DETAILED ACTION
This communication is in response to the Amendments and Arguments filed on June 15, 2026. Claims 1-3, 6-20, and 23-34 are pending and have been examined. Hence, this action has been made FINAL.
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Arguments
The reply filed on June 15, 2026 has been entered. Applicant’s arguments with respect to claims 1-3, 6-20, and 23-34 have been considered but are not persuasive.
With respect to the applicant’s arguments to claim objections because of informalities, Applicant asserts that “The Examiner's suggestion to change 'maps' to 'mapped' would convert the verb to past tense or a past participle, which would actually create a grammatical issue because the clause would then lack a proper finite verb.” The examiner agrees with these assertions and has thus withdrawn the claim objections to claims 7 and 24. Further, the amendments to claims 9, 10, 26, and 27 have been noted by the examiner and thus, claim objections to each of the respective claims has been withdrawn. However, as noted in further detail below, the amendments to the independent claims has raised new minor informalities, necessitating claim objections to claims 2, 6-7, 9-11, 13, 19, 23-24, 26-28, and 30.
With respect to the applicant’s arguments to claim rejections under 35 U.S.C § 101, Applicant has amended each of the independent claims and asserts that “A human cannot detect hotwords in streaming audio captured by a microphone-this requires specialized audio processing hardware and software that continuously monitors audio signals and identifies specific acoustic patterns corresponding to predetermined hotwords.” The examiner respectfully disagrees with these assertions. As expanded upon further below with respect to claim rejections under 35 U.S.C. § 101, the limitations of a “microphone” and “data processing hardware” are recited at a high level of generality (¶ [0028] and ¶ [0077], respectively) and merely equate to “apply it” or otherwise merely uses a generic computer as a tool to perform an abstract which are not indicative of integration into a practical application as per MPEP 2106.05(f).
Applicant further asserts that “The system must maintain a predefined mapping between specific acoustic patterns (hotwords) and corresponding LLM adaptation configurations - a technical architecture that cannot be replicated mentally.” The examiner respectfully disagrees with these assertions. As expanded upon further below with respect to claim rejections under 35 U.S.C. § 101, the limitation of an “LLM” is recited at a high level of generality (¶ [0037]) and merely equates to “apply it” or otherwise merely uses a generic computer as a tool to perform an abstract which are not indicative of integration into a practical application as per MPEP 2106.05(f). In other words, a human could perform the same tasks as the LLM in the specific scenario; that is, a human could assign different spoken utterances (e.g. hotwords) to different functionalities or tasks (e.g. the hotword “play music” could indicate to a human that the user wishes to play music, while the hotword “open page” could indicate to a human that the user wishes to view a webpage).
Applicant further asserts that “The Examiner's analogy of a human manager formulating a task list for an employee fundamentally mischaracterizes the technical nature of LLM adaptation, which involves computational operations on neural network parameters or embeddings.” The examiner respectfully disagrees with these assertions. The applicant alludes to “computational operations on neural network parameters or embeddings” relating to LLM adaptation, but no such language can be found within the claim limitations. Merely “obtaining an adaptation input” that is “formulated for adapting the assistant LLM” cannot be construed as technically adapting an LLM’s parameters or embeddings because no such steps are recited within the claim language that would support such an assertion.
Applicant further asserts that “LLM inference involves massive matrix multiplications across many neural network layers and cannot be performed in the human mind. The Examiner's analogy to a human employee performing tasks ignores the computational nature of LLM processing, which requires specialized hardware to execute billions of mathematical operations.” The examiner respectfully disagrees with these assertions. Similar to above, the limitation of an “LLM” is recited at a high level of generality (¶ [0037]) and merely equates to “apply it” or otherwise merely uses a generic computer as a tool to perform an abstract which are not indicative of integration into a practical application as per MPEP 2106.05(f). The assertion that LLM inference involves massive matrix multiplications across many neural network layers cannot be considered valid justification for overcoming claim rejections under 35 U.S.C. § 101, as no such language exists within the claim limitations that would restrict the invention to those massive matrix multiplications. See MPEP 2173.05(q), “a claim which read: "[a] process for using monoclonal antibodies of claim 4 to isolate and purify human fibroblast interferon" was held to be indefinite because it merely recites a use without any active, positive steps delimiting how this use is actually practiced. ... “Use” claims that do not purport to claim a process, machine, manufacture, or composition of matter fail to comply with 35 U.S.C. 101.” As amended, there is no language in the independent claims that would prevent a human from performing these steps, as addressed in further detail below with respect to claim rejections under 35 USC § 101.
With respect to the applicant’s arguments to claim rejections under 35 U.S.C § 102 and 103, the applicant’s arguments with respect to claims 1-3, 6-20, and 23-34 have been considered but are not persuasive. The applicant asserts that “Claims 4-5 and 21-22 are rejected under 35 U.S.C. §103 as being obvious over Xie in view of U.S. Patent Application No. 11,100,922 to Mutagi et al. ("Mutagi"). This rejection is respectfully traversed.” However, the applicant fails to provide arguments or support for why the rejection is traversed. Further, the applicant fails to explain why Mutagi et al. cannot teach the claim limitations of amended independent claims 1 and 18. Thus, as described further below with respect to claim rejections under 35 U.S.C. § 103, the examiner maintains that Xie et al. in view of Mutagi et al. discloses all of the claim limitations of independent claims 1 and 18.
Claim Objections
Claims 2, 6-7, 9-11, 13, 19, 23-24, 26-28, and 30 are objected to because of the following informalities:
Claim 2, lines 1-2, should be “the particular hotword”
Claims 6-7, 9-11, 13, 19, 23-24, 26-28, and 30 are rejected for similar language to claim 2.
Appropriate correction is required.
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 1-3, 10, 12-14, 16-20, 27, 29-31, and 33-34 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. All of the claims are method claims (1-3, 9-14, 16-17), apparatus/machine claims (18-20, 27, 29-31, 33-34) or manufacture claim under (Step 1), but under Step 2A all of these claims recite abstract ideas and specifically mental processes. These mental processes are more particularly recited in claims 1 and 18 as:
receiving, from a user, a hotword detection event indication indicating detection of a particular hotword in streaming audio captured by a microphone in communication with the data processing hardware…
based on the received hotword detection event indication indicating detection of the particular hotword, obtaining an adaptation input specifically formulated for adapting the assistant LLM to undertake the respective functionality specified by the particular hotword…
receiving, from the user, the follow-on query, the follow-on query comprising a natural language query specifying an action for the assistant LLM to perform…
providing, for input to the assistant LLM, the adaptation input…
processing, using the adapted assistant LLM undertaking the respective functionality specified by the particular hotword, the follow-on query to fulfill performance of the action specified by the natural language query…
Under Step 2A Prong One, claims 1 and 18 are directed to an abstract idea and specifically a mental process. As detailed above, the steps of generating, classifying, determining, identifying, updating, etc. may be practically performed in the human mind with the use of a physical aid such as a pen and paper. For example, a human manager could receive a hotword and natural language service request from a human client corresponding to to-be-serviced data, and based on the received hotword, formulate a task list for a human employee. Upon receiving the task list and the to-be-serviced data from their manager, the employee could perform the tasks on the to-be-serviced data, yielding serviced data, and then return the serviced data to their manager.
Under Step 2A Prong Two, this judicial exception is not integrated into a practical application because claims 1-3, 10, 12-14, 16-20, 27, 29-31, and 33-34 do not recite additional elements that integrate the exception into a practical application. In particular, claims 1 and 18 recite the additional elements of a microphone (¶ [0028]), data processing hardware (¶ [0077]), memory hardware (¶ [0071]), and an assistant large language model (LLM) (¶ [0037]). These additional elements are recited at a high level of generality and merely equate to “apply it” or otherwise merely uses a generic computer as a tool to perform an abstract which are not indicative of integration into a practical application as per MPEP 2106.05(f). Further, claims 1 and 18 recite the additional elements of “receiving…” which amounts to insignificant extra-solution activities which are not indicative of integration into a practical application as per MPEP 2106.05(g). Accordingly, these additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. The claim is directed to an abstract idea.
Under Step 2B, the claims do not recite additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional elements of using a computer is noted as a general computer {data processing hardware (¶ [0077]); memory hardware (¶ [0071])}. Mere instructions to apply an exception using a generic computer component cannot provide an inventive concept. Further, the additional limitations in the claims noted above are directed towards insignificant extra-solution activities. The claims are not patent eligible.
With respect to claims 2-3 and 19-20, the claim relates to receiving user input using a UI element. This limitation amount to insignificant pre-solution activities which are not indicative of integration into a practical application as per MPEP 2106.05(g). The additional limitation of a “UI element” is recited at a high level of generality (¶ [0031]) and merely equates to “apply it” or otherwise merely uses a generic computer as a tool to perform an abstract which are not indicative of integration into a practical application as per MPEP 2106.05(f). The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claims 10 and 27, the claim relates to identifying a matching prompt phrase, concatenating the prompt phrase to the second instruction, and then processing the concatenated query to fulfill a task. This relates to the manager brainstorming additional instructions to the client’s task request that would assist their employee towards a specific task result, concatenating the additional instructions to the task list, and sending the concatenated instructions to their employee. . No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claims 12 and 29, the claim relates to processing the adaptation input before processing the follow-on query. This relates to the manager writing the task list before receiving the to-be-serviced data from the client. No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claims 13 and 30, the claim relates to retrieving content that was previously used by the LLM for inferencing. This relates to the manager recalling the client’s previous service requests and to-be-serviced data in order to implement more context within the task list. . No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claims 14 and 31, the claim relates to using an auxiliary LLM to preprocess the retrieved data for the assistant LLM. This relates to the manager asking a human shadow manager to organize and format the to-be-serviced data in such a way that would make it better suited for the employee to execute. The element “auxiliary LLM” is recited at a high level of generality and merely equates to “apply it” or otherwise merely uses a generic computer as a tool to perform an abstract which are not indicative of integration into a practical application as per MPEP 2106.05(f). No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claims 16 and 33, the claim relates to processing the adaptation input while receiving the follow-on query. This relates to the manager writing the task list while receiving the to-be-serviced data from the client. No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
With respect to claims 17 and 34, the claim relates to presenting content and based on presenting content, obtaining a subsequent adaptation input and a subsequent follow-on query. This relates to the manager presenting their employee’s work to the client, receiving client feedback, formulating a second task list for their employee that differs from the first task list, and then sending the second task list to their employee alongside modified to-be-serviced data in order to meet expectations from client feedback. No additional limitations are present. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
For all of the above reasons, taken alone or in combination, claims 1-3, 10, 12-14, 16-20, 27, 29-31, and 33-34 recite a non-statutory mental process.
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, 12, 16-18, 29, and 33-34 are rejected under 35 U.S.C. 103 as obvious over US Patent Publication 20250077780 A1 (Xie et al.) in view of US Patent 11100922 B1 (Mutagi et al.).
Claim 1
Regarding claim 1, Xie et al. disclose a computer-implemented method executed on data processing hardware (Xie et al. ¶ [0228], "The one or more computer programs are executable and/or interpretable on a programmable system including at least one programmable processor.") that causes the data processing hardware to perform operations comprising:
receiving, from a user, [a hotword detection event indication indicating detection of] a particular [hotword] input (Xie et al. ¶ [0037]-[0038], "In S101, natural language content is acquired. The natural language content may refer to the natural language content input by the user during a human-computer interaction process.") [in streaming audio captured by a microphone in communication with the data processing hardware], the particular [hotword] input [comprising one of at least two different predetermined hotwords each] specifying a [different respective] functionality for an assistant large language model (LLM) to undertake for processing a follow-on query from the user (Xie et al. ¶ [0038], "The natural language content may be understood as content provided by the user that includes user intent to call the large language model to implement required functions." User-provided natural language content that specifies a required LLM functionality is considered analogous to a particular input specifying a functionality);
based on the received [hotword detection event indication indicating detection of the] particular [hotword] input, obtaining an adaptation input specifically formulated for adapting the assistant LLM to undertake the respective functionality specified by the particular [hotword] input (Xie et al. ¶ [0039]-[0040], "In S102, semantic understanding is performed on the natural language content, and whether the natural language content hits a plugin is detected to obtain a first plugin pointed to by the plugin hit result. ... The first plugin hit by the natural language content may refer to a plugin that implements the function of the user intent." A plugin is considered analogous to an adaptation input");
receiving, from the user, the follow-on query, the follow-on query comprising a natural language query specifying an action for the assistant LLM to perform (Xie et al. ¶ [0040], "Semantic understanding may be performed on the natural language content to obtain the user intent. The user intent may be matched with the function of each available plugin, and the plugin corresponding to the user intent may be determined as the first plugin.");
providing, for input to the assistant LLM, the adaptation input specifically formulated for adapting the assistant LLM to undertake the respective functionality specified by the particular [hotword] input (Xie et al. ¶ [0044]-[0047], "In S104, ... the related information of the third plugin and the natural language content are provided to the large language model. ... The related information of the third plugin may include information such as the function and input and output of the third plugin. The language understanding content is sent to the large language model to obtain the input parameter of the third plugin and the parameter value corresponding to the input parameter." The third plugin and associated related information is considered analogous to an adaptation input); and
processing, using the adapted assistant LLM undertaking the respective functionality specified by the particular [hotword] input, the follow-on query to fulfill performance of the action specified by the natural language query (Xie et al. ¶ [0049], "In S105, the third plugin is called according to the input parameter of the third plugin to obtain the calling result of the to-be-executed session understanding task.").
Xie et al. do not explicitly disclose all of hotword detection.
However, Mutagi et al. disclose receiving, from a user, a hotword detection event indication indicating detection of a particular hotword (Mutagi et al. ¶ (30), "electronic device 100a may be capable of being activated in response to detecting a specific sound, such as ... one or more inputs. After detecting a specific sound ..., electronic device 100a may recognize commands (e.g., audible commands, inputs, etc.) within captured audio, and may perform one or more actions in response to the received commands." One or more inputs that activate the electronic device is considered analogous to a hotword detection event. One or more inputs includes the captured audio. Thus, recognized commands within the captured audio is considered analogous to a particular hotword) in streaming audio captured by a microphone in communication with the data processing hardware (Mutagi et al. ¶ (32), "electronic device 100a may include, or be in communication with, one or more microphones that listen for a wakeword by continually monitoring local audio. In response to the wakeword being detected, electronic device 100a may establish a connection with computing system 200, send audio data to computing system 200, and await/receive a response from computing system 200"), the particular hotword comprising one of at least two different predetermined hotwords each specifying a different respective functionality for an assistant [large language model (LLM)] to undertake for processing a follow-on query from the user (Mutagi et al. ¶ (76), "An intent classification (“IC”) module 274 may parse the query to determine an intent or intents for each identified domain, where the intent corresponds to the action to be performed that is responsive to the query. Each domain is associated with a database of words linked to intents for that domain. For example, a music intent database of a music domain may link words and phrases such as “play” to a play music intent, “stop” to a stop playing music intent, and “mute” to a mute volume intent." Words and phrases such as "play", "stop", and "mute" are considered analogous to predetermined hotwords that specify different respective functionalities for an assistant to undertake); …
receiving, from the user, the follow-on query, the follow-on query comprising a natural language query specifying an action for the assistant [LLM] to perform (Mutagi et al. ¶ (65), "ASR module 258 may include an expression detector that analyzes audio signals received by the speech-processing system, such as the expression detector mentioned above with regards to electronic device 100a.... The expression detector may then ... determine whether the wakeword or phrase will be declared as having been spoken."); … and
processing, using the [adapted] assistant [LLM] undertaking the respective functionality specified by the particular hotword, the follow-on query to fulfill performance of the action specified by the natural language query (Mutagi et al. ¶ (70), "if a spoken utterance processed by ASR module 258 results in an output of text, “call Mom,” NLU module 260 may determine that the intent of the utterance is to activate a telephone, or telephone functionality, and to initiate a call with a contact represented by an identifier “Mom.”").
It would have been obvious to a person having ordinary skill in the art before the time of the effective filing date of the claimed invention of the instant application to modify Xie et al.’s LLM adaptation method to incorporate Mutagi et al.’s functionality-based hotword detection because such a modification is the result of combining prior art elements according to known methods to yield predictable results. More specifically, Xie et al.’s LLM adaptation method as modified by Mutagi et al.’s functionality-based hotword detection can yield a predictable result of improving user experience since users could assign specific hotwords to certain functionalities, allowing a user to customize the LLM to their personal workflow and preferences. Thus, a person of ordinary skill would have appreciated including in Xie et al.’s LLM adaptation method the ability to do Mutagi et al.’s functionality-based hotword detection since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art would have recognized that the results of the combination were predictable.
Claim 12
Regarding claim 12, the rejection of claim 1 is incorporated.
Xie et al. further disclose wherein the operations further comprise, prior to commencing the processing of the follow-on query using the assistant LLM, commencing processing of the adaptation input to adapt the assistant LLM to undertake the particular functionality (Xie et al. ¶ [0043] "the first plugin is compared with the second plugin to detect whether the intent of the previous session and the intent of the subsequent session are the same, thereby detecting whether the session state is interrupted." Comparing current plugins to previous plugins based on session intent is considered analogous to processing an adaptation input).
Claim 16
Regarding claim 16, the rejection of claim 1 is incorporated.
Xie et al. further disclose wherein:
the operations further comprise processing, using the assistant LLM, the adaptation input (Xie et al. ¶ [0085] "The language understanding content of the to-be-executed session understanding task is acquired and sent to the large language model to obtain the input parameter of the third plugin"); and
the assistant LLM processes the adaptation input while receiving the follow-on query from the user (Xie et al. ¶ [0086] "When the to-be-executed session understanding task is executed, the plugin tasks corresponding to the to-be-executed session understanding task are executed in sequence or in parallel according to the sorting result. The plugin tasks executed in parallel are independent of each other." Performing multiple plugin tasks simultaneously for a single user is considered analogous to processing adaptation inputs while receiving follow-on queries from a user).
Claim 17
Regarding claim 17, the rejection of claim 1 is incorporated.
Xie et al. further disclose wherein:
based on processing the follow-on query to fulfill performance of the action, generating presentation content responsive to the follow-on query (Xie et al. ¶ [0088] "The input parameter of the third plugin corresponding to the currently executed plugin task fed back by the large language model specifically includes the identifier, parameter value, and parameter type of the input parameter. The input parameter corresponding to the currently executed plugin task is acquired, and the third plugin corresponding to the currently executed plugin task is called to obtain the calling result." A calling result is considered analogous to presentation content); and
based on the presentation content, obtaining another adaptation input specifically formulated for adapting the assistant LLM to undertake another particular functionality (Xie et al. ¶ [0088] "After the calling result is fed back, the next plugin task is selected and used as the currently executed plugin task.") specified by a subsequent follow-on query (Xie et al. ¶ [0088] "the natural language content may be the natural language content input by the user during multiple rounds of sessions. For example, during the i-th round of session, the initial session understanding task in the i-th round of session is the current session understanding task, that is, the to-be-executed session understanding task determined in the i-th round of session. The to-be-executed session understanding task is determined based on the natural language content input by the user in the i-th round, that is, the to-be-executed session understanding task determined in the i-th round of session is also used as the initial conversation understanding task in the (i+1)-th round of session.").
Claim 18
Regarding claim 18, Xie et al. disclose data processing hardware (Xie et al. ¶ [0228], "The one or more computer programs are executable and/or interpretable on a programmable system including at least one programmable processor."); and
memory hardware in communication with the data processing hardware (Xie et al. ¶ [0225], "The computing unit 701 may perform various types of appropriate operations and processing based on a computer program stored in a read-only memory (ROM) 702 or a computer program loaded from a storage unit 708 to a random-access memory (RAM) 703.").
The remaining limitations of claim 18 are similar to that of claim 1 and therefore are rejected for similar reasons as described above.
Claim 29
Regarding claim 29, the rejection of claim 18 is incorporated. The limitations of claim 29 are similar to that of claim 12 and therefore are rejected for similar reasons as described above.
Claim 33
Regarding claim 33, the rejection of claim 18 is incorporated. The limitations of claim 33 are similar to that of claim 16 and therefore are rejected for similar reasons as described above.
Claim 34
Regarding claim 34, the rejection of claim 18 is incorporated. The limitations of claim 34 are similar to that of claim 17 and therefore are rejected for similar reasons as described above.
Claims 2-3, 15, 19-20, and 32 are rejected under 35 U.S.C. 103 as obvious over Xie et al. in view of Mutagi et al. as applied to claims 1 and 18 above, and further in view of "Low-code LLM: Graphical User Interface over Large Language Models" (Cai et al.).
Claim 2
Regarding claim 2, the rejection of claim 1 is incorporated.
Xie et al. in view of Mutagi et al. do not explicitly disclose all of a user interface element.
However, Cai et al. disclose wherein receiving a particular trigger input from the user comprises receiving a user input indication indicating selection of a particular user interface (Ul) element displayed on a screen in communication with data processing hardware (Cai et al. pg. 14, Section 2.3, Paragraph 1-2, "To more intuitively present users with the workflow, a flowchart is utilized to visualize the workflow and presented it to users. ... We define six types of low-code interactions for users to edit the workflow").
It would have been obvious to a person having ordinary skill in the art before the time of the effective filing date of the claimed invention of the instant application to modify Xie et al. in view of Mutagi et al. to incorporate Cai et al.’s user interface.
The suggestion/motivation for doing so would have been that, “Through visual interaction with a graphical user interface, users can incorporate their ideas into the process without writing trivial prompts,” as noted by Cai et al. in the abstract.
Claim 3
Regarding claim 3, the rejection of claim 2 is incorporated.
Cai et al. further disclose wherein the particular UI element is one of at least two different UI elements displayed on the screen, each UI element of the at least two different UI elements specifying a different respective particular functionality for the assistant LLM to undertake (Cai et al. Figures 2 and 3 illustrate a user editing steps related to separate and differing tasks that an LLM plans to take).
Claim 15
Regarding claim 15, the rejection of claim 12 is incorporated.
Xie et al. in view of Mutagi et al. do not explicitly disclose all of a user interface element.
However, Cai et al. disclose loading a user interface (Ul) element that was previously generated by the assistant LLM when the assistant LLM was adapted to undertake the particular functionality during fulfillment of a previous query (Cai et al. pg. 13, Section 2.2., Paragraph 2, "We implement the Planning LLM with Chat-GPT1 and educate it to draft a plan with education prompts, which consists of ... (3) Generation of sub-workflow: If a user intends to extend a step, the Planning LLM is provided with the dialogue history of the previous generation of the overall workflow to ensure logical consistency and prevent duplication of content between the sub-workflow the other steps of the overall workflow."); and displaying, on a screen in communication with the data processing hardware, the UI element (Cai et al. pg. 13, Section 2.2., Paragraph 2. See Figure 3 for an illustrative example of a user creating a sub-workflow), wherein processing the follow-on query to fulfill performance of the action specified by the natural language query comprises interacting with the UI element displayed on the screen based on the action specified by the natural language query (Cai et al. pg. 16, Figure 3 illustrates a user manually editing a generated workflow by interacting with UI elements displayed on-screen).
It would have been obvious to a person having ordinary skill in the art before the time of the effective filing date of the claimed invention of the instant application to modify Xie et al. in view of Mutagi et al. to incorporate Cai et al.’s user interface.
The suggestion/motivation for doing so is similar to the suggestion/motivation described above with respect to claim 15.
Claim 19
Regarding claim 19, the rejection of claim 18 is incorporated. The limitations of claim 19 are similar in scope to that of claim 2 and therefore are rejected for similar reasons as described above.
Claim 20
Regarding claim 20, the rejection of claim 19 is incorporated. The limitations of claim 20 are similar in scope to that of claim 3 and therefore are rejected for similar reasons as described above.
Claim 32
Regarding claim 32, the rejection of claim 18 is incorporated. The limitations of claim 32 are similar in scope to that of claim 15 and therefore are rejected for similar reasons as described above.
Claims 6-7 and 23-24 are rejected under 35 U.S.C. 103 as obvious over Xie et al. in view of Mutagi et al. as applied to claims 1 and 18 above, and further in view of US Patent Publication 20250384272 B1 (Pentyala et al.).
Claim 6
Regarding claim 6, the rejection of claim 1 is incorporated.
Xie et al. further disclose obtaining the adaptation input based on the received particular trigger input comprises processing the particular trigger input to identify a particular set [of fine-tuned weights] that map to the particular trigger input, the particular set [of fine-tuned weights] comprising the adaptation input [and trained] to adapt the assistant LLM model to undertake the particular functionality specified by the particular trigger input (Xie et al. ¶ [0040] "Semantic understanding may be performed on the natural language content to obtain the user intent. The user intent may be matched with the function of each available plugin, and the plugin corresponding to the user intent may be determined as the first plugin.") [while the set of pretrained weights of the pretrained assistant LLM are frozen].
Xie et al. in view of Mutagi et al. do not explicitly disclose all of freezing LLM weights.
However, Pentyala et al. disclose wherein:
the assistant LLM comprises a pretrained assistant LLM having a set of pre-trained weights (Pentyala et al. ¶ [0030], "FIG. 1 is a simplified diagram illustrating an example training process 100 of adapting a base neural network model via an adapter, according to embodiments described herein. To adapt a base neural network model 100 with pretrained weights for a specific task... an adapter module 120 may be used to learn the task-specific and/or domain-specific features.");
[processing a particular trigger input to identify] identifying a particular set of fine-tuned weights [that map to the particular trigger input], the particular set of fine-tuned weights comprising the adaptation input and trained to adapt the assistant LLM model to undertake a particular functionality [specified by the particular trigger input] while the set of pretrained weights of the pretrained assistant LLM are frozen (Pentyala et al. ¶ [0023], "During training, instead of updating the weights and/or parameters of the entire LLM, only the weights of the adapter modules are updated while keeping the bulk of the original LLM unchanged."); and
providing the adaptation input for input to the assistant LLM comprises activating the particular set of fine-tuned weights for adapting the assistant LLM to undertake the particular functionality (Pentyala et al. ¶ [0024], "After being updated during training, the adapter module may then be merged with a base LLM such that the merged new model is adapted to perform the specific task or on the specific domain.") [specified by the particular trigger input].
It would have been obvious to a person having ordinary skill in the art before the time of the effective filing date of the claimed invention of the instant application to modify Xie et al. in view of Mutagi et al. to include Pentyala et al.’s LLM weight-based adaptation because such a modification is the result of simple substitution of one known element for another producing a predictable result. More specifically, Xie et al.’s plugins and Pentyala et al.’s weight-based adapter modules perform the same general and predictable function, the predictable function being adapting a general-purpose LLM to a specific task. Since each individual element and its function are shown in the prior art, albeit shown in separate references, the difference between the claimed subject matter and the prior art rests not on any individual element or function but in the very combination itself - that is in the substitution of Xie et al.’s plugins by replacing it with Pentyala et al.’s weight-based adapter modules. Thus, the simple substitution of one known element for another producing a predictable result renders the claim obvious.
Claim 7
Regarding claim 7, the rejection of claim 6 is incorporated.
Pentyala et al. further disclose wherein the particular set of fine-tuned weights comprises one of multiple sets of fine-tuned weights (Pentyala et al. ¶ [0028], "developers may select compatible LLMs as base models to merge with trained adapters to produce new models without additional training and/or finetuning the adapter with different LLMs." A trained adapter is considered analogous to a set of fine-tuned weights. Therefore, a plurality of trained adapters is considered analogous to multiple sets of fine-tuned weights), each corresponding set of fine-tuned weights of the multiple sets of fine-tuned weights:
mapped to a different corresponding trigger input that specifies a different corresponding functionality for the pretrained assistant LLM to undertake (Pentyala et al. ¶ [0023], "To train or adapt a neural network such as an LLM to perform a specific task or on a specific domain, e.g., to understand and generate a legal document, to understand and answer mathematical questions, and/or the like, instead of training and retraining the entire LLM using a training dataset for the specific tasks or domain, an adapter module may be used."); and
trained to adapt the pretrained assistant LLM to undertake the corresponding functionality specified by the corresponding trigger input while the set of pretrained weights of the pretrained assistant LLM are frozen (Pentyala et al. ¶ [0023], "During training, instead of updating the weights and/or parameters of the entire LLM, only the weights of the adapter modules are updated while keeping the bulk of the original LLM unchanged.").
Claim 23
Regarding claim 23, the rejection of claim 18 is incorporated. The limitations of claim 23 are similar in scope to that of claim 6 and therefore are rejected for similar reasons as described above.
Claim 24
Regarding claim 24, the rejection of claim 23 is incorporated. The limitations of claim 24 are similar in scope to that of claim 7 and therefore are rejected for similar reasons as described above.
Claims 8 and 25 are rejected under 35 U.S.C. 103 as obvious over Xie et al. in view of Mutagi et al. in view of Pentyala et al. as applied to claims 6 and 23 above, and further in view of US Patent Publication 20250131209 A1 (Calapodescu et al.).
Claim 8
Regarding claim 8, the rejection of claim 6 is incorporated.
Xie et al. in view of Mutagi et al. in view of Pentyala et al. do not explicitly disclose all of layer disposal between attention heads.
However, Calapodescu et al. disclose wherein:
the pretrained assistant LLM comprises a plurality of multi-head attention layers (Calapodescu et al. ¶ [0048], "The pretrained NMT model may include the transformer encoder-decoder architecture 200 shown in FIG. 2 (transformer cell) or other encoder-decoder architecture, including the encoder 202 and the decoder 204." ¶ [0034], "The self-attention layer 220 inside the encoder 202, and the self-attention layer 230 and cross-attention layer 234 inside the decoder 204, for instance, can be embodied in multi-head attention mechanisms, such as disclosed in the '978 Patent."); and
the particular set of fine-tuned weights are implemented by one or more adaptor layers each disposed within a respective one of the plurality of multi-head attention layers of the pretrained assistant LLM or between a respective pair of the plurality of multi-head attention layers of the pretrained assistant LLM (Calapodescu et al. ¶ [0036], "The MR adapter layer 240a is disposed (e.g., inserted) in the encoder 202 downstream of the encoder feed-forward layer 222, and the MR adapter layer 240b is disposed (e.g., inserted) in the decoder 204 downstream of the decoder feed-forward layer 232.").
It would have been obvious to a person having ordinary skill in the art before the time of the effective filing date of the claimed invention of the instant application to modify Xie et al. in view of Mutagi et al. in view of Pentyala et al. to include Calapodescu et al.’s adapter layer disposal because such a modification is the result of simple substitution of one known element for another producing a predictable result. More specifically, Pentyala et al.’s LoRA adapter integration and Calapodescu et al.’s adapter layer disposal feature perform the same general and predictable function, the predictable function being adapting a base LLM to a specific task. Since each individual element and its function are shown in the prior art, albeit shown in separate references, the difference between the claimed subject matter and the prior art rests not on any individual element or function but in the very combination itself - that is in the substitution of Pentyala et al.’s LoRA adapter integration by replacing it with Calapodescu et al.’s adapter layer disposal. Thus, the simple substitution of one known element for another producing a predictable result renders the claim obvious.
Claim 25
Regarding claim 25, the rejection of claim 23 is incorporated. The limitations of claim 25 are similar in scope to that of claim 8 and therefore are rejected for similar reasons as described above.
Claims 9 and 26 are rejected under 35 U.S.C. 103 as obvious over Xie et al. in view of Mutagi et al. as applied to claims 1 and 18 above, and further in view of "The Power of Scale for Parameter-Efficient Prompt Tuning" (Lester et al.).
Claim 9
Regarding claim 9, the rejection of claim 1 is incorporated.
Xie et al. in view of Mutagi et al. do not explicitly disclose all of an adaptation input that involves concatenation.
However, Lester et al. disclose wherein:
obtaining an adaptation input based on a received particular trigger input comprises processing the particular trigger input to identify a particular fine-tuned user prompt embedding that maps to the particular trigger input (Lester et al. pg. 3, Section 2, Paragraph 3, "Given a series of
n
tokens,
{
x
1
,
x
2
,
.
.
.
,
x
n
}
, the first thing T5 does is embed the tokens, forming a matrix
X
e
∈
R
n
×
e
where
e
is the dimension of the embedding space. Our soft-prompts are represented as a parameter
P
e
∈
R
p
×
e
, where
p
is the length of the prompt. Our prompt is then concatenated to the embedded input forming a single matrix
P
e
;
X
e
∈
R
(
p
+
n
)
×
e
which then flows though the encoder-decoder as normal."), the particular fine-tuned user prompt embedding comprising the adaptation input (Lester et al. pg. 2, Section 1, Paragraph 6, "We freeze the entire pre-trained model and only allow an additional
k
tunable tokens per downstream task to be prepended to the input text." Additional tokens tuned to adapt a language model for a downstream task is considered analogous to an adaptation input); and
providing the adaptation input for input to the assistant LLM comprises:
concatenating the follow-on query with the particular fine-tuned user prompt embedding that maps to the particular trigger input (Lester et al. pg. 3, Section 2, Paragraph 3, "Given a series of
n
tokens,
{
x
1
,
x
2
,
.
.
.
,
x
n
}
, the first thing T5 does is embed the tokens, forming a matrix
X
e
∈
R
n
×
e
where
e
is the dimension of the embedding space. Our soft-prompts are represented as a parameter
P
e
∈
R
p
×
e
, where
p
is the length of the prompt. Our prompt is then concatenated to the embedded input forming a single matrix
P
e
;
X
e
∈
R
(
p
+
n
)
×
e
which then flows though the encoder-decoder as normal." The series of
n
tokens
{
x
1
,
x
2
,
.
.
.
,
x
n
}
is considered analogous to a follow-on query. The soft-prompt is considered analogous to a fine-tuned user prompt); and
providing the concatenation of the follow-on query with the particular fine-tuned prompt embedding as input to the assistant LLM (Lester et al. pg. 3, Section 2, Paragraph 3, "Our prompt is then concatenated to the embedded input forming a single matrix
P
e
;
X
e
∈
R
(
p
+
n
)
×
e
which then flows though the encoder-decoder as normal."),
wherein, when processing the follow-on query to fulfill performance of the action specified by the natural language query, the particular fine-tuned user prompt embedding is configured to guide the assistant LLM to undertake the particular functionality while parameters of the assistant LLM are held fixed (Lester et al. pg. 3, Section 2.1, Paragraph 1, "Conceptually, our soft-prompt modulates the frozen network’s behavior in the same way as text preceding the input.... For classification tasks, a third option is to initialize the prompt with embeddings that enumerate the output classes, similar to the “verbalizers” of Schick and Schütze (2021). Since we want the model to produce these tokens in the output, initializing the prompt with the embeddings of the valid target tokens should prime the model to restrict its output to the legal output classes.").
It would have been obvious to a person having ordinary skill in the art before the time of the effective filing date of the claimed invention of the instant application to modify Xie et al. in view of Mutagi et al. to incorporate Lester et al.’s prompt tuning.
The suggestion/motivation for doing so would have been that, “In Figure 1 (p. 1), we see that prompt tuning becomes more competitive with model tuning as scale increases. At the XXL size (11 billion parameters), prompt tuning matches even the stronger multi-task model tuning baseline, despite having over 20,000 times fewer task-specific parameters,” as noted by Lester et al. in pg. 4, Section 3.1, Paragraph 2.
Claim 26
Regarding claim 26, the rejection of claim 18 is incorporated. The limitations of claim 26 are similar in scope to that of claim 9 and therefore are rejected for similar reasons as described above.
Claims 10 and 27 are rejected under 35 U.S.C. 103 as obvious over Xie et al. in view of Mutagi et al. as applied to claims 1 and 18 above, and further in view of "The Power of Scale for Parameter-Efficient Prompt Tuning" (Lester et al.) in view of "Exploiting Cloze Questions for Few Shot Text Classification and Natural Language Inference" (Schick and Schütze).
Claim 10
Regarding claim 10, the rejection of claim 1 is incorporated.
Xie et al. in view of Mutagi et al. do not explicitly disclose all of an adaptation input that involves concatenation.
However, Lester et al. disclose wherein:
obtaining the adaptation input based on the received particular trigger input comprises processing the particular trigger input to identify a particular [natural language] prefix prompt that maps to the particular trigger input (Lester et al. pg. 3, Section 2, Paragraph 3, "Given a series of
n
tokens,
{
x
1
,
x
2
,
.
.
.
,
x
n
}
, the first thing T5 does is embed the tokens, forming a matrix
X
e
∈
R
n
×
e
where
e
is the dimension of the embedding space. Our soft-prompts are represented as a parameter
P
e
∈
R
p
×
e
, where
p
is the length of the prompt. Our prompt is then concatenated to the embedded input forming a single matrix
P
e
;
X
e
∈
R
(
p
+
n
)
×
e
which then flows though the encoder-decoder as normal."), the particular [natural language] prefix prompt comprises the adaptation input (Lester et al. pg. 2, Section 1, Paragraph 6, " We freeze the entire pre-trained model and only allow an additional
k
tunable tokens per downstream task to be prepended to the input text." Additional tokens that are prepended to input text is considered analogous to a [natural language] prefix prompt); and
providing the adaptation input for input to the assistant LLM comprises:
concatenating the follow-on query with the particular [natural language] prefix prompt that maps to the particular trigger input (Lester et al. pg. 3, Section 2, Paragraph 3, "Given a series of
n
tokens,
{
x
1
,
x
2
,
.
.
.
,
x
n
}
, the first thing T5 does is embed the tokens, forming a matrix
X
e
∈
R
n
×
e
where
e
is the dimension of the embedding space. Our soft-prompts are represented as a parameter
P
e
∈
R
p
×
e
, where
p
is the length of the prompt. Our prompt is then concatenated to the embedded input forming a single matrix
P
e
;
X
e
∈
R
(
p
+
n
)
×
e
which then flows though the encoder-decoder as normal." The series of
n
tokens
{
x
1
,
x
2
,
.
.
.
,
x
n
}
is considered analogous to a follow-on query. The soft-prompt is considered analogous to a natural language prefix prompt); and
providing the concatenation of the follow-on query with the particular [natural language] prefix prompt as input to the assistant LLM (Lester et al. pg. 3, Section 2, Paragraph 3, "Our prompt is then concatenated to the embedded input forming a single matrix
P
e
;
X
e
∈
R
(
p
+
n
)
×
e
which then flows though the encoder-decoder as normal."),
wherein, when processing the follow-on query to fulfill performance of the action specified by the natural language query, the particular [natural language] prefix prompt is configured to instruct the assistant LLM to undertake the particular functionality (Lester et al. pg. 3, Section 2.1, Paragraph 1, "Conceptually, our soft-prompt modulates the frozen network’s behavior in the same way as text preceding the input.... For classification tasks, a third option is to initialize the prompt with embeddings that enumerate the output classes, similar to the “verbalizers” of Schick and Schütze (2021). Since we want the model to produce these tokens in the output, initializing the prompt with the embeddings of the valid target tokens should prime the model to restrict its output to the legal output classes.").
It would have been obvious to a person having ordinary skill in the art before the time of the effective filing date of the claimed invention of the instant application to modify Xie et al. in view of Mutagi et al. to incorporate Lester et al.’s prompt tuning.
The suggestion/motivation for doing so is similar to the suggestion/motivation described above with respect to claim 9.
Xie et al. in view of Mutagi et al. in view of Lester et al. do not explicitly disclose all of a particular natural language prompt for concatenation.
However, Schick and Schütze disclose processing the particular trigger input to identify a particular natural language prefix prompt that maps to the particular trigger input (Schick and Schütze pg. 2-3, Section 3, Paragraph 2, "consider the task of identifying whether two sentences
a
and
b
contradict each other (label
y
0
) or agree with each other (
y
1
). For this task, we may choose the pattern
P
(
a
,
b
)
=
a
?
_
_
_
_
,
b
.
combined with a verbalizer
v
that maps
y
0
to “Yes” and
y
1
to “No”. Given an example input pair
x
=
(
M
i
a
l
i
k
e
s
p
i
e
,
M
i
a
h
a
t
e
s
p
i
e
)
, the task now changes from having to assign a label without inherent meaning to answering whether the most likely choice for the masked position in
P
(
x
)
=
M
i
a
l
i
k
e
s
p
i
e
?
_
_
_
_
,
M
i
a
h
a
t
e
s
p
i
e
.
is "Yes" or "No".")
It would have been obvious to a person having ordinary skill in the art before the time of the effective filing date of the claimed invention of the instant application to modify Xie et al. in view of Mutagi et al. in view of Lester et al. to include Schick and Schütze’s natural language prefix because such a modification is the result of simple substitution of one known element for another producing a predictable result. More specifically, Lester et al.’s fine-tuned user prompt and Schick and Schütze’s natural language prefix perform the same general and predictable function, the predictable function being adding trainable parameters to the input of an LLM. Since each individual element and its function are shown in the prior art, albeit shown in separate references, the difference between the claimed subject matter and the prior art rests not on any individual element or function but in the very combination itself - that is in the substitution of Lester et al.’s fine-tuned user prompt by replacing it with Schick and Schütze’s natural language prefix. Thus, the simple substitution of one known element for another producing a predictable result renders the claim obvious.
Claim 27
Regarding claim 27, the rejection of claim 18 is incorporated. The limitations of claim 27 are similar in scope to that of claim 10 and therefore are rejected for similar reasons as described above.
Claims 11 and 28 are rejected under 35 U.S.C. 103 as obvious over Xie et al. in view of Mutagi et al. as applied to claims 1 and 18 above, and further in view of "Language Models are Few-Shot Learners" (Brown et al.).
Claim 11
Regarding claim 11, the rejection of claim 1 is incorporated.
Xie et al. in view of Mutagi et al. do not explicitly disclose all of utilizing few-shot examples.
However, Brown et al. disclose wherein:
the assistant LLM comprises a pretrained assistant LLM having a set of pretrained weights (Brown et al. pg. 5, Section 1, Paragraph 9, "In this paper, we test this hypothesis by training a 175 billion parameter autoregressive language model, which we call GPT-3, and measuring its in-context learning abilities."); and
obtaining the adaptation input based on the received particular trigger input comprises processing the particular trigger input to identify a particular set of one or more few-shot learning examples that maps to the particular trigger input (Brown et al. pg. 10, Section 2.4, Paragraph 1, "For few-shot learning, we evaluate each example in the evaluation set by randomly drawing K examples from that task’s training set as conditioning, delimited by 1 or 2 newlines depending on the task."), the particular set of one or more few-shot learning examples comprises an adaptation input (Brown et al. pg. 10, Section 2.4, Paragraph 3, "On tasks that involve choosing one correct completion from several options (multiple choice), we provide K examples of context plus correct completion, followed by one example of context only, and compare the LM likelihood of each completion." Few-shot examples constructed for a particular task is considered analogous to an adaptation input),
wherein each few-shot learning example in the particular set of the one or more few-shot learning examples depicts an example query input paired with a ground-truth response of the example query input (Brown et al. pg. 50, Appendix G, "The following figures illustrate the formatting and phrasing of all the tasks included in the paper. All data comes from the ground truth datasets in this section, and no samples from GPT-3 are included here." See Figure G.1. The section labelled "Context" is considered analogous to an example query. The section labelled with "Answers" is considered analogous to a ground-truth response) to provide in-context learning for adapting the assistant LLM to generalize to the particular functionality [specified by the trigger input] (Brown et al. pg. 3, Figure 1.1 description, "During unsupervised pre-training, a language model develops a broad set of skills and pattern recognition abilities. It then uses these abilities at inference time to rapidly adapt to or recognize the desired task. We use the term “in-context learning” to describe the inner loop of this process, which occurs within the forward-pass upon each sequence.").
It would have been obvious to a person having ordinary skill in the art before the time of the effective filing date of the claimed invention of the instant application to modify Xie et al. in view of Mutagi et al. to incorporate Brown et al.’s few-shot examples.
The suggestion/motivation for doing so would have been that, “the main advantages of few-shot are a major reduction in the need for task-specific data and reduced potential to learn an overly narrow distribution from a large but narrow fine-tuning dataset,” as noted by Brown et al. in pg. 6, Section 2, Paragraph 1.
Claim 28
Regarding claim 28, the rejection of claim 18 is incorporated. The limitations of claim 28 are similar in scope to that of claim 11 and therefore are rejected for similar reasons as described above.
Claims 13-14 and 30-31 are rejected under 35 U.S.C. 103 as obvious over Xie et al. in view of Mutagi et al. as applied to claims 12 and 29 above, and further in view of US Patent Publication 20240289863 (Smith Lewis et al.).
Claim 13
Regarding claim 13, the rejection of claim 12 is incorporated.
Xie et al. further disclose wherein commencing the processing of the adaptation input comprises [performing vector index lookups to retrieve] retrieving content relevant to the particular functionality specified by the particular trigger input for use by the assistant LLM once processing of the follow-on query commences (Xie et al. ¶ [0095] "in the case where the to-be-executed session understanding task is different from the new session understanding task, the context of the to-be-executed session understanding task is acquired; and the language understanding content of the to-be-executed session understanding task is determined according to the context of the to-be-executed session understanding task and the natural language content."), the retrieved content comprising at least one of:
…
one or more documents that were previously accessed by the assistant LLM (Xie et al. ¶ [0101] "All pre-recorded context and the natural language content are provided to the large language model. ... For example, after the user successfully books a flight ticket, the user enters the ticket booking request again, and the large language model may reply: Do you want to re-book or modify the history booking? Thus, whether the user calls the third plugin to modify the order or generate a new order is clarified to provide more accurate input parameters.") to fulfill one or more previous queries when the assistant LLM was adapted to undertake the same particular functionality (Xie et al. ¶ [0073]-[0076] "When it is determined that the session is interrupted, it may be detected whether the first plugin is the same as a fourth plugin corresponding to a historical session understanding task to determine whether the session is recovered. ... The historical session understanding task may be a session understanding task generated through a historical session that hits the fourth plugin but is not in the current execution state.")….
Xie et al. in view of Mutagi et al. do not explicitly disclose all of performing vector index lookups to retrieve content.
However, Smith Lewis et al. disclose performing vector index lookups to retrieve content (Smith Lewis et al. ¶ [0045] "the knowledge base may be used to store information specific to an individual user and their past interactions with the system. For example, the conversation history between the system and the user, ... or similar text, audio or visual representations of the user's history with the system, may be embedded as indexed vectors in a vector database or similar storage structure for later reference by the system. In this way the system may develop a persistent and accessible record of the user's past interaction history analogous to human long-term memory.")
It would have been obvious to a person having ordinary skill in the art before the time of the effective filing date of the claimed invention of the instant application to modify Xie et al. in view of Mutagi et al. to incorporate Smith Lewis et al.’s vector indexing.
The suggestion/motivation for doing so would have been that, “Together, embeds and indexes may afford embodiments of the system to quickly retrieve relevant information from the knowledge base and use it to inform a decision-making process and deliver a personalized user experience. Embeds and indexes improve upon the efficiency of search and access of information in the knowledge base, reducing response latency for the system while affording the ability to deliver relevant and personalized responses to users.” as noted by the Smith Lewis et al. disclosure in paragraph [0044].
Claim 14
Regarding claim 14, the rejection of claim 13 is incorporated.
Xie et al. further disclose instructing an auxiliary LLM to preprocess the retrieved content (Xie et al. ¶ [0197] "The intent recognition model is a small language model having 100 million parameters and is configured to determine whether an input message includes the call intent to one or several plugins."); and
receiving, from the auxiliary LLM, preprocessed results for the retrieved content (Xie et al. ¶ [0197] "The intent recognition model... is configured to determine whether an input message includes the call intent to one or several plugins."), wherein commencing the processing of the adaptation input to adapt the assistant LLM to undertake the particular functionality comprises using the preprocessed results to adapt the assistant LLM to undertake the particular functionality (Xie et al. ¶ [0198] "For the intent that requires multiple plugins to satisfy by orchestration, the task planning module may generate an ordered plugin execution plan according to the input message and the related plugin list.").
Claim 30
Regarding claim 30, the rejection of claim 29 is incorporated. The limitations of claim 30 are similar in scope to that of claim 13 and therefore are rejected for similar reasons as described above.
Claim 31
Regarding claim 31, the rejection of claim 30 is incorporated. The limitations of claim 31 are similar in scope to that of claim 14 and therefore are rejected for similar reasons as described above.
Reference Cited
The prior art made of record and not relied upon is considered pertinent to applicant’s disclosure.
US Patent Publication 20230030738 A1 to Seo et al. discloses receiving and interpreting a plurality of hotwords, each hotword corresponding to a different respective functionality for an assistant to undertake.
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 JACOB B VOGT whose telephone number is (571)272-7028. The examiner can normally be reached Monday - Friday, 11am - 8pm 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, PARAS D SHAH can be reached at (571)270-1650. 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.
/JACOB B VOGT/ Examiner, Art Unit 2653
/Paras D Shah/ Supervisory Patent Examiner, Art Unit 2653
08/25/2026
