer-implemented steps, there is nothing in the claims themselves that foreclose them from being performed by a human, mentally or with pen and paper."); Mortgage Grader, Inc. v. First Choice Loan Servs. Inc., 811 F.3d 1314, 1324, 117 USPQ2d 1693, 1699 (Fed. Cir. 2016) (holding that computer-implemented method for "anonymous loan shopping" was an abstract idea because it could be "performed by humans without a computer"). Here, as applied in view of MPEP2106.04(a)(2) Ill MENTAL PROCESSES, C. "A Claim That Requires a Computer May Still Recite a Mental Process," the Applicant's invention is merely (1) performing a mental process on a generic computer, as evidenced by the Applicant's written description of the specification in the published application in para. [0142]: " Some examples of the computing unit 1201 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units that run machine learning objective function algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc."; (2) performing a mental process in a computer environment, is analogous to how a specialist or pharmacist can perform the claimed steps. Per the Applicant's written description of the specification in the published application in para. [0038]: "first type of drug knowledge question answering system, an answer usually is manually given by a specialist or medical professional … Such a question answering system requires a number of specialists or medical professionals, leading to a high labor cost. The problem such as an inefficient or untimely artificial response results in a less responsive system, causing poor user experience". This clearly implies using the computer to be used to alleviate the burden of "high labor cost" and “inefficient or untimely”. Finally, SRI Int'l, Inc. v. Cisco Systems, Inc., 930 F.3d 1295, 1304 (Fed. Cir. 2019), relates to monitoring network data which on its face requires a computer. The Applicant's argument is unclear as to how it is on point with the Applicant's claimed subject matter, since the Applicant's "method and system" have been historically performed in the analog by pharmacists and doctors for decades. As such, the argument is not persuasive.
Applicant argues: For instance, the 2024 Guidance provides an example in MPEP 2106.04(a)(2), subsection III.A which claims a specific data encryption method for computer communication involving a several-step manipulation of data, Synopsys., 839 F.3d at 1148, 120 USPQ2d at 1481. Such an encryption method cannot be practically performed in the human mind and therefore does not fall under the category of a "mental process." Similarly, the method of amended Claim 1 in the present application performs data processing operations that go beyond what can be mentally performed by a human. Specifically, the method involves vectors and texts generated by a neural network (e.g., in a large language model). This type of data processing (e.g., cosine similarity calculation or generating a rephrased drug question text by a large language model) requires a level of computational precision, complexity, and scale that cannot be achieved mentally.
The examiner submits that the limitations generating a rephrased drug question text based on an output of a large language model by inputting the complete contextual dialog and a prompt requirement together into the large language model; performing retrieval in question-answer text in a drug knowledge base based on the rephrased drug question text, to obtain answer text corresponding to the drug question text; performing retrieval in drug instructions in the drug knowledge base based on the rephrased drug question text, to obtain instruction text corresponding to the drug question text, comprising: obtaining the instruction text corresponding to the drug question text by calculating a cosine similarity between a vector of the drug question text and a vector of the drug instructions in the drug knowledge base are not inherently complex. There is no requirement to the size of the large language model (LLM) and drug knowledge base; nor any encryption of the data in the claims. None of the steps require any type of complexity; the amount of data; nor require any precision on the calculation of the cosine similarity and rephrasing a drug question (summarizing drug question) in the claims; a human can perform the steps and analysis in his/her mind by using pen and paper.
Applicant argues: These additional elements are integrated into the overall method in a way that improves the functioning of the underlying hardware. Specifically, in response to user input, the computer-by leveraging a large language model-accurately interprets the user's retrieval intent and automatically generates a reply based on the answer text and instruction text. This allows a combination of the retrieval in the question-answer text and retrieval in drug instruction paragraphs, where the answer related to the user query may be quickly and accurately retrieved by the retrieval in the question-answer text, and when no accurate answer is found, the retrieval in the drug instruction paragraphs may be performed as a supplement, so as to ensure the recall of the question answering system, improving the answering accuracy and the recall of the system.
The examiner submits that the use of LLM and using computer to calculate cosine similarity between vectors, as recited in claim 1, would not necessarily improve accuracy in generating answer(s) to medical questions over a human doctors and pharmacists. Just like any artificial intelligent model, the answer(s) from AI like chatGPT does not necessary give the user an accurate answer, since LLMs may be trained on inaccurate and outdated databases. Answers generated by AI (LLM) models can be bias as well. There is not indication in the claim that answers from LLM is more accuracy than human specialists.
Second, simply making a process faster (more efficient) through the use of computer technology is not sufficient to establish patent eligibility. See Credit Acceptance Corp. v. Westlake Servs., 859 F.3d 1044, 1055 (Fed. Cir. 2017) ("[M]ere automation of manual processes using generic computers does not constitute a patentable improvement in computer technology."); Bancorp Servs. L.L.C. v. Sun Life Assurance Co. of Can. (US.), 687 F.3d 1266, 1279 (Fed. Cir. 2012) ("Using a computer to accelerate an ineligible mental process does not make that process patent-eligible.").
Applicant argues: the Applicant submit that neither Alkhalifa nor Nguyen discloses "predefined question-answer text pairs" and "obtaining an answer text corresponding to a question text of the corresponding question-answer text pair". In amended claim 1, a question-answer text pair is first determined based on the drug question text, and then an answer text is determined based on this question-answer text pair … Alkhalifa generates the answer by finding documents in database 126 that are similar to a particular regulatory question, rather than looking for a predetermined question-answer text pair.
The Examiner submits that Alkhalifa teaches a predefined question-answer text pair. For example, in para. [0025], a database contains an archive of past regulatory questions and their answers and in para. [0010], relevant historical answers by first identifying similar questions. There are past achieve / historical questions and their associated answers stored in the database, so that the obtained a drug question text input by a user can be compared with the achieved /historical questions.
Applicant argues: Applicant submits that neither Alkhalifa nor Nguyen discloses the limitation of "combining the answer text and the instruction text to obtain evidence text." The Examiner alleges that Nguyen discloses that "the answer output unit 57 may output, as the answer text, links to a plurality of pieces of similar knowledge information output by the knowledge database 60 in response to the question text and a fixed phrase" (Office Action, p. 10; see also Nguyen, para. [0064] and [0082]), and asserts that these passages teach the feature of "combining the answer text and the instruction text to obtain evidence text." However, Nguyen merely discloses combining two similar pieces of knowledge information to generate an answer text. Even if the Examiner interprets the "knowledge information" as the "instruction text" recited in amended claim 1, Nguyen at most discloses generating an answer text by combining two instruction texts. It does not disclose or suggest combining the answer text with another instruction text to obtain evidence text.
According to MPEP 2111 [R-5], during patent examination, the pending claims must be “given their broadest reasonable interpretation consistent with the specification.” The Federal Circuit’s en banc decision in Phillips v. AWH Corp., 415 F.3d 1303, 75 USPQ2d 1321 (Fed. Cir. 2005) expressly recognized that the USPTO employs the “broadest reasonable interpretation” standard. Applicant has not clearly defined what are “answer text”, “instruction text” and “evidence text”. The newly cited reference: Iwamasa (US 2025/0094730 A1) teaches limitations: generating the response text corresponding to the drug question text based on an output of the large language model by inputting the prompt text into the large language model.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JACK YIP whose telephone number is (571)270-5048. The examiner can normally be reached Monday thru Friday; 9:00 AM - 5:00 PM 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, XUAN THAI can be reached at (571) 272-7147. 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.
/JACK YIP/Primary Examiner, Art Unit 3715
Prosecution Insights