DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Arguments
Applicant's arguments filed 6/12/2026 have been fully considered
35 USC § 102 & 35 USC § 103:
Regarding Applicant’s Argument (pages: 9-11): Examiner’s response:- Regarding the amendments for independent claim 1 and 15: Applicant’s arguments with respect to the rejection(s) of under 35 USC § 102/103 have been fully considered, upon further consideration, a new ground(s) of rejection is made in view of US 20250086467 A1; Yee; Victor et al. (hereinafter Yee). Regarding the amendments for independent claim 8: Applicant’s arguments with respect to the rejection(s) of under 35 USC § 102/103 have been fully considered, upon further consideration, a new ground(s) of rejection is made in view of US 20250373574 A1; Deutsch; Noah (hereinafter Deutsch)
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 1,3-5,15-17, and 19 are rejected under 35 U.S.C. 103 as being unpatentable over US 20250117410 A1; AGHAJANYAN; Viktor et al. (hereinafter Agah) in view of US 20230237114 A1; Bosarge; Jason (hereinafter Bosarge), US 20190190803 A1; Joshi; Prabodh et al. (hereinafter Joshi) and US 20250086467 A1; Yee; Victor et al. (hereinafter Yee)
Regarding claim 1, Agah teaches A computer-implemented method, comprising: associating a first artificial intelligence (AI) agent with a first container comprising a first set of files and a first set of embeddings representing content of the first set of files; (Agah [0030] The autonomous LLM agent 102 is a computer program that may be instantiated as one or more computer processes. In certain examples, each instance of the autonomous LLM agent 102 may be executed within its own container (e.g., a docker container or the like). In other examples, separate instances of the autonomous LLM agent 102 may be instantiated for each query that is being processed. The autonomous LLM agent 102 may act as a controller or the like that processes a query submitted by a user, determines a workflow to execute (e.g., a dynamically generated and/or optimized set of states for the workflow), and then manages the execution of that workflow. In certain examples, the autonomous LLM agent 102 is configured to parse a given query, dynamically generate a task list that includes one or more states, and then execute tasks within that list using one or more sub-agents that are associated with that state.[0032] The autonomous LLM agent 102 can include multiple different sub-agents (which may also be called agents or tools herein) that can be individually used in connection with processing a query. The different sub-agents that the autonomous LLM agent 102 may use are discussed below...[0040] Turning now more specifically to the autonomous LLM agent 102 and the sub-agents thereof. The autonomous LLM agent 102 includes or has access to sub-agents that may be individually selected and used based on the particular nature of the query being processed by the autonomous LLM agent 102. The sub-agents of the autonomous LLM agent 102 can be thought of as individual tools that the agent 102 can employ in connection with the dynamically constructed workflow. Accordingly, the autonomous LLM agent 102 can be thought of as a multi-tool agent in certain examples. The sub-agents include any or all of the following:..[55-64] elaborates on the different agents with different data storage/containers and corresponding different data/embeddings [FIG.1] shows corresponding visual) wherein the first Al agent includes a first language model; (Agah [0029] System 100 includes an autonomous LLM agent 102, database 106, an application server 108, and one or more large language models (LLMs) 104, or interfaces to such LLMs.[0032] The autonomous LLM agent 102 can include multiple different sub-agents (which may also be called agents or tools herein) that can be individually used in connection with processing a query. The different sub-agents that the autonomous LLM agent 102 may use are discussed below. [0033] The LLMs 104 may be maintained within system 100, external to system 100, or both. Non-limiting illustrative examples of LLMs (or services that interface with LLMs) include ChatGPT from OpenAI, Claude/Claude 2 from Anthropic, and Amazon Titan from Amazon.[0035] Databases 106 may also include a relational database 142 that includes one or more fields that have been extracted from documents. In some examples, relational database 142 stores the results of prompts that have been processed against one or more of the LLMs 104. [0075] Next, at 406, the process determined which sub-agent(s) to use in connection with each of the states for the given workflow. In some example embodiments, as with the generation of the states, the determination of which sub-agent to use for a given state may also use LLM 104 (which may be the same or a different LLM that those prompted previously in connection with 402 and 404). [71-75 and 123] further elaborate on the matter [FIG.1] shows the different LLM(s) which can be integrated with the different agents) associating a second AI agent with a second container comprising a second set of files and a second set of embeddings representing content of the second set of files; (Agah [0030] The autonomous LLM agent 102 is a computer program that may be instantiated as one or more computer processes. In certain examples, each instance of the autonomous LLM agent 102 may be executed within its own container (e.g., a docker container or the like). In other examples, separate instances of the autonomous LLM agent 102 may be instantiated for each query that is being processed. The autonomous LLM agent 102 may act as a controller or the like that processes a query submitted by a user, determines a workflow to execute (e.g., a dynamically generated and/or optimized set of states for the workflow), and then manages the execution of that workflow. In certain examples, the autonomous LLM agent 102 is configured to parse a given query, dynamically generate a task list that includes one or more states, and then execute tasks within that list using one or more sub-agents that are associated with that state.[0032] The autonomous LLM agent 102 can include multiple different sub-agents (which may also be called agents or tools herein) that can be individually used in connection with processing a query. The different sub-agents that the autonomous LLM agent 102 may use are discussed below...[0040] Turning now more specifically to the autonomous LLM agent 102 and the sub-agents thereof. The autonomous LLM agent 102 includes or has access to sub-agents that may be individually selected and used based on the particular nature of the query being processed by the autonomous LLM agent 102. The sub-agents of the autonomous LLM agent 102 can be thought of as individual tools that the agent 102 can employ in connection with the dynamically constructed workflow. Accordingly, the autonomous LLM agent 102 can be thought of as a multi-tool agent in certain examples. The sub-agents include any or all of the following:..[55-64] elaborates on the different agents with different data storage/containers and corresponding different data/embeddings [FIG.1] shows corresponding visual) wherein the second Al agent includes a second language model different from the first language model; (Agah [0029] System 100 includes an autonomous LLM agent 102, database 106, an application server 108, and one or more large language models (LLMs) 104, or interfaces to such LLMs.[0032] The autonomous LLM agent 102 can include multiple different sub-agents (which may also be called agents or tools herein) that can be individually used in connection with processing a query. The different sub-agents that the autonomous LLM agent 102 may use are discussed below. [0033] The LLMs 104 may be maintained within system 100, external to system 100, or both. Non-limiting illustrative examples of LLMs (or services that interface with LLMs) include ChatGPT from OpenAI, Claude/Claude 2 from Anthropic, and Amazon Titan from Amazon.[0035] Databases 106 may also include a relational database 142 that includes one or more fields that have been extracted from documents. In some examples, relational database 142 stores the results of prompts that have been processed against one or more of the LLMs 104. [0075] Next, at 406, the process determined which sub-agent(s) to use in connection with each of the states for the given workflow. In some example embodiments, as with the generation of the states, the determination of which sub-agent to use for a given state may also use LLM 104 (which may be the same or a different LLM that those prompted previously in connection with 402 and 404). [71-75 and 123] further elaborate on the matter [FIG.1] shows the different LLM(s) which can be integrated with the different agents) receiving, via a user interface, a first query and an indication of the first container; (Agah [0030] The autonomous LLM agent 102 is a computer program that may be instantiated as one or more computer processes. In certain examples, each instance of the autonomous LLM agent 102 may be executed within its own container (e.g., a docker container or the like). In other examples, separate instances of the autonomous LLM agent 102 may be instantiated for each query that is being processed. The autonomous LLM agent 102 may act as a controller or the like that processes a query submitted by a user, determines a workflow to execute (e.g., a dynamically generated and/or optimized set of states for the workflow), and then manages the execution of that workflow. In certain examples, the autonomous LLM agent 102 is configured to parse a given query, dynamically generate a task list that includes one or more states, and then execute tasks within that list using one or more sub-agents that are associated with that state. [0048] Web search & parsing agent 126 is used to interface with the Internet and search engines. Web search & parsing agent 126 may be used to validate datapoints or retrieve, for example, information on a company or organization from websites and the like. [31-37] elaborates on the matter [FIG.1] shows corresponding visual) generating a first prompt for the first AI agent, the first prompt including at least a portion of the first query; providing the first prompt to the first AI agent; (Agah [0016] FIG. 7 is an example of a prompt configuration file that may be generated by the Prompt Config File Generation Module of FIG. 1 according to certain example embodiments;[0017] FIG. 8 is a flowchart of a process in which the Prompt Config File Execution Module of FIG. 1 executes a generated prompt configuration file according to certain example embodiments; [0025] a prompt configuration file that may be generated by the Prompt Config File Generation Module of FIG. 1 and FIG. 8 illustrates a process in which the Prompt Config File Execution Module of FIG. 1 executes a generated prompt configuration file. FIG. 9 is a flowchart of a process for automatically processing documents and generating contextual data that may be displayed as part of the illustrative graphical user interfaces in FIGS. 10A-10C. And FIG. 11 shows an example computing device that may be used in some embodiments, such as FIG. 1, to implement features described herein.[0033] Large language models (LLMs) 104 are used by the system 100 to extract information, via generated prompts, from documents, text, or other electronically stored data. For example, a prompt may be submitted by the autonomous LLM agent 102 to an LLM 104 to generate or find where a particular fact (e.g., a data item) is located within one or more documents. As discussed elsewhere herein, the prompts may be automatically generated based on the query...[0077] In certain example embodiments, for each determined state in the workflow, the process constructs a prompt to select one (or more) of the plurality of sub-agents to use in carrying out the task for that state [125-131] elaborates on the matter) receiving a first response from the first AI agent… and presenting the first response in the user interface. (Agah [FIG.1] shows corresponding visual on the response/output from the ai agent [0025] FIG. 1 is an architecture diagram of the example systems used in connection with certain example embodiments, this includes the computer system that is configured to receive and process a query and generate a responsive output. FIG. 2 is a flowchart of a process for generating an agent configuration file that includes state and tool definitions that are used by the system of FIG. 1. FIGS. 3A-3B are examples of state and tool definitions that may be created for the agent configuration file discussed in FIG. 2. FIG. 4[0030] The autonomous LLM agent 102 is a computer program that may be instantiated as one or more computer processes. In certain examples, each instance of the autonomous LLM agent 102 may be executed within its own container (e.g., a docker container or the like). In other examples, separate instances of the autonomous LLM agent 102 may be instantiated for each query that is being processed. The autonomous LLM agent 102 may act as a controller or the like that processes a query submitted by a user, determines a workflow to execute (e.g., a dynamically generated and/or optimized set of states for the workflow), and then manages the execution of that workflow. In certain examples, the autonomous LLM agent 102 is configured to parse a given query, dynamically generate a task list that includes one or more states, and then execute tasks within that list using one or more sub-agents that are associated with that state. [0038] the autonomous LLM agent 102. As an illustrative example, the application program 138 may generate a web page that provides users with the ability to submit a query. As discussed below in connection with the various examples, application program 138 may provide responsive output that may be presented to the user that submitted the query. In certain example embodiments, the application server 108 may also be used to generate webpages (or other graphical user interfaces) that are communicated and displayed on client devices 110. Examples of different webpages that may be delivered to and displayed on client device are shown in FIGS. 10A-10C.[0042] In some examples, a sub-agent may leverage or include a dynamic hybrid RAG pipeline. This functionality may include an adaptive weighting technique that fine-tunes the emphasis on keyword(s) and semantic search based on query specifics. It then may combine these results using a flexible aggregation method influenced by the dynamic weights. This approach allows for tailored query handling and result scoring, leading to enhanced content relevancy and precision in subsequent processing stages. Examples of tools that may be incorporate or use such functionality...[69-74] elaborate on the response/output from the ai agent) Agah lacks explicitly and orderly teaching wherein the first response includes information from the first container and excludes information from files that are excluded from the first container; However Bosarge teaches wherein the first response includes information from the first container and excludes information from files that are excluded from the first container; (Bosarge [0008] Accordingly, in view of the foregoing and other problems associated with conventional Internet browsers, there is an ongoing need and desire for improved browsers that can facilitate access to and navigation of Internet search results [0034] In view of the foregoing and subsequent disclosure, it will be appreciated that the disclosed invention provides many technical benefits, including the preservation of browsing status while refraining from caching irrelevant content, the automatic grouping and organization of search results based on search term context within corresponding containers, and the ability to navigate quickly between search results of existing and previous queries.[0043] When the search results are rendered for a particular search, according to the disclosed embodiments, they are grouped into a container and the search result content and tabbed pages are cached for easy access. Each search result webpage is associated with a different tab that is automatically created by the browser. In this manner, the user does not need to individually open a separate tab for each search result. The different tabs are automatically created and rendered on the browser within a tab view bar 410 [60-70] further elaborate on the matter [FIG.15-18] show wherein the first response includes information from the first container and excludes information from files that are excluded from the first container) Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to take all prior methods and make the addition of Bosarge in order to facilitate and improve the overall search/output results (Bosarge [AB] Systems, methods, and devices include browsers for facilitating and preserving query seminality in web browsing. Navigation between different tabbed content is facilitated by corresponding query containers that group tabbed search result content and that monitor and maintain the browsing state of the different search results. [0010] For instance, in some embodiments, systems are configured with browsers that utilize query containers and corresponding container icons for navigating search results that are grouped within respective containers as tabbed search results. Each search query corresponds to a different query container/container icon and a different corresponding set of tabbed search results for that search query. These query containers can help facilitate the management of different browsing session states and the navigation to prior and existing search results, as described herein. [60-70] further elaborate on the matter [FIG.15-18] show wherein the first response includes information from the first container and excludes information from files that are excluded from the first container) the combination lack explicitly and orderly teaching selecting, based on the indication of the first container, the first Al agent generating, based on the selection of the first container, However Joshi teaches selecting, based on the indication of the first container, the first Al agent generating, based on the selection of the first container (Joshi [0008] multi-container application, selecting, with one or more processors, a plurality of infrastructure or application performance monitoring agents based on the composition record defining the multi-container application; causing, with one or more processors, the selected agents to be deployed on one or more computing devices executing the multi-container application; receiving, with one or more processors, metrics or events from the agents indicative [0072] the agent deployment module 50 may be operative to receive the selected agents (or agent specified by the user) and the configurations from the agent configure 48 (or default configurations) and cause the corresponding agents to be deployed. Causing deployment may include sending an instruction to the application monitor 16 or to the container manager 20 that causes the selected agents to be deployed, the configurations to be applied, or configurations to be changed on existing agents. Causing deployment does not require that the entity causing deployment itself implement the deployment [0073] the multi-container application may include containers or services therein that are not identified by the agent selector 46 or manually by the user, but which a user may still wish to monitor [113] when executed, implement respective services of the multi-container application, selecting, with one or more processors, a plurality of infrastructure or application performance monitoring agents based on the composition record defining the multi-container application; causing, with one or more processors, the selected agents to be deployed on one or more computing devices executing the multi-container application; receiving, with one or more processors, metrics or events from the agents indicative of performance of at least part of the multi-container application or at least some of the one or more computing devices executing the multi-container application [115-124] further elaborate on the matter [FIG.1] shows the corresponding system which is capable of selecting, based on the indication of the first container, the first Al agent generating, based on the selection of the first container) Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to take all prior methods and make the addition of Joshi in order to create a more efficient system via container applications manipulations (Joshi [0008] the selected agents to be deployed on one or more computing devices executing the multi-container application; receiving, with one or more processors, metrics or events from the agents indicative of performance of at least part of the multi-container application or at least some of the one or more computing devices executing the multi-container application; and causing, with one or more processors, an indication of the received metrics or events to be presented [0027] As discussed, distributed applications are often relatively complex and difficult for developers and operations engineers to reason about. To help make these applications more manageable, often monitoring applications are installed alongside the distributed application to gather information about the underlying computers upon which the distributed application is executing or performance of application components [FIG.1] shows the corresponding system which is capable of selecting, based on the indication of the first container, the first Al agent generating, based on the selection of the first container) the combination lack explicitly and orderly teaching wherein the first response is generated based on a set of grounding data identified by executing a grounding query over the first set of files of the first container However Yee teaches wherein the first response is generated based on a set of grounding data identified by executing a grounding query over the first set of files of the first container (Yee [0019] support techniques for using organization-specific metadata to “ground” responses and content returned by generative AI applications. As described herein, grounding refers to the processing of making a prompt (e.g., a request or instruction) more specific, clear, and unambiguous so the generative AI application or LLM can generate more accurate and contextually relevant responses. Additionally or alternatively, the process of grounding may include providing the LLM with a corpus of data from which it is to generate the content of a response (as opposed to allowing the LLM to rely on the vast amount of training data to derive the actual content of the response). The process of grounding involves providing additional context or details to inform the LLM's understanding of the requested task/query and/or the universe of information from which the LLM may draw content. Examples of grounding include adding contextual information to the prompt, defining the expected input/output format, instructing the model to avoid certain terminology, etc. Using organization or user-specific metadata for prompt grounding may improve the accuracy, coherence, and consistency of responses provided by generative AI applications [0034] For example, a user associated with a tenant may submit a query to the cloud platform 115 or other system to generate a response to the query using generative AI, such as an LLM. In some examples, the user may authenticate or provide proof of an identity for identify verification or authentication of access to information to be used for grounding the generative AI prompt. The cloud platform 115 may process the query and may leverage metadata (e.g., configured or indicated by an administrator) that indicates documents, files, or information to which the tenant may have access. The cloud platform 115 may have previously indexed or otherwise made record of such documents, files, or information using vectorization or embedding techniques to allow for a comparison between the query and the documents, files, or information to locate relevant portions to be included in a prompt for generative AI. ....[33-41] elaborate on the matter [FIG.2&4] elaborate on the system which includes teaching the steps for performing wherein the first response is generated based on a set of grounding data identified by executing a grounding query over the first set of files of the first container ) Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to take all prior methods and make the addition of Yee in order to improve the accuracy, coherence, and consistency of the system output (Yee [0019] Aspects of the present disclosure support techniques for using organization-specific metadata to “ground” responses and content returned by generative AI applications. As described herein, grounding refers to the processing of making a prompt (e.g., a request or instruction) more specific, clear, and unambiguous so the generative AI application or LLM can generate more accurate and contextually relevant responses. Additionally or alternatively, the process of grounding may include providing the LLM with a corpus of data from which it is to generate the content of a response (as opposed to allowing the LLM to rely on the vast amount of training data to derive the actual content of the response). The process of grounding involves providing additional context or details to inform the LLM's understanding of the requested task/query and/or the universe of information from which the LLM may draw content. Examples of grounding include adding contextual information to the prompt, defining the expected input/output format, instructing the model to avoid certain terminology, etc. Using organization or user-specific metadata for prompt grounding may improve the accuracy, coherence, and consistency of responses provided by generative AI applications.)
Corresponding system claim 15 is rejected similarly as claim 1 above. Additional Limitations: Device with processor(s) and memory (Agah [FIG.1 in conjunction with Fig.11] show Device with processor(s) and memory)
Regarding claim 3, Agah, Bosarge, Joshi and Yee teach The method of claim 1, further comprising: receiving, via the user interface, a second query and an indication of the second container; (Bosarge [0003] It is typical for search engines to utilize tabs to facilitate access to different search results that correspond to a single query or to different queries. For instance, a user can select a control for generating a new tab (e.g., the “+” icon within the tab view bar of the browser) to launch a new tab from which the user can perform a new search to navigate new content. Likewise, a user can use menu controls to trigger the generation of a new tab corresponding to a selected term or link from within one webpage to spawn an additional tabbed webpage. The different tabs within the browser window are selectable to enable a user to quickly navigate to the content of a corresponding tab is selected. The use of such tabs are well-known to those of skill in the art.[0036] In this illustration, a user can enter a search term or URL into a URL field 110 or other type of query field 120. Results of the search can be rendered as webpages (such as webpage 130). When a search is performed, the new search result content will replace the content shown in webpage 130. A tab view bar 140 contains a tab 150 that is associated with the currently displayed content and the search results that will be displayed. Once the web content is updated, the tab will also be updated to reflect the corresponding association.[0040] As shown, a browser window 210 is rendered with a query field (e.g., URL field 210 and/or a search term/URL field 220). When a user enters one or more search terms into the query field, the system will perform a search for related web content that is accessible through the Internet or other network databases corresponding to the entered search term(s). Notably, the search terms entered may comprises a single set of search terms for a single search or a ballistic multi-query that includes two or more sets of search terms for a plurality of respective searches [60-69] further elaborates on the matter [FIG.14] shows corresponding visual) generating a second prompt for the second AI agent, the second prompt including at least a portion of the second query; providing the second prompt to the second AI agent; (Agah [0043] trained LLM with domain specific or up-to-date knowledge for when the LLM generates a response to one or more prompts. [0071] At 402, the query is processed to generate a workflow that includes a list of one or more of the pre-defined states (e.g., as discussed in connection with FIG. 2). The query that is generated is modular in that the states that makeup any given query can change based on the nature of the query. The processing of the query may include submitting one or more prompts to LLM 104 to determine which states are applicable for the query that is received at 400 [0072] Determination of which states are applicable for a workflow for a query that has been received may include generating and submitting one or more prompts to LLM 104. A first prompt may be used to analyze the intent and goals of the query. An illustrative example prompt that is submitted to an LLM may include “Classify the intent and goals of this query: [query text of the submitted query].” The responsive answer from the LLM 104 may be (for the example query): “Intent: Extract metrics. Goal: Retrieve latest report and extract sustainability metrics.[0073] With an intent and goal of the query determined, the process then determines which states, as defined in the agent configuration file (e.g., 300 in FIG. 3A), are applicable to the query. To determine which states are applicable, the process constructs one or more prompts and submits those prompt(s) ... [99-108] elaborate on the matter [FIG.1] shows generating a second prompt for the second AI agent, the second prompt including at least a portion of the second query; providing the second prompt to the second AI agent) receiving a second response from the second AI agent, wherein the response includes information from the second container and excludes information from files that are excluded from the first container; (Bosarge [0008] Accordingly, in view of the foregoing and other problems associated with conventional Internet browsers, there is an ongoing need and desire for improved browsers that can facilitate access to and navigation of Internet search results [0034] In view of the foregoing and subsequent disclosure, it will be appreciated that the disclosed invention provides many technical benefits, including the preservation of browsing status while refraining from caching irrelevant content, the automatic grouping and organization of search results based on search term context within corresponding containers, and the ability to navigate quickly between search results of existing and previous queries.[0043] When the search results are rendered for a particular search, according to the disclosed embodiments, they are grouped into a container and the search result content and tabbed pages are cached for easy access. Each search result webpage is associated with a different tab that is automatically created by the browser. In this manner, the user does not need to individually open a separate tab for each search result. The different tabs are automatically created and rendered on the browser within a tab view bar 410 [60-70] further elaborate on receiving a second response from the second AI agent, wherein the response includes information from the second container and excludes information from files that are excluded from the first container [FIG.15-18] show receiving a second response from the second AI agent, wherein the response includes information from the second container and excludes information from files that are excluded from the first container) and displaying the second response in the user interface. (Agah [FIG.1] shows corresponding visual on the response/output from the ai agent [0025] FIG. 1 is an architecture diagram of the example systems used in connection with certain example embodiments, this includes the computer system that is configured to receive and process a query and generate a responsive output. FIG. 2 is a flowchart of a process for generating an agent configuration file that includes state and tool definitions that are used by the system of FIG. 1. FIGS. 3A-3B are examples of state and tool definitions that may be created for the agent configuration file discussed in FIG. 2. FIG. 4. [0038] the application server 108 may also be used to generate webpages (or other graphical user interfaces) that are communicated and displayed on client devices 110. Examples of different webpages that may be delivered to and displayed on client device are shown in FIGS. 10A-10C.[0042] In some examples, a sub-agent may leverage or include a dynamic hybrid RAG pipeline. This functionality may include an adaptive weighting technique that fine-tunes the emphasis on keyword(s) and semantic search based on query specifics. It then may combine these results using a flexible aggregation method influenced by the dynamic weights. This approach allows for tailored query handling and result scoring, leading to enhanced content relevancy and precision in subsequent processing stages. Examples of tools that may be incorporate or use such functionality...[69-74] elaborate on the response/output from the ai agent)
Corresponding system claim 17 is rejected similarly as claim 3 above.
Regarding claim 4, Agah, Bosarge, Joshi and Yee teach The method of claim 1, further comprising the combination lacks explicitly and orderly teaching receiving a request to delete the first container or the first AI agent; and in response to the request, deleting the first container and the first AI agent. However Joshi teaches receiving a request to delete the first container or the first AI agent; and in response to the request, deleting the first container and the first AI agent. (Joshi [0056] Some embodiments of the container manager 20 may further be configured to determine when containers have ceased to operate, are operating at greater than a threshold capacity, or are operating at less than a threshold capacity, and take responsive action, for instance by terminating containers that are underused, re-instantiating containers that have crashed, and adding additional instances of containers that are at greater than a threshold capacity. Some embodiments of the container manager 20 may further be configured to deploy new versions of images of containers, for instance, to rollout updates or revisions to application code. Some embodiments may be configured to roll back to a previous version responsive to a failed version or a user command. In some embodiments, the container manager 20 may facilitate discovery of other services within a multi-container application, for instance, indicating to one service executing in one container where and how to communicate with another service executing in other containers, like indicating to a web server service an Internet Protocol address of a database management service used by the web server service to formulate a response to a webpage request. In some cases, these other services may be on the same computing device and accessed via a loopback address or on other computing devices.[55-58] elaborate on the matter [FIG.1] shows corresponding visual) Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to take all prior methods and make the addition of Joshi in order to create a more efficient system via delete of unnecessary containers (Joshi [0056] Some embodiments of the container manager 20 may further be configured to determine when containers have ceased to operate, are operating at greater than a threshold capacity, or are operating at less than a threshold capacity, and take responsive action, for instance by terminating containers that are underused, re-instantiating containers that have crashed, and adding additional instances of containers that are at greater than a threshold capacity. Some embodiments of the container manager 20 may further be configured to deploy new versions of images of containers, for instance, to rollout updates or revisions to application code. Some embodiments may be configured to roll back to a previous version responsive to a failed version or a user command. In some embodiments, the container manager 20 may facilitate discovery of other services within a multi-container application, for instance, indicating to one service executing in one container where and how to communicate with another service executing in other containers, like indicating to a web server service an Internet Protocol address of a database management service used by the web server service to formulate a response to a webpage request. In some cases, these other services may be on the same computing device and accessed via a loopback address or on other computing devices.[55-58] elaborate on the matter [FIG.1] shows corresponding visual)
Corresponding system claim 18 is rejected similarly as claim 4 above.
Regarding claim 5, Agah, Bosarge, Joshi and Yee teach The method of claim 1, wherein the first container comprises a first custom prompt that configures the first AI agent to respond to prompts, the method further comprising: providing the first custom prompt to the first AI agent before providing the first prompt to the first AI agent, wherein the first custom prompt configures the first AI agent to generate a response to the first prompt. (Agah [0040] Turning now more specifically to the autonomous LLM agent 102 and the sub-agents thereof. The autonomous LLM agent 102 includes or has access to sub-agents that may be individually selected and used based on the particular nature of the query being processed by the autonomous LLM agent 102. The sub-agents of the autonomous LLM agent 102 can be thought of as individual tools that the agent 102 can employ in connection with the dynamically constructed workflow. Accordingly, the autonomous LLM agent 102 can be thought of as a multi-tool agent in certain examples. The sub-agents include any or all of the following: 1) a document disclosure RAG (retrieval-augmented generation) tool 120; 2) a regulatory RAG tool 122; 3) a tabular data agent 124; 4) a web search & parsing agent 126; 5) a report segment composer sub-agent agent 128 (which may also be a database query agent in some examples); 6) a report retrieval & conversion tool agent 130 (which may also be a PDF extraction agent in some examples); and 7) a prompt engineering pipeline module 132. Other sub-agents may also be included or be accessed by the autonomous LLM agent 102 in connection with processing a query and the workflow generated therefrom. [0053] Prompt engineering pipeline module 132 includes two sub-components: 1) Prompt Config File Execution Module 134; and 2) Prompt Config File Generation Module 136. The Prompt Config File Generation Module 136 is configured to generate a prompt config file that is then executed by the Prompt Config File Execution Module 134. Details of these two modules are discussed, among other places, in FIGS. 7 and 8 and the illustrative examples discussed below [0064] In certain example embodiments, each tool that is defined in 350 may have a corresponding sub-agent module that implements that processing associated with that tool. Thus, for example, the “WebSearch Sub-Agent” tool that is defined in FIG. 3B may correspond or reference Web Search & Parsing Agent 126 that is discussed in connection with FIG. 1. Similarly, the “PromptEngineeringPipeline” may correspond to the Prompt Engineering Pipeline Module...[89&170] further elaborate [FIG.1] shows overall visual of the system)
Corresponding system claim 19 is rejected similarly as claim 5 above.
Claims 2 and 16 are rejected under 35 U.S.C. 103 as being unpatentable over Agah in view of Bosarge, Joshi, Yee and US 20210232632 A1; HOWARD; Todd A (hereinafter Howard).
Regarding claim 2, Agah, Bosarge, Joshi and Yee teach The method of claim 1, the combination lacks explicitly and orderly teaching wherein the first container is a first file folder and associating the first AI agent with the first container comprises linking the first AI agent with the first file folder. However Howard teaches wherein the first container is a first file folder and associating the first AI agent with the first container comprises linking the first AI agent with the first file folder. (Howard [0215] Returning now to FIG. 4A, the virtual experience container may be provided to the user experience device (404). A virtual experience container, such as the one depicted in the example structure in FIG. 5, may be arranged in a variety of ways. For example, it may be arranged as a multi-dimensional array containing both binary data (e.g., content segment streams in their native formats) and metadata arranged in time series according to the direction of flow of the virtual experience container playback. Each dimension may approximate a “layer” as described in FIG. 5, with layers operating in parallel to render content or provide instructions depending on the nature of the layer. A virtual experience container may contain pointers to outside content segments, e.g., using embedded URLs to initiate streaming of content from external services. Other arrangements may also be used, such an XML metadata file that stores properties and the direction of flow, along with embedded binary large object (BLOB) data or pointers to separately stored BLOB data. A virtual experience container may be a proprietary file format which may embed content, reference content, stream content, or any combination thereof. It is contemplated that, while used in the singular form herein, a virtual experience container can comprise one or more physical files, temporary files, storages, and/or dynamically-generated streams, which may in fact be stored on different computer-readable storage media across one or more services or systems. Depending on the embodiment, the nature of the virtual experience, and the user experience device parameters, a virtual experience container may be provided in a number of ways, such as by a file or stream download to the user experience device (including a file that can begin to be “played” before the download is complete). A virtual experience container may also be streamed by being downloaded into player software that buffers content in a temporary cache in advance of playback. Other variations of arrangement still within the scope of described embodiments may suggest themselves to the ordinarily skilled practitioner. [Fig.1] shows corresponding visual)
Corresponding system claim 16 is rejected similarly as claim 2 above.
Claims 6-7 and 20 are rejected under 35 U.S.C. 103 as being unpatentable over US Agah in view of Bosarge, Joshi, Yee and US 20180329990 A1; SEVERN; Robert et al. (hereinafter Severn).
Regarding claim 6, Agah, Bosarge, Joshi and Yee teach The method of claim 5, the combination lacks explicitly and orderly teaching receiving a request to add a third file to the first container; in response to receiving the request to add the third file to the first container: adding the third file to the first container; generating a second set of embeddings based on the third file; and adding the second set of embeddings to the first container. However Severn teaches receiving a request to add a third file to the first container; in response to receiving the request to add the third file to the first container: adding the third file to the first container; (Severn [0053] The query processing module 140 interprets the queries as geometric constraints on the embedding space, and narrows or otherwise modifies a catalog of documents obtained from the embedding space to develop a set of candidate documents which satisfy the geometric constraints. These candidate documents are written into a candidate space database 150. Candidate spaces as used herein are also embedding spaces, and for example may constitute a portion of the embedding space of the document catalog database 120. [0517] In another implementation, the implicit user action can be one or more of (i) requesting a more detailed description of the document, (ii) requesting the document to be added to a cart, (iii) requesting the document to be added to a list (iv) the user lingering on a view of the document for more than a predetermined amount of time and (v) the user lingering on a view of the document for less than another predetermined amount of time. Further, the meaning assigned to the implicit user action is weighted in dependence on whether the implicit user action is one or more of (i), (ii), (iii), (iv) and (v). [0486] In operation 2818 a meaning is assigned to the user action. This meaning can be any of the types of user interactions described above. A score or value can then be determined for a document based on various types of implicit user actions associated with the document. A weighting can also be assigned based on the type of implicit user action. For example, opening up a product display page could be given a 0.5 weighting, adding a document to a wish list, a favorites list or the cart could be given a 1.0 weighting, lingering on a product in a carousel of displayed products or extending a mouse hover can be given an initial weighting of 0.1 which could increase over time, scrolling past a document, which would be similar to disliking a document, could be given a rating of 0. [485-487] elaborate on the matter [FIG.1] shows corresponding flow ) generating a second set of embeddings based on the third file; and adding the second set of embeddings to the first container. (Severn [0053] Referring to FIG. 1, a block diagram 100 of a visual interactive search system includes an embedding module 110 which calculates an embedding of source documents into an embedding space, and writes embedding information, in association with an identification of the documents, into a document catalog database (e.g., document catalog) 120. A user interaction module 130 receives queries and query refinement input (such as relevance feedback) from a user, and provides the received queries and query refinement input to a query processing module 140. In an implementation, the user interaction module 130 includes a computer terminal, whereas in another implementation the user interaction module 130 includes only certain network connection components through which the system communicates with an external computer terminal. The query processing module 140 interprets the queries as geometric constraints on the embedding space, and narrows or otherwise modifies a catalog of documents obtained from the embedding space to develop a set of candidate documents which satisfy the geometric constraints. These candidate documents are written into a candidate space database 150. Candidate spaces as used herein are also embedding spaces, and for example may constitute a portion of the embedding space of the document catalog database 120. [0059] Alternatively, the embedding module 110 may derive a library of image classifications (axes on which a given photograph may be placed), each in association with an algorithm for recognizing in a given photograph whether (or with what probability) the given photograph satisfies that classification. Then the embedding module 110 may apply its pre-developed library to a smaller set of newly provided photographs, such as the photos currently on the user computer 210, in order to determine embedding information applicable to each photograph. Either way, the embedding module 110 writes into the document catalog database 120 the identifications of the catalog of documents that the user may search, each in association with the corresponding embedding information.[0342] As can be seen, identification of a desired document may include providing, accessibly to a computer system, a database identifying a catalog of documents in an embedding space, calculating... [485-487] elaborate on the matter [FIG.1] shows corresponding flow ) Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to take all prior methods and make the addition of Severn in order to use embedding module methods in order to efficiently organize the system (Severn [0005] Some libraries are annotated with metadata, such as date and location (for a photo library), or type and features of products (for a product catalog). But many are not annotated, and even those which are may not be sufficiently specific to allow the server to efficiently hone in on the desired document quickly. As touched on above, some search technologies allow the server to perform searches iteratively, thereby gradually narrowing the field of possible documents until the target document is found. But these often still take a long time, and cause the server to consume an unnecessarily high level of bandwidth and processing power by requiring the server to offer many different collections of candidate documents before the target document is found. [0009] The new results provided to the user are driven by the “likes,” “dislikes” or “neutral” opinions provided by the user. As a result of implementing these features, a conventional computer system (e.g., the server and/or user device providing the results) is improved because the user is able to get to what they are looking for faster, with fewer iterations and more confidence. This is achievable because the computer system implementing these features is able to intelligently display a limited set of information to the user by gathering the information mentioned above and summarizing the user's interests to provide a limited set of valuable information to the user. [11-13] elaborate on the matter)
Corresponding system claim 20 is rejected similarly as claim 6 above.
Regarding claim 7, Agah, Bosarge, Joshi and Yee teach The method of claim 6, wherein the second container includes a second custom prompt that configures the second AI agent to respond to prompts, the method further comprising: providing the second custom prompt to the first AI agent; (Agah [0043] trained LLM with domain specific or up-to-date knowledge for when the LLM generates a response to one or more prompts. [0071] At 402, the query is processed to generate a workflow that includes a list of one or more of the pre-defined states (e.g., as discussed in connection with FIG. 2). The query that is generated is modular in that the states that makeup any given query can change based on the nature of the query. The processing of the query may include submitting one or more prompts to LLM 104 to determine which states are applicable for the query that is received at 400 [0072] Determination of which states are applicable for a workflow for a query that has been received may include generating and submitting one or more prompts to LLM 104. A first prompt may be used to analyze the intent and goals of the query. An illustrative example prompt that is submitted to an LLM may include “Classify the intent and goals of this query: [query text of the submitted query].” The responsive answer from the LLM 104 may be (for the example query): “Intent: Extract metrics. Goal: Retrieve latest report and extract sustainability metrics.[0073] With an intent and goal of the query determined, the process then determines which states, as defined in the agent configuration file (e.g., 300 in FIG. 3A), are applicable to the query. To determine which states are applicable, the process constructs one or more prompts and submits those prompt(s) ... [99-108] elaborate on the matter [FIG.1] show corresponding visual) Figure 1 and corresponding paragraphs in Agah show that there can be multiple prompts and queries created and utilized receiving, via the user interface, a third query and an indication of the first container; (Bosarge [0003] It is typical for search engines to utilize tabs to facilitate access to different search results that correspond to a single query or to different queries. For instance, a user can select a control for generating a new tab (e.g., the “+” icon within the tab view bar of the browser) to launch a new tab from which the user can perform a new search to navigate new content. Likewise, a user can use menu controls to trigger the generation of a new tab corresponding to a selected term or link from within one webpage to spawn an additional tabbed webpage. The different tabs within the browser window are selectable to enable a user to quickly navigate to the content of a corresponding tab is selected. The use of such tabs are well-known to those of skill in the art.[0036] In this illustration, a user can enter a search term or URL into a URL field 110 or other type of query field 120. Results of the search can be rendered as webpages (such as webpage 130). When a search is performed, the new search result content will replace the content shown in webpage 130. A tab view bar 140 contains a tab 150 that is associated with the currently displayed content and the search results that will be displayed. Once the web content is updated, the tab will also be updated to reflect the corresponding association.[0040] As shown, a browser window 210 is rendered with a query field (e.g., URL field 210 and/or a search term/URL field 220). When a user enters one or more search terms into the query field, the system will perform a search for related web content that is accessible through the Internet or other network databases corresponding to the entered search term(s). Notably, the search terms entered may comprises a single set of search terms for a single search or a ballistic multi-query that includes two or more sets of search terms for a plurality of respective searches [60-69] further elaborates on the matter [FIG.14] shows corresponding visual) generating a third prompt for the first AI agent, the third prompt including at least a portion of the third query; providing the third prompt to the first AI agent; (Agah [0043] trained LLM with domain specific or up-to-date knowledge for when the LLM generates a response to one or more prompts. [0071] At 402, the query is processed to generate a workflow that includes a list of one or more of the pre-defined states (e.g., as discussed in connection with FIG. 2). The query that is generated is modular in that the states that makeup any given query can change based on the nature of the query. The processing of the query may include submitting one or more prompts to LLM 104 to determine which states are applicable for the query that is received at 400 [0072] Determination of which states are applicable for a workflow for a query that has been received may include generating and submitting one or more prompts to LLM 104. A first prompt may be used to analyze the intent and goals of the query. An illustrative example prompt that is submitted to an LLM may include “Classify the intent and goals of this query: [query text of the submitted query].” The responsive answer from the LLM 104 may be (for the example query): “Intent: Extract metrics. Goal: Retrieve latest report and extract sustainability metrics.[0073] With an intent and goal of the query determined, the process then determines which states, as defined in the agent configuration file (e.g., 300 in FIG. 3A), are applicable to the query. To determine which states are applicable, the process constructs one or more prompts and submits those prompt(s) ... [99-108] elaborate on the matter [FIG.1] show corresponding visual) Figure 1 and corresponding paragraphs in Agah show that there can be multiple prompts and queries created and utilized receiving a third response from the first AI agent, wherein the third response includes information from the third file and excludes information in files that are excluded from the first container. (Bosarge [0008] Accordingly, in view of the foregoing and other problems associated with conventional Internet browsers, there is an ongoing need and desire for improved browsers that can facilitate access to and navigation of Internet search results [0034] In view of the foregoing and subsequent disclosure, it will be appreciated that the disclosed invention provides many technical benefits, including the preservation of browsing status while refraining from caching irrelevant content, the automatic grouping and organization of search results based on search term context within corresponding containers, and the ability to navigate quickly between search results of existing and previous queries.[0043] When the search results are rendered for a particular search, according to the disclosed embodiments, they are grouped into a container and the search result content and tabbed pages are cached for easy access. Each search result webpage is associated with a different tab that is automatically created by the browser. In this manner, the user does not need to individually open a separate tab for each search result. The different tabs are automatically created and rendered on the browser within a tab view bar 410 [60-70] further elaborate on the matter [FIG.15-18] show receiving a third response from the first AI agent, wherein the third response includes information from the third file and excludes information in files that are excluded from the first container)
Claims 8-9 and 12 are rejected under 35 U.S.C. 103 as being unpatentable over US 20250117410 A1; AGHAJANYAN; Viktor et al. (hereinafter Agah) in view of US 20250373574 A1; Deutsch; Noah (hereinafter Deutsch).
Regarding Claim 8, Agah teaches A method, comprising: associating a first AI agent with a first container comprising: a first set of files and a first set of embeddings, and a second container comprising a nested AI agent, a second set of files, and a second set of embeddings, (Agah [0030] The autonomous LLM agent 102 is a computer program that may be instantiated as one or more computer processes. In certain examples, each instance of the autonomous LLM agent 102 may be executed within its own container (e.g., a docker container or the like). In other examples, separate instances of the autonomous LLM agent 102 may be instantiated for each query that is being processed. The autonomous LLM agent 102 may act as a controller or the like that processes a query submitted by a user, determines a workflow to execute (e.g., a dynamically generated and/or optimized set of states for the workflow), and then manages the execution of that workflow. In certain examples, the autonomous LLM agent 102 is configured to parse a given query, dynamically generate a task list that includes one or more states, and then execute tasks within that list using one or more sub-agents that are associated with that state.[0032] The autonomous LLM agent 102 can include multiple different sub-agents (which may also be called agents or tools herein) that can be individually used in connection with processing a query. The different sub-agents that the autonomous LLM agent 102 may use are discussed below...[0040] Turning now more specifically to the autonomous LLM agent 102 and the sub-agents thereof. The autonomous LLM agent 102 includes or has access to sub-agents that may be individually selected and used based on the particular nature of the query being processed by the autonomous LLM agent 102. The sub-agents of the autonomous LLM agent 102 can be thought of as individual tools that the agent 102 can employ in connection with the dynamically constructed workflow. Accordingly, the autonomous LLM agent 102 can be thought of as a multi-tool agent in certain examples. The sub-agents include any or all of the following:..[0099] The prompt configuration file 700 includes a prompt pipeline that is composed of a plurality of sequenced prompts 706 (e.g., “prompt 1”, “prompt 2”, “prompt 3”, etc.). Each of these prompts includes one or more templates 708. Each of the items in brackets within the templates (e.g., “{disclosure_item}”, “{unit_of_measurement1}”, etc.) are variables that will be defined when the prompt is executed by the Prompt Engineering Pipeline Module 132.[0104] Next, at 802, a chain linking prompt is used to dive deeper into the data that is being extracted by the LLM. For this 3 concurrent prompts are generated and submitted at 804, 806, and 808. [55-64] elaborates on the different agents with different data storage/containers and corresponding different data/embeddings [FIG.1] shows corresponding visual) and receiving, via a user interface, a query for the first AI agent; (Agah [0030] The autonomous LLM agent 102 is a computer program that may be instantiated as one or more computer processes. In certain examples, each instance of the autonomous LLM agent 102 may be executed within its own container (e.g., a docker container or the like). In other examples, separate instances of the autonomous LLM agent 102 may be instantiated for each query that is being processed. The autonomous LLM agent 102 may act as a controller or the like that processes a query submitted by a user, determines a workflow to execute (e.g., a dynamically generated and/or optimized set of states for the workflow), and then manages the execution of that workflow. In certain examples, the autonomous LLM agent 102 is configured to parse a given query, dynamically generate a task list that includes one or more states, and then execute tasks within that list using one or more sub-agents that are associated with that state. [0048] Web search & parsing agent 126 is used to interface with the Internet and search engines. Web search & parsing agent 126 may be used to validate datapoints or retrieve, for example, information on a company or organization from websites and the like. [31-37] elaborates on the matter [FIG.1] shows corresponding visual) generating a first prompt for the first AI agent, the first prompt including at least a portion of the query; generating a first prompt for the first AI agent, the first prompt including at least a portion of the query; providing the first prompt to the first AI agent; (Agah [0016] FIG. 7 is an example of a prompt configuration file that may be generated by the Prompt Config File Generation Module of FIG. 1 according to certain example embodiments;[0017] FIG. 8 is a flowchart of a process in which the Prompt Config File Execution Module of FIG. 1 executes a generated prompt configuration file according to certain example embodiments; [0025] a prompt configuration file that may be generated by the Prompt Config File Generation Module of FIG. 1 and FIG. 8 illustrates a process in which the Prompt Config File Execution Module of FIG. 1 executes a generated prompt configuration file. FIG. 9 is a flowchart of a process for automatically processing documents and generating contextual data that may be displayed as part of the illustrative graphical user interfaces in FIGS. 10A-10C. And FIG. 11 shows an example computing device that may be used in some embodiments, such as FIG. 1, to implement features described herein.[0033] Large language models (LLMs) 104 are used by the system 100 to extract information, via generated prompts, from documents, text, or other electronically stored data. For example, a prompt may be submitted by the autonomous LLM agent 102 to an LLM 104 to generate or find where a particular fact (e.g., a data item) is located within one or more documents. As discussed elsewhere herein, the prompts may be automatically generated based on the query...[0077] In certain example embodiments, for each determined state in the workflow, the process constructs a prompt to select one (or more) of the plurality of sub-agents to use in carrying out the task for that state [125-131] elaborates on the matter) receiving a first output from the first AI agent; (Agah [FIG.1] shows corresponding visual on the response/output from the ai agent [0025] FIG. 1 is an architecture diagram of the example systems used in connection with certain example embodiments, this includes the computer system that is configured to receive and process a query and generate a responsive output. FIG. 2 is a flowchart of a process for generating an agent configuration file that includes state and tool definitions that are used by the system of FIG. 1. FIGS. 3A-3B are examples of state and tool definitions that may be created for the agent configuration file discussed in FIG. 2. FIG. 4[0030] The autonomous LLM agent 102 is a computer program that may be instantiated as one or more computer processes. In certain examples, each instance of the autonomous LLM agent 102 may be executed within its own container (e.g., a docker container or the like). In other examples, separate instances of the autonomous LLM agent 102 may be instantiated for each query that is being processed. The autonomous LLM agent 102 may act as a controller or the like that processes a query submitted by a user, determines a workflow to execute (e.g., a dynamically generated and/or optimized set of states for the workflow), and then manages the execution of that workflow. In certain examples, the autonomous LLM agent 102 is configured to parse a given query, dynamically generate a task list that includes one or more states, and then execute tasks within that list using one or more sub-agents that are associated with that state. [0038] the autonomous LLM agent 102. As an illustrative example, the application program 138 may generate a web page that provides users with the ability to submit a query. As discussed below in connection with the various examples, application program 138 may provide responsive output that may be presented to the user that submitted the query. In certain example embodiments, the application server 108 may also be used to generate webpages (or other graphical user interfaces) that are communicated and displayed on client devices 110. Examples of different webpages that may be delivered to and displayed on client device are shown in FIGS. 10A-10C.[0042] In some examples, a sub-agent may leverage or include a dynamic hybrid RAG pipeline. This functionality may include an adaptive weighting technique that fine-tunes the emphasis on keyword(s) and semantic search based on query specifics. It then may combine these results using a flexible aggregation method influenced by the dynamic weights. This approach allows for tailored query handling and result scoring, leading to enhanced content relevancy and precision in subsequent processing stages. Examples of tools that may be incorporate or use such functionality...[69-74] elaborate on the response/output from the ai agent) analyzing the first output to determine the first output indicates additional data from the nested AI agent; (Agah [0099] The prompt configuration file 700 includes a prompt pipeline that is composed of a plurality of sequenced prompts 706 (e.g., “prompt 1”, “prompt 2”, “prompt 3”, etc.). Each of these prompts includes one or more templates 708. Each of the items in brackets within the templates (e.g., “{disclosure_item}”, “{unit_of_measurement1}”, etc.) are variables that will be defined when the prompt is executed by the Prompt Engineering Pipeline Module 132.[0100] It will be appreciated that different prompt engineering configurations can be developed depending on application need. The example discussed below in connection with FIG. 8 relates to scope 1 emissions. However, another prompt example that follows from the techniques herein is provided below in the context of insights into a Board of Directors Oversight of Climate related topics. Accordingly, the different prompt engineering configurations[0104] Next, at 802, a chain linking prompt is used to dive deeper into the data that is being extracted by the LLM. For this 3 concurrent prompts are generated and submitted at 804, 806, and 808.[0152] In certain example embodiments, a system is provided that allows for automatically processing documents, and extracting data from those documents, in a more efficient manner. The processing may be more efficient as the relevant contextual data is determined automatically based on, for example, the defined prompt pipeline configuration files. In certain examples, the configuration provided by the prompt pipeline provides more accurate results (e.g., a decreased error rate in comparison to other approaches, including manual review of the documents [FIG.1] show corresponding flow) generating a second prompt for the nested AI agent based on the first output from the first AI agent; providing the second prompt to the nested AI agent; (Agah [0043] trained LLM with domain specific or up-to-date knowledge for when the LLM generates a response to one or more prompts. [0071] At 402, the query is processed to generate a workflow that includes a list of one or more of the pre-defined states (e.g., as discussed in connection with FIG. 2). The query that is generated is modular in that the states that makeup any given query can change based on the nature of the query. The processing of the query may include submitting one or more prompts to LLM 104 to determine which states are applicable for the query that is received at 400 [0072] Determination of which states are applicable for a workflow for a query that has been received may include generating and submitting one or more prompts to LLM 104. A first prompt may be used to analyze the intent and goals of the query. An illustrative example prompt that is submitted to an LLM may include “Classify the intent and goals of this query: [query text of the submitted query].” The responsive answer from the LLM 104 may be (for the example query): “Intent: Extract metrics. Goal: Retrieve latest report and extract sustainability metrics.[0073] With an intent and goal of the query determined, the process then determines which states, as defined in the agent configuration file (e.g., 300 in FIG. 3A), are applicable to the query. To determine which states are applicable, the process constructs one or more prompts and submits those prompt(s) ... [99-108] elaborate on the matter [FIG.1] show corresponding visual) Figure 1 and corresponding paragraphs in Agah show that there can be multiple prompts and queries created and utilized receiving an output from the nested AI agent; (Agah [FIG.1] shows corresponding visual on the response/output from the ai agent [0025] FIG. 1 is an architecture diagram of the example systems used in connection with certain example embodiments, this includes the computer system that is configured to receive and process a query and generate a responsive output. FIG. 2 is a flowchart of a process for generating an agent configuration file that includes state and tool definitions that are used by the system of FIG. 1. FIGS. 3A-3B are examples of state and tool definitions that may be created for the agent configuration file discussed in FIG. 2. FIG. 4[0030] The autonomous LLM agent 102 is a computer program that may be instantiated as one or more computer processes. In certain examples, each instance of the autonomous LLM agent 102 may be executed within its own container (e.g., a docker container or the like). In other examples, separate instances of the autonomous LLM agent 102 may be instantiated for each query that is being processed. The autonomous LLM agent 102 may act as a controller or the like that processes a query submitted by a user, determines a workflow to execute (e.g., a dynamically generated and/or optimized set of states for the workflow), and then manages the execution of that workflow. In certain examples, the autonomous LLM agent 102 is configured to parse a given query, dynamically generate a task list that includes one or more states, and then execute tasks within that list using one or more sub-agents that are associated with that state. [0038] the autonomous LLM agent 102. As an illustrative example, the application program 138 may generate a web page that provides users with the ability to submit a query. As discussed below in connection with the various examples, application program 138 may provide responsive output that may be presented to the user that submitted the query. In certain example embodiments, the application server 108 may also be used to generate webpages (or other graphical user interfaces) that are communicated and displayed on client devices 110. Examples of different webpages that may be delivered to and displayed on client device are shown in FIGS. 10A-10C.[0042] In some examples, a sub-agent may leverage or include a dynamic hybrid RAG pipeline. This functionality may include an adaptive weighting technique that fine-tunes the emphasis on keyword(s) and semantic search based on query specifics. It then may combine these results using a flexible aggregation method influenced by the dynamic weights. This approach allows for tailored query handling and result scoring, leading to enhanced content relevancy and precision in subsequent processing stages. Examples of tools that may be incorporate or use such functionality...[69-74] elaborate on the response/output from the ai agent) generating a subsequent prompt for the first AI agent, wherein the subsequent prompt includes the output from the nested AI agent; providing the subsequent prompt to the first AI agent; (Agah [0043] trained LLM with domain specific or up-to-date knowledge for when the LLM generates a response to one or more prompts. [0071] At 402, the query is processed to generate a workflow that includes a list of one or more of the pre-defined states (e.g., as discussed in connection with FIG. 2). The query that is generated is modular in that the states that makeup any given query can change based on the nature of the query. The processing of the query may include submitting one or more prompts to LLM 104 to determine which states are applicable for the query that is received at 400 [0072] Determination of which states are applicable for a workflow for a query that has been received may include generating and submitting one or more prompts to LLM 104. A first prompt may be used to analyze the intent and goals of the query. An illustrative example prompt that is submitted to an LLM may include “Classify the intent and goals of this query: [query text of the submitted query].” The responsive answer from the LLM 104 may be (for the example query): “Intent: Extract metrics. Goal: Retrieve latest report and extract sustainability metrics.[0073] With an intent and goal of the query determined, the process then determines which states, as defined in the agent configuration file (e.g., 300 in FIG. 3A), are applicable to the query. To determine which states are applicable, the process constructs one or more prompts and submits those prompt(s) ... [99-108] elaborate on the matter [FIG.1] show corresponding visual) Figure 1 and corresponding paragraphs in Agah show that there can be multiple prompts and queries created and utilized receiving a second output from the first AI agent, wherein the second output includes information generated from the nested AI agent. (Agah [FIG.1] shows corresponding visual on the response/output from the ai agent [0025] FIG. 1 is an architecture diagram of the example systems used in connection with certain example embodiments, this includes the computer system that is configured to receive and process a query and generate a responsive output. FIG. 2 is a flowchart of a process for generating an agent configuration file that includes state and tool definitions that are used by the system of FIG. 1. FIGS. 3A-3B are examples of state and tool definitions that may be created for the agent configuration file discussed in FIG. 2. FIG. 4[0030] The autonomous LLM agent 102 is a computer program that may be instantiated as one or more computer processes. In certain examples, each instance of the autonomous LLM agent 102 may be executed within its own container (e.g., a docker container or the like). In other examples, separate instances of the autonomous LLM agent 102 may be instantiated for each query that is being processed. The autonomous LLM agent 102 may act as a controller or the like that processes a query submitted by a user, determines a workflow to execute (e.g., a dynamically generated and/or optimized set of states for the workflow), and then manages the execution of that workflow. In certain examples, the autonomous LLM agent 102 is configured to parse a given query, dynamically generate a task list that includes one or more states, and then execute tasks within that list using one or more sub-agents that are associated with that state. [0038] the autonomous LLM agent 102. As an illustrative example, the application program 138 may generate a web page that provides users with the ability to submit a query. As discussed below in connection with the various examples, application program 138 may provide responsive output that may be presented to the user that submitted the query. In certain example embodiments, the application server 108 may also be used to generate webpages (or other graphical user interfaces) that are communicated and displayed on client devices 110. Examples of different webpages that may be delivered to and displayed on client device are shown in FIGS. 10A-10C.[0042] In some examples, a sub-agent may leverage or include a dynamic hybrid RAG pipeline. This functionality may include an adaptive weighting technique that fine-tunes the emphasis on keyword(s) and semantic search based on query specifics. It then may combine these results using a flexible aggregation method influenced by the dynamic weights. This approach allows for tailored query handling and result scoring, leading to enhanced content relevancy and precision in subsequent processing stages. Examples of tools that may be incorporate or use such functionality...[69-74] elaborate on the response/output from the ai agent) Agah lacks explicitly and orderly teaching wherein the first AI agent lacks permission to directly access the second set of files in the second container and the nested AI agent lacks permission to directly access the first set of files in the first container; However Deutsch teaches wherein the first AI agent lacks permission to directly access the second set of files in the second container and the nested AI agent lacks permission to directly access the first set of files in the first container; (Deutsch [0006] FIG. 4 illustrates an example method for generating a plan and determining that the AI agent instance has permission to execute the plan in accordance with some embodiments of the present technology. 0022] When viewed through the lens of having AI agent instances interacting with each other on behalf of different user accounts, there are challenges in the principal (user account)—AI agent interaction experience, in determining the AI agent's permission or authority to act, and in determining whether consent exists from the recipient of a communication to interact with or through an AI agent, etc. [0033] FIG. 1 also illustrates that the sender AI agent instance 102 can have access to one or more sender user account apps 112. Likewise, recipient AI agent instance 104 can have access to one or more recipient user account apps 110. In particular, user account can provide access to some apps like a calendar application, a document management system, a workflow application, etc., to their respective AI agent instance. In this way, the respective AI agent instances can at least learn more information about a user account's context, and in some instances can take actions using these apps on behalf of the user account. In some embodiments, the respective AI agent instance might access the respective apps through sender user account front end 118 or recipient user account front end 122. [82-90] elaborate on the matter [FIG.4] shows the corresponding flow which can include wherein the first AI agent lacks permission to directly access the second set of files in the second container and the nested AI agent lacks permission to directly access the first set of files in the first container [FIG.1] shows an overall visual of the system which can perform the corresponding limitation ) Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to take all prior methods and make the addition of Deutsch’s agent methods in order to efficiently create a more secure system (Deutsch [0094] In some embodiments, an AI agent can avoid unnecessary communications or approvals with the user account by using the memory file. The memory file can include information learned from the user account, so just like a human assistant can learn preferences about when someone they are assisting wants to be consulted, so too can the AI agent. In some embodiments, when a fact or permission is recorded in the memory, the AI agent can avoid messaging the user account, unless the memory also indicates that the user account wants to be consulted [0135] The task agents 816 are autonomous AI agents that are generally trained to perform a specific type of task or that might be trained on a particular knowledge set. Task agents 816 might be smaller (less trainable parameters) and more efficient than a more generalized knowledge model such as coordinator agent ... [FIG.4] shows the corresponding flow which can include wherein the first AI agent lacks permission to directly access the second set of files in the second container and the nested AI agent lacks permission to directly access the first set of files in the first container [FIG.1] shows an overall visual of the system which efficiently create a more secure system )
Regarding Claim 9, Agah and Deutsch teach The method of claim 8, wherein the first container comprises a first custom prompt that configures the first AI model to respond to prompts, the method further comprising: providing the first custom prompt to the first AI model before providing the first prompt to the first AI model. (Agah [0040] Turning now more specifically to the autonomous LLM agent 102 and the sub-agents thereof. The autonomous LLM agent 102 includes or has access to sub-agents that may be individually selected and used based on the particular nature of the query being processed by the autonomous LLM agent 102. The sub-agents of the autonomous LLM agent 102 can be thought of as individual tools that the agent 102 can employ in connection with the dynamically constructed workflow. Accordingly, the autonomous LLM agent 102 can be thought of as a multi-tool agent in certain examples. The sub-agents include any or all of the following: 1) a document disclosure RAG (retrieval-augmented generation) tool 120; 2) a regulatory RAG tool 122; 3) a tabular data agent 124; 4) a web search & parsing agent 126; 5) a report segment composer sub-agent agent 128 (which may also be a database query agent in some examples); 6) a report retrieval & conversion tool agent 130 (which may also be a PDF extraction agent in some examples); and 7) a prompt engineering pipeline module 132. Other sub-agents may also be included or be accessed by the autonomous LLM agent 102 in connection with processing a query and the workflow generated therefrom. [0053] Prompt engineering pipeline module 132 includes two sub-components: 1) Prompt Config File Execution Module 134; and 2) Prompt Config File Generation Module 136. The Prompt Config File Generation Module 136 is configured to generate a prompt config file that is then executed by the Prompt Config File Execution Module 134. Details of these two modules are discussed, among other places, in FIGS. 7 and 8 and the illustrative examples discussed below.[0064] In certain example embodiments, each tool that is defined in 350 may have a corresponding sub-agent module that implements that processing associated with that tool. Thus, for example, the “WebSearch Sub-Agent” tool that is defined in FIG. 3B may correspond or reference Web Search & Parsing Agent 126 that is discussed in connection with FIG. 1. Similarly, the “PromptEngineeringPipeline” may correspond to the Prompt Engineering Pipeline Module...[89 &170] further elaborate [FIG.1] shows overall visual of the system )
Regarding Claim 12, Agah and Deutsch teach The method of claim 8, wherein the second container comprises a second custom prompt that configures the nested AI agent to respond to prompts, the method further comprising: providing the second custom prompt to the nested AI agent before providing the second prompt to the nested AI agent. (Agah [0043] trained LLM with domain specific or up-to-date knowledge for when the LLM generates a response to one or more prompts. [0071] At 402, the query is processed to generate a workflow that includes a list of one or more of the pre-defined states (e.g., as discussed in connection with FIG. 2). The query that is generated is modular in that the states that makeup any given query can change based on the nature of the query. The processing of the query may include submitting one or more prompts to LLM 104 to determine which states are applicable for the query that is received at 400 [0072] Determination of which states are applicable for a workflow for a query that has been received may include generating and submitting one or more prompts to LLM 104. A first prompt may be used to analyze the intent and goals of the query. An illustrative example prompt that is submitted to an LLM may include “Classify the intent and goals of this query: [query text of the submitted query].” The responsive answer from the LLM 104 may be (for the example query): “Intent: Extract metrics. Goal: Retrieve latest report and extract sustainability metrics.[0073] With an intent and goal of the query determined, the process then determines which states, as defined in the agent configuration file (e.g., 300 in FIG. 3A), are applicable to the query. To determine which states are applicable, the process constructs one or more prompts and submits those prompt(s) ... [0099] The prompt configuration file 700 includes a prompt pipeline that is composed of a plurality of sequenced prompts 706 (e.g., “prompt 1”, “prompt 2”, “prompt 3”, etc.). Each of these prompts includes one or more templates 708. Each of the items in brackets within the templates (e.g., “{disclosure_item}”, “{unit_of_measurement1}”, etc.) are variables that will be defined when the prompt is executed by the Prompt Engineering Pipeline Module 132.[0100] It will be appreciated that different prompt engineering configurations can be developed depending on application need. The example discussed below in connection with FIG. 8 relates to scope 1 emissions. However, another prompt example that follows from the techniques herein is provided below in the context of insights into a Board of Directors Oversight of Climate related topics. Accordingly, the different prompt engineering configurations[99-108] elaborate on the matter [FIG.1] show corresponding visual)
Claims 10-11 are rejected under 35 U.S.C. 103 as being unpatentable over Agah in view of Deutsch and Joshi
Regarding claim 10, Agah and Deutsch teach The method of claim 8, Agah lacks explicitly and orderly teaching further comprising: receiving a request to delete the first container or the first AI agent; and in response to receiving the request, deleting the first AI agent and deleting the first container including the second container and the nested AI agent. However Joshi teaches receiving a request to delete the first container or the first AI agent; and in response to receiving the request, deleting the first AI agent and deleting the first container including the second container and the nested AI agent. (Joshi [0056] Some embodiments of the container manager 20 may further be configured to determine when containers have ceased to operate, are operating at greater than a threshold capacity, or are operating at less than a threshold capacity, and take responsive action, for instance by terminating containers that are underused, re-instantiating containers that have crashed, and adding additional instances of containers that are at greater than a threshold capacity. Some embodiments of the container manager 20 may further be configured to deploy new versions of images of containers, for instance, to rollout updates or revisions to application code. Some embodiments may be configured to roll back to a previous version responsive to a failed version or a user command. In some embodiments, the container manager 20 may facilitate discovery of other services within a multi-container application, for instance, indicating to one service executing in one container where and how to communicate with another service executing in other containers, like indicating to a web server service an Internet Protocol address of a database management service used by the web server service to formulate a response to a webpage request. In some cases, these other services may be on the same computing device and accessed via a loopback address or on other computing devices.[55-58] elaborate on the matter [FIG.1] shows corresponding visual) Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to take all prior methods and make the addition of Joshi in order to create a more efficient system via delete of unnecessary containers (Joshi [0056] Some embodiments of the container manager 20 may further be configured to determine when containers have ceased to operate, are operating at greater than a threshold capacity, or are operating at less than a threshold capacity, and take responsive action, for instance by terminating containers that are underused, re-instantiating containers that have crashed, and adding additional instances of containers that are at greater than a threshold capacity. Some embodiments of the container manager 20 may further be configured to deploy new versions of images of containers, for instance, to rollout updates or revisions to application code. Some embodiments may be configured to roll back to a previous version responsive to a failed version or a user command. In some embodiments, the container manager 20 may facilitate discovery of other services within a multi-container application, for instance, indicating to one service executing in one container where and how to communicate with another service executing in other containers, like indicating to a web server service an Internet Protocol address of a database management service used by the web server service to formulate a response to a webpage request. In some cases, these other services may be on the same computing device and accessed via a loopback address or on other computing devices.[55-58] elaborate on the matter [FIG.1] shows corresponding visual)
Regarding claim 11, Agah and Deutsch teach The method of claim 8, Agah lacks explicitly and orderly teaching receiving a request to delete the second container or the nested AI agent; and in response to receiving the request, deleting the second container and the nested AI agent without deleting the first container and the first AI agent. However Joshi teaches receiving a request to delete the second container or the nested AI agent; and in response to receiving the request, deleting the second container and the nested AI agent without deleting the first container and the first AI agent (Joshi [0056] Some embodiments of the container manager 20 may further be configured to determine when containers have ceased to operate, are operating at greater than a threshold capacity, or are operating at less than a threshold capacity, and take responsive action, for instance by terminating containers that are underused, re-instantiating containers that have crashed, and adding additional instances of containers that are at greater than a threshold capacity. Some embodiments of the container manager 20 may further be configured to deploy new versions of images of containers, for instance, to rollout updates or revisions to application code. Some embodiments may be configured to roll back to a previous version responsive to a failed version or a user command. In some embodiments, the container manager 20 may facilitate discovery of other services within a multi-container application, for instance, indicating to one service executing in one container where and how to communicate with another service executing in other containers, like indicating to a web server service an Internet Protocol address of a database management service used by the web server service to formulate a response to a webpage request. In some cases, these other services may be on the same computing device and accessed via a loopback address or on other computing devices.[55-58] elaborate on the matter [FIG.1] shows corresponding visual) Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to take all prior methods and make the addition of Joshi in order to create a more efficient system via delete of unnecessary containers (Joshi [0056] Some embodiments of the container manager 20 may further be configured to determine when containers have ceased to operate, are operating at greater than a threshold capacity, or are operating at less than a threshold capacity, and take responsive action, for instance by terminating containers that are underused, re-instantiating containers that have crashed, and adding additional instances of containers that are at greater than a threshold capacity. Some embodiments of the container manager 20 may further be configured to deploy new versions of images of containers, for instance, to rollout updates or revisions to application code. Some embodiments may be configured to roll back to a previous version responsive to a failed version or a user command. In some embodiments, the container manager 20 may facilitate discovery of other services within a multi-container application, for instance, indicating to one service executing in one container where and how to communicate with another service executing in other containers, like indicating to a web server service an Internet Protocol address of a database management service used by the web server service to formulate a response to a webpage request. In some cases, these other services may be on the same computing device and accessed via a loopback address or on other computing devices.[55-58] elaborate on the matter [FIG.1] shows corresponding visual)
Claim 13 is rejected under 35 U.S.C. 103 as being unpatentable over Agah in view of Deutsch and US 20180329990 A1; SEVERN; Robert et al. (hereinafter Severn)
Regarding claim 13, Agah and Deutsch teach The method of claim 8, further comprising: Agah lacks explicitly and orderly teaching receiving a request to add a first file to the second container; in response to receiving the request to add the first file to the second container: adding the first file to the second container; generating a third set of embeddings based on the first file; and adding the third set of embeddings to the second container. However Severn teaches receiving a request to add a first file to the second container; in response to receiving the request to add the first file to the second container: adding the first file to the second container; (Severn [0053] The query processing module 140 interprets the queries as geometric constraints on the embedding space, and narrows or otherwise modifies a catalog of documents obtained from the embedding space to develop a set of candidate documents which satisfy the geometric constraints. These candidate documents are written into a candidate space database 150. Candidate spaces as used herein are also embedding spaces, and for example may constitute a portion of the embedding space of the document catalog database 120. [0517] In another implementation, the implicit user action can be one or more of (i) requesting a more detailed description of the document, (ii) requesting the document to be added to a cart, (iii) requesting the document to be added to a list (iv) the user lingering on a view of the document for more than a predetermined amount of time and (v) the user lingering on a view of the document for less than another predetermined amount of time. Further, the meaning assigned to the implicit user action is weighted in dependence on whether the implicit user action is one or more of (i), (ii), (iii), (iv) and (v). [0486] In operation 2818 a meaning is assigned to the user action. This meaning can be any of the types of user interactions described above. A score or value can then be determined for a document based on various types of implicit user actions associated with the document. A weighting can also be assigned based on the type of implicit user action. For example, opening up a product display page could be given a 0.5 weighting, adding a document to a wish list, a favorites list or the cart could be given a 1.0 weighting, lingering on a product in a carousel of displayed products or extending a mouse hover can be given an initial weighting of 0.1 which could increase over time, scrolling past a document, which would be similar to disliking a document, could be given a rating of 0. [485-487] elaborate on the matter [FIG.1] shows corresponding flow ) generating a third set of embeddings based on the first file; and adding the third set of embeddings to the second container. (Severn [0053] Referring to FIG. 1, a block diagram 100 of a visual interactive search system includes an embedding module 110 which calculates an embedding of source documents into an embedding space, and writes embedding information, in association with an identification of the documents, into a document catalog database (e.g., document catalog) 120. A user interaction module 130 receives queries and query refinement input (such as relevance feedback) from a user, and provides the received queries and query refinement input to a query processing module 140. In an implementation, the user interaction module 130 includes a computer terminal, whereas in another implementation the user interaction module 130 includes only certain network connection components through which the system communicates with an external computer terminal. The query processing module 140 interprets the queries as geometric constraints on the embedding space, and narrows or otherwise modifies a catalog of documents obtained from the embedding space to develop a set of candidate documents which satisfy the geometric constraints. These candidate documents are written into a candidate space database 150. Candidate spaces as used herein are also embedding spaces, and for example may constitute a portion of the embedding space of the document catalog database 120. [0059] Alternatively, the embedding module 110 may derive a library of image classifications (axes on which a given photograph may be placed), each in association with an algorithm for recognizing in a given photograph whether (or with what probability) the given photograph satisfies that classification. Then the embedding module 110 may apply its pre-developed library to a smaller set of newly provided photographs, such as the photos currently on the user computer 210, in order to determine embedding information applicable to each photograph. Either way, the embedding module 110 writes into the document catalog database 120 the identifications of the catalog of documents that the user may search, each in association with the corresponding embedding information.[0342] As can be seen, identification of a desired document may include providing, accessibly to a computer system, a database identifying a catalog of documents in an embedding space, calculating... [485-487] elaborate on the matter [FIG.1] shows corresponding flow ) Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to take all prior methods and make the addition of Severn in order to use embedding module methods in order to efficiently organize the system (Severn [0005] Some libraries are annotated with metadata, such as date and location (for a photo library), or type and features of products (for a product catalog). But many are not annotated, and even those which are may not be sufficiently specific to allow the server to efficiently hone in on the desired document quickly. As touched on above, some search technologies allow the server to perform searches iteratively, thereby gradually narrowing the field of possible documents until the target document is found. But these often still take a long time, and cause the server to consume an unnecessarily high level of bandwidth and processing power by requiring the server to offer many different collections of candidate documents before the target document is found. [0009] The new results provided to the user are driven by the “likes,” “dislikes” or “neutral” opinions provided by the user. As a result of implementing these features, a conventional computer system (e.g., the server and/or user device providing the results) is improved because the user is able to get to what they are looking for faster, with fewer iterations and more confidence. This is achievable because the computer system implementing these features is able to intelligently display a limited set of information to the user by gathering the information mentioned above and summarizing the user's interests to provide a limited set of valuable information to the user. [11-13] elaborate on the matter)
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ARYAN D TOUGHIRY whose telephone number is (571)272-5212. The examiner can normally be reached Monday - Friday, 9 am - 5 pm.
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, Aleksandr Kerzhner can be reached at (571) 270-1760. 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.
/ARYAN D TOUGHIRY/Examiner, Art Unit 2165
Prosecution Insights