Notice of Pre-AIA or AIA Status
1. The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Arguments
2. Applicant's arguments filed 03/02/2026 have been fully considered but they are not persuasive. Applicant argues that Mielke does not disclose the features of claim 1, specifically: “obtaining, for each of the one or more services, a plurality of examples, each example providing an example command suitable for execution by the respective service; generating one or more instructions based on the user query, the one or more services, and the one or more pluralities of examples; providing the one or more instructions to a trained large language model ("LLM");”
The examiner respectfully disagrees. First applicant asserts that in paragraph 56; however, there is no discussion of any "example providing an example command suitable for execution by the respective service." Office Action, p. 3. Instead, paragraph 56 generally describes that the system may analyze the user input using natural language understanding techniques and that the system "may interact with different agents to obtain information or services that are associated with the resolved
entities." Mielke, ¶ 56. However, there is no discussion of any example commands, nor does it discuss obtaining any such examples. Moreover, paragraph 56 does not even mention the word "example." Therefore, Mielke does not disclose or make obvious "obtaining, for each of the one or more services, a plurality of examples, each example providing an example command suitable for execution by the respective service." However, Mielke does disclose at ¶[0251] that developers can annotate code with “assistant entity, assistant actions and assistant parameters” and in ¶[0252] that the system extracts the annotations into a structured representation. ¶[0253] discloses that once the action graph ontology is available, the system may easily synthesize utterance examples. Therefore, Mielke teaches obtaining service/domain specific prompting a language model with “a small set of example dialogues” as disclosed in ¶[0262].
Applicant further asserts Mielke also does not disclose or make obvious "generating one or more instructions based on the user query, the one or more services, and the one or more pluralities of examples." Here the Office Action cites to paragraph 77; however, paragraph 77 has no discussion of any instructions being generated, nor that any such instructions are generated "based on the user query, the one or more services, and the one or more pluralities of examples." Office Action, p. 3. Instead, this paragraph merely describes that actions may be selected and triggered. However, the actions do not involve generating any instructions "based on the user query, the one or more services, and the one or more pluralities of examples." Instead, Mielke describes that these actions are predetermined and may be selected; they are not generated. Mielke, ¶ 77 ("The action selector 222 a/b may determine and trigger a set of general executable actions."). Critically, as discussed above, Mielke does not even describe that examples may be obtained and used to "generate one or more instructions." Thus, Mielke cannot disclose or make obvious "generating one or more instructions based on the user query, the one or more services, and the one or more pluralities of examples." However, Mielke discloses at ¶[0251] that the system extracts these into a structured representation or manifests and ¶[0253] discloses that from the resulting action graph ontology the system may easily synthesize utterances examples. ¶[0262] further teaches prompting a large language model with a small set of example dialogues and ¶[0266] discloses prompting the model with tokenized example dialogues in conjunction with the current dialogue/query. Therefore, Mielke teaches the functional equivalent of "generating one or more instructions based on the user query, the one or more services, and the one or more pluralities of examples." Also See Fig. 3 block 380 “Response Generation.”
Accordingly, the rejection of claims 1-20 is maintained.
Claim Rejections - 35 USC § 102
3. 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 the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
4. Claims 1-20 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Mielke (US 2023/0135179).
Regarding Claim 1:
Mielke discloses a method comprising:
receiving, by an artificial intelligence ("Al") assistant, a user query comprising one or more tasks (Mielke: ¶[0041] discloses an assistant application interpreted as an AI assistant receiving a task or input from the user);
determining one or more services based on the user query (Mielke: ¶[0041] discloses receiving natural language input such as spoken or typed queries and determining; ¶[0047] discloses the system may include a third party content object provider which connects to various different services);
obtaining, for each of the one or more services, a plurality of examples, each example providing an example command suitable for execution by the respective service (Mielke: ¶[0251]-[0253] discloses that the developers annotate code with assistant action and assistant parameters, which are extracted into a structured representation and from the resulting action graph ontology the system may then synthesize utterance examples, these examples are tied to corresponding actions and services and therefore teach examples suitable for execution by the respective service);
generating one or more instructions based on the user query, the one or more services, and the one or more pluralities of examples (Mielke: ¶[0262] discloses prompting a language model with a small set of example dialogues and a user turn corresponding to the user query and ¶[0266] discloses prompting the model with tokenized example dialogues, therefore Mielke teaches constructing input to the language model based on the user query together with service and domain specific examples, which corresponds to generating as claimed);
providing the one or more instructions to a trained large language model ("LLM") (Mielke: ¶[0056] the NLU/NLG stems are machine learning based and represent a trained large language model processing the instructions and generating outputs);
receiving, from the LLM, one or more commands corresponding to the user query (Mielke: ¶[0056] generated output here corresponds to commands ready for execution by services);
for each command of the one or more commands, issuing the respective command to a corresponding service of the one or more services (Mielke: ¶[0077] each generated command is given to a corresponding service i.e., it issues the command to the service);
generating a response to the user query based on results of the one or more commands (Mielke: ¶[0081] discloses forming a final response);
and outputting the response (Mielke: ¶[0081] provides the generated response to the user).
Regarding Claim 2:
Mielke further discloses the method of claim 1, further comprising:
generating, using a trained machine learning (“ML”) model, one or more first embeddings based on the user query (Mielke: ¶[0056] establishes that a trained model is used to analyze user inputs; ¶[0071] explicitly references task-specific user representations (first embeddings) generated by the system, confirming that embeddings are created using a trained ML model);
generating, using the trained ML model, one or more second embeddings based on descriptions of the one or more services (Mielke: ¶[0071] discloses federated user representation which shows embeddings being created for different tasks or entities; ¶[0077] discloses that services have structured templates and descriptions, which are suitable inputs for generating second embeddings corresponding to those services, e.g., second embeddings);
and wherein determining the one or more services is based on the one or more first embeddings and the one or more second embeddings (Mielke: ¶[0074] discloses the system resolves user input into candidate tasks (services) by comparing user data and service/task specifications, this is consistent with matching user query embeddings AND service embeddings to select which services to invoke).
Regarding Claim 3:
Mielke further discloses the method of claim 2, further comprising:
determining, for each service, a confidence based on the one or more first embeddings and one or more second embeddings corresponding to the respective service (Mielke: ¶[0170], ¶[0186] discloses explicitly using model embeddings and internal representations to generate a confidence score (probability of correctness));
and determining the one or more services based on the respective confidences and a confidence threshold (Mielke: ¶[0191] explicitly teaches using thresholds to determine outcomes based on the predicted probabilities (confidence levels)).
Regarding Claim 4:
Mielke further discloses the method of claim 2, further comprising:
determining, for each service, a confidence based on the one or more first embeddings and one or more second embeddings corresponding to the respective service (Mielke: ¶[0170], ¶[0186] discloses explicitly using model embeddings and internal representations to generate a confidence score (probability of correctness));
and wherein obtaining, for each of the one or more services, the plurality of examples is based on the respective confidence for the respective service (Mielke: ¶[0168], ¶[0191] discloses using the confidence probabilities to determine how well the system selects how outputs are generated and how training examples are used to further finetune itself).
Regarding Claim 5:
Mielke further discloses the method of claim 1, further comprising:
for each example, generating a relevancy based on the user query (Mielke: ¶[0170] discloses a process where each candidate example (answer) is compared to the user query (question) using model embeddings and internal representations, ¶[0186] discloses the performance is further evaluated for correctness/relevancy);
and wherein obtaining, for each of the one or more services, the plurality of examples is based on the respective relevancies (Mielke: ¶[0168] discloses using the probabilities (relevancies) to influence how outputs are chose and adjusted; ¶[0191] discloses selecting or filtering outputs based on those calculated probabilities).
Regarding Claim 6:
Mielke further discloses the method of claim 5, further comprising:
generating, using a trained ML model, one or more first embeddings based on the user query (Mielke: ¶[0071] discloses generating task-specific embeddings; ¶[0170] discloses using model representations for the user query (question) which aligns with creating first embeddings from the query);
generating, using the trained ML model, one or more second embeddings based on the pluralities of examples (Mielke: ¶[0168]-[0170] the answer within this portion corresponds to the example in the claim);
and wherein generating the relevancy is based on at least a subset of the one or more first embeddings and the respective one or more second embeddings associated with the respective example (Mielke: ¶[0170] discloses that the calibrator takes both the query (first embedding) and example (second embeddings) as input to determine a probability of correctness, which is the relevancy).
Regarding Claim 7:
Mielke further discloses the method of claim 1, wherein the issuing the respective command is performed by the Al assistant (Mielke: ¶[0077] discloses how the assistance system itself routes and issues commands to the appropriate service agents for execution).
Regarding Claim 8:
The method of claim 1, wherein the issuing the respective command is performed by the trained LLM (Mielke: ¶[0077] ¶[0056] establishes the assistant uses machine learning-based language models to interpret input ang generate appropriate output; ¶[0077] shows that the generated outputs (commands) are issued to agents by the same modules).
Regarding Claim 9:
Claim 9 has been analyzed with regard to claim 1 (see rejection above) and is rejected for the same reasons of anticipation used above.
It is noted that Mielke discloses a system comprising a non-transitory computer readable storage media at least at paragraphs [0354]-[0359].
Regarding Claim 10:
Claim 10 has been analyzed with regard to claim 2 (see rejection above) and is rejected for the same reasons of anticipation used above.
Regarding Claim 11:
Claim 11 has been analyzed with regard to claim 3 (see rejection above) and is rejected for the same reasons of anticipation used above.
Regarding Claim 12:
Claim 12 has been analyzed with regard to claim 4 (see rejection above) and is rejected for the same reasons of anticipation used above.
Regarding Claim 13:
Claim 13 has been analyzed with regard to claim 5 (see rejection above) and is rejected for the same reasons of anticipation used above.
Regarding Claim 14:
Claim 14 has been analyzed with regard to claim 6 (see rejection above) and is rejected for the same reasons of anticipation used above.
Regarding Claim 15:
Claim 15 has been analyzed with regard to claim 7 (see rejection above) and is rejected for the same reasons of anticipation used above.
Regarding Claim 16:
Claim 16 has been analyzed with regard to claim 8 (see rejection above) and is rejected for the same reasons of anticipation used above.
Regarding Claim 17:
Claim 17 has been analyzed with regard to claim 1 (see rejection above) and is rejected for the same reasons of anticipation used above.
It is noted that Mielke discloses a system comprising a non-transitory computer readable storage media at least at paragraphs [0354]-[0359].
Regarding Claim 18:
Claim 18 has been analyzed with regard to claim 2 (see rejection above) and is rejected for the same reasons of anticipation used above.
Regarding Claim 19:
Claim 19 has been analyzed with regard to claim 3 (see rejection above) and is rejected for the same reasons of anticipation used above.
Regarding Claim 20:
Claim 20 has been analyzed with regard to claim 5 (see rejection above) and is rejected for the same reasons of anticipation used above.
Conclusion
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to IAN SCOTT MCLEAN whose telephone number is (703)756-4599. The examiner can normally be reached "Monday - Friday 8:00-5:00 EST, off Every 2nd Friday".
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, Hai Phan can be reached at (571) 272-6338. 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.
/IAN SCOTT MCLEAN/Examiner, Art Unit 2654
/HAI PHAN/Supervisory Patent Examiner, Art Unit 2654
Prosecution Insights