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 .
Specification
The abstract of the disclosure is objected to because the format is in legalese/claim form, as well as, a statement in the improvement over the current state-of-the-art. A corrected abstract of the disclosure is required and must be presented on a separate sheet, apart from any other text. See MPEP § 608.01(b).
Claim Objections
Claims 10&20 are objected to because of the following informalities: The claims contain a Trademark moniker; the claim is open-ended based on that”DeepSeek” can change meanings in the future. Examiner recommends replacing with a suitable explainable definition. Appropriate correction is required.
Claim Rejections - 35 USC § 102
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)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claim(s) 1,2,4-8,11-12,14-18 are rejected under 35 U.S.C. 102(a)(2) as being anticipated by Goel et al (20220269870).
As per claim 1, Goel et al (20220269870) discloses a system comprising:
a communication interface configured to receive a request for information from a device over a network, the request including text-input and context data (para [0035] - "Alternatively, if a wireless connection between the rendering device 137 and the companion device 138 is available, the application on the rendering device 137 may communicate the user request (optionally including additional data and/or contextual information available to the rendering device 137) to the companion application on the companion device 138 via the wireless connection. The companion application on the companion device 138 may then communicate the user request (optionally including additional data and/or contextual information available to the companion device 138) to the assistant system 140 via the network 110."; para [0079] - "In particular embodiments, computer system 1000 includes a processor 1002, memory 1004, storage 1006, an input/output (I/O) interface 1008, a communication interface 1010, and a bus 1012."); a system bus for determining a large language model from a plurality of large language models based at least in part on the context data (para [0031] - In particular embodiments, the user may interact with the assistant system 140 by providing user input to the assistant application 136 via various modalities (e.g., audio, voice, text, vision, image, video, gesture, motion, activity, location, orientation)." para [0103] - "In particular embodiments, the NLG component 372 may use different language models and/or language templates to generate natural-language outputs."; para [0109] - "The personalized language model may be a model of the probabilities that various word sequences may occur in the language. The sounds of the phonetic units in the audio input may be matched with word sequences using the personalized language model, and greater weights may be assigned to the word sequences that are more likely to be phrases in the language As an example and not by way of limitation, the assistant system 140 may pre-compute a plurality of personalized language models for a plurality of possible subjects a user may talk about."; see also para [0035] and [0079]); providing the text- input, or input data derived from the text-input into the large language model (para [0031] and [0109]); and sending the output of the large language model, or data derived from the output of the large language model to the device for further processing or output by the device or another device (para [0031], [0035], [0103] and [0109]).
As per claim 2, Goel et al (20220269870) teaches the system of claim 1, further disclosing wherein the context data includes location data (para [0052] - "In particular embodiments, the user input may comprise non-speech data, which may be received at a local context engine 220a. As an example and not by way of limitation, the non-speech data may comprise locations, visuals, touch, gestures, world updates, social updates, contextual information, information related to people, activity data, and/or any other suitable type of non- speech data.").
As per claim 4, Goel et al (20220269870) teaches the system of claim 1 further comprising:
wherein the plurality of large language models are customized for a particular culture or language or service (para [0080] - "In particular embodiments, an assistant service module 305 may access a request manager 310 upon receiving a user input. In particular embodiments, the request manager 310 may comprise a context extractor 312 and a conversational understanding object generator (CU object generator) 314. The context extractor 312 may extract contextual information associated with the user input."; para [0082] - "In particular embodiments, the NLU module 210 may identify one or more of a domain, an intent, or a slot from the user input in a personalized and context-aware manner." - The custom language models are utilized for personal assistant services.; see also para [0103] and [0109]).
As per claim 5, Goel et al (20220269870) teaches the system of claim 1. Meta further discloses wherein the LLM restructures the request for information into a form suitable for processing and adds an additional request to consult additional context data that are available on or accessible from the network (para [0031], [0035], [0103] and [0109]).
As per claim 6, Goel et al (20220269870) teaches the system of claim 1 further comprising wherein the request for information is converted to a natural language representation, and the request for information encapsulates and text or speech input from a user with metadata about the form of the actual input (para [0046] - "In particular embodiments, the assistant system 140 may create and store a user profile comprising both personal and contextual information associated with the user. In particular embodiments, the assistant system 140 may analyze the user input using natural-language understanding (NLU) techniques."; para [0081] - "The domain classification/selection results may be further processed based on two related procedures. In one procedure, the NLU module 210 may process the domain classification/selection results using a meta-intent classifier 336a. The meta-intent classifier 336a may determine categories that describe the user's intent."; see also para [0031] and [0035]).
As per claim 7, Goel et al (20220269870) teaches the system of claim 1 further comprising:
wherein the system bus: invokes at least one agent, tool, or service on the device based on the request for information (para [0031], [0035], [0103] and [0109]); consults the LLM to determine if the intent of the request for information has been met (para [0107] - "Completing a task correctly and successfully to deliver the value to the user may be the goal that the assistant system 140 is optimized for."; para [0112] - "In particular embodiments, the entity resolution module 212 may provide one or more of the intents, slots, entities, events, context, or user memory to the dialog state tracker 218. The dialog state tracker 218 may identify a set of state candidates for a task accordingly, conduct interaction with the user to collect necessary information to fill the state, and call the action selector 222 to fulfill the task. The task state may comprise all the current information about a task execution status, such as arguments, confirmation status, confidence score, etc. Any incorrect or outdated information in the task state may lead to failure or incorrect task execution. The task state may also serve as a set of contextual information for many other components such as the ASR module 208, the NLU module 210, etc."; see also para [0035]); in accordance with the intent of the request not fully being met, resending the request for information (para [0035], [0107] and [0112]); in accordance with the intent of the request being fully met, requesting the LLM to format a response for presentation on the device (para [0107] - "Completing a task correctly and successfully to deliver the value to the user may be the goal that the assistant system 140 is optimized for. In particular embodiments, an assistant task may be defined as a capability or a feature. The assistant task may be shared across multiple product surfaces if they have exactly the same requirements so it may be easily tracked."; para [0127] - "In particular embodiments, the client system 130 may comprise one or more rendering devices and one or more companion devices. Correspondingly, the one or more formats may comprise rendering the readout at one or more destination devices selected from the rendering devices and the companion devices; see also para [0035], [0109] and [0112]).
Regarding claim 8, Goel et al (20220269870) teaches the system of claim 1 further comprising:
wherein the LLM categorizes the request for information and forwards the request for information to another LLM based on the categorization (para [0081] - "The domain classification/selection results may be further processed based on two related procedures. In one procedure, the NLU module 210 may process the domain classification/selection results using a meta-intent classifier 336a. The meta-intent classifier 336a may determine categories that describe the user's intent.": para [0083] - "The generic entity resolution 342 may resolve the entities by categorizing the slots and meta slots into different generic topics. The domain entity resolution 340 may resolve the entities by categorizing the slots and meta slots into different domains." - The system may process the request based on the intent and domain of the request.; see also para [0031], [0035], [0103] and [0109]).
Regarding claim 11, Goel et al (20220269870) teaches a method comprising: receiving, from a device, a request for information, the request including text input and context data (para [0035]); determining, with at least one processor, a large language model from a plurality of large language models based at least in part on the context data (para [0031], [0079], [0103] and [0109]); providing, with the at least one processor, the text- input, or input data derived from the text-input into the large language model (para [0031], [0079] and [0109]); and sending the output of the large language model, or data derived from the output of the large language model to the device for further processing or output by the device or another device (para [0031], [0035], [0103] and [0109).
Regarding claim 12, Goel et al (20220269870) teaches the method of claim 11, further comprising: wherein the context data includes location data (para [0052]).
Regarding claim 14, Goel et al (20220269870) teaches the method of claim 11, further comprising wherein the plurality of large language models are customized for a particular culture or language or service (para [0080], [0103] and [0109]).
Regarding claim 15, Goel et al (20220269870) teaches the method of claim 11 further comprising wherein the LLM restructures the request for information into a form suitable for processing and adds an additional request to consult additional context data that are available on or accessible from the network (para [0031], [0035], [0103] and [0109]).
Regarding claim 16, Goel et al (20220269870) teaches the method of claim 11 further comprising wherein the request for information is converted to a natural language representation, and the request for information encapsulates and text or speech input from a user with metadata about the form of the actual input (para [0031], [0035], [0046] and [0081]).
Regarding claim 17, Goel et al (20220269870) teaches the method of claim 11 further comprising wherein the system bus: invokes at least one agent, tool, or service on the device based on the request for information (para [0031], [0035], [0103] and [0109]); consults the LLM to determine if the intent of the request for information has been met (para [0035], [0107] and [0112]); in accordance with the intent of the request not fully being met, resending the request for information (para [0035], [0107] and [0112]); in accordance with the intent of the request being fully met, requesting the LLM to format a response for presentation on the device(para [0035], [0107], [0109], [0112] and [0127]).
Regarding claim 18, Goel et al (20220269870) teaches the method of claim 11 further comprising wherein the LLM categorizes the request for information and forwards the request for information to another LLM based on the categorization (para [0031], [0035], [0081], [0083], [0103] and [0109]).
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.
Claim(s) 10,20 are rejected under 35 U.S.C. 103 as being unpatentable over Goel et al (20220269870) in view of admitted prior art (applicants spec, para 0026).
Claims 10 and 20 lack are obvious over Goel et al (20220269870). In view of applicants admitted prior art:
Regarding claim 10, Goel et al (20220269870) discloses the system of claim 1 further comprising:
suggests that the LLM could be any known type of LLM (para [0103] - "In particular embodiments, the NLG component 372 may use different language models and/or language templates to generate natural-language outputs.") but does not explicitly mention a DeepSeek(TM) LLM. Nevertheless, DeepSeek(TM) LLMs were known in the prior art as admitted by applicant in para [0026] of the instant specification. It would have been obvious to a skilled artisan to utilize a DeepSeek(TM) LLM in Meta's systems and methods to advantageously provide multiple alternative known LLMs, as suggested by Goel et al (20220269870).
Regarding claim 20, Meta discloses the method of claim 11. Meta further suggests that the LLM could be any known type of LLM (para [0103]) but does not explicitly mention a DeepSeek(TM) LLM. Nevertheless, DeepSeek(TM) LLMs were known in the prior art as admitted by applicant in para [0026] of the instant specification. It would have been obvious to a skilled artisan to utilize a DeepSeek(TM) LLM in Meta's systems and methods to advantageously provide multiple alternative known LLMs, as suggested by Goel et al (20220269870).
Claim(s) 3,13 are rejected under 35 U.S.C. 103 as being unpatentable over Goel et al (20220269870) in view of Yadgar et al (20190130912).
As per claim 3, Goel et al (20220269870) teaches the system of claim 1; however, Goel et al (20220269870) does not disclose wherein the plurality of large language models are customized for a particular geographic region. However, Yadgar et al (20190130912) docs teaches wherein the plurality of large language models are customized for a particular geographic region (para [0021] - "In various embodiments of the present disclosure, the sentence-level understanding module 130 may also be configured to receive a domain-specific plug-in, such as domain-specific language model 182. For example, the domain-specific language model 182 may include grammar rules and semantic information (such as proper nouns, people's names, place names, email addresses, phrases, telephone numbers, dates, times, addresses, and the like) which are specific to the particular domain. In one embodiment, the domain-specific language model 182 may comprise an ontology."; para [0036] - "In some embodiments, the generic virtual personal assistant platform 110 is configured to utilize additional information such as personalization information, date and time information, geographic or other location information, and other information."; para [0058] - "The run-time specification may also include one or more domain-specific language models."). It would have been obvious to one of ordinary skill in the art to disclose wherein the plurality of large language models are customized for a particular geographic region, as taught by Yadgar et al (20190130912) to the system of Goel et al (20220269870), as it would allow the system to generate a custom LLM based on the user location, as specified by the tenant application (see SRI para [0021], [0036] and [0058]).
Regarding claim 13, Goel et al (20220269870) discloses the method of claim 11; however, Goel et al (20220269870) does not explicitly teach wherein the plurality of large language models are customized for a particular geographic region. However, Yadgar et al (20190130912) does disclose wherein the plurality of large language models are customized for a particular geographic region (para [0021], [0036] and [0058]). It would have been obvious to one of ordinary skill in the art to disclose wherein the plurality of large language models are customized for a particular geographic region, as taught by Yadgar et al (20190130912) to the system of Goel et al (20220269870), as it would allow the system to generate a custom LLM based on the user location, as specified by the tenant application (see Yadgar et al (20190130912) para [0021], [0036] and [0058]).
Claim(s) 9,19 are rejected under 35 U.S.C. 103 as being unpatentable over Goel et al (20220269870) in view of Mimassi (20220383859).
As per claim 9, Goel et al (20220269870) teaches the system of claim 1 Goel et al (20220269870) teaches the method of claim 11 further comprising wherein the LLM or a portion thereof is downloaded to the device, and processes at least a portion of the text data or context data on the device. However, Mimassi (20220383859) does disclose wherein the LLM or a portion thereof is downloaded to the device, and processes at least a portion of the text data or context data on the device (para [0061] - "The system 100 operating both on a mobile device and on federated language model server 110, allow the central and local language models to continuously learn in a nested way. Initially, the local language models operating on an individual's mobile device are using default model parameters generated via the central language models 204 running on the server 110 and downloaded to the mobile device. As the local language models operate on the mobile device, they become tailored to the device user as the user interacts with the device and as the device gathers device and user-specific context data."). It would have been obvious to one of ordinary skill in the art to disclose wherein the LLM or a portion thereof is downloaded to the device, and processes at least a portion of the text data or context data on the device, as taught by Mimassi (20220383859) to the system of Goel et al (20220269870), as it would allow the system to process contextual data and store the customized LLM for the user, as specified by the tenant application (see Mimassi (20220383859) para [0061]). Regarding claim 19, Goel et al (20220269870) discloses the method of claim 11. Goel et al (20220269870) does not disclose wherein the LLM or a portion thereof is downloaded to the device, and processes at least a portion of the text data or context data on the device. However, Mimassi (20220383859) does disclose wherein the LLM or a portion thereof is downloaded to the device, and processes at least a portion of the text data or context data on the device (para [0061]). It would have been obvious to one of ordinary skill in the art to disclose wherein the LLM or a portion thereof is downloaded to the device, and processes at least a portion of the text data or context data on the device, as taught by Mimassi (20220383859) to the system of Goel et al (20220269870), as it would allow the system to process contextual data and store the customized LLM for the user, as specified by the tenant application (see Goel et al (20220269870) para [0061]).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Please see related art listed on the PTO-892 form.
Furthermore, the following references were found, to be pertinent to applicants claims/specification.
Wu et al (20240419903) teaches LLM to address a multimodal input (para 0034).
Gupta et al (20240410453) teaches generation of the prompts to check for potential extra-matching according to the desired model.
Pathak et al (20250157473) teaches neural network language models to reduce the amount of data processing using more efficient compact models – para 0018.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Michael Opsasnick, telephone number (571)272-7623, who is available Monday-Friday, 9am-5pm.
If attempts to reach the examiner by telephone are unsuccessful, the examiner's supervisor, Mr. Richemond Dorvil, can be reached at (571)272-7602. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free).
/Michael N Opsasnick/Primary Examiner, Art Unit 2658 09/03/2026
Prosecution Insights