Unified network protocol for specialized artificial intelligence agents

Abstract

A system for providing artificial intelligence services includes a user agent corresponding to a user of the system; a plurality of worker agents, each worker agent of the plurality of worker agents being configured to communicate with a different artificial intelligence system; and at least one core node, the at least one core node being configured to discover and rank worker agents. In such a case, the system is configured to perform the following steps: responsive to a request initiated by the user agent, the at least one core node discovers worker agents and provides details of the worker agents to the user agent, the details including rankings; based on the details of the worker agents provided by the core node, the user agent requests task execution from the worker agents based on the details provided by the core node; and the user agent receives results of the task execution and displays the results to the user.

Description

UNIFIED NETWORK PROTOCOLAND SYSTEM FOR SPECIALIZED ARTIFICIAL INTELLIGENCE AGENTSCROSS REFERENCE TO RELATED APPLICATIONS

[0001] This application claims priority to U.S. Provisional Patent Application Serial No.: 63 / 537.979, filed on September 12, 2023, which is incorporated herein by reference in its entirety.TECHNICAL FIELD

[0002] The present disclosure pertains to the field of artificial general intelligence (AGI), particularly to a decentralized network-based system of distinct Artificial Intelligence (Al) agents that communicate and collaborate to achieve specific tasks, catering to online service interactions and providing personalized responses based on user preferences.

[0003] In essence, this disclosure seeks to revolutionize the way online services are accessed and utilized, improving user experience by leveraging the capabilities of Al and decentralization to provide tailored, efficient, and responsive service interactions. By doing so, this disclosure is also relevant to the fields of distributed computing, e-commerce, personal digital assistants, service-oriented architecture, and data privacy.BACKGROUND

[0004] With the rapid expansion of the digital world, users today interact with a multitude of online services on a daily basis, ranging from e-commerce platforms, entertainment services, productivity tools, to various other digital utilities. Each of these platforms operates in isolation, requiring individual user inputs, account management, and separate interfacing. This fragmented digital experience often results in inefficiencies, redundancy, and a steep learning curve as users navigate through different interfaces and workflows for every platform they interact with.

[0005] Furthermore, as the number of online platforms grows, users often find themselves overwhelmed with the sheer number of choices and the challenge of discovering and integrating new services becomes increasingly daunting.

[0006] From the perspective of Al-driven automation and personalization, current systems are siloed. While many platforms employ Al to enhance user experience, these Al systems operate within their confined environments from the same ecosystem or company, thus locking users into a specific set of services and limiting their options, unable to share insights or learnings across platforms. This results in a lack of continuity' in user experience and missed opportunities for holistic personalization.

[0007] Moreover, the centralized nature of many existing Al solutions poses challenges related to data privacy, trust, and control. Users, more than ever, are concerned about how their data is used, where it's stored, and who has access to it.

[0008] There exists, therefore, a pressing need for an integrated system that allows for seamless interaction between various online services, one that can cater to user needs in a holistic manner while prioritizing user data privacy and customization. Such a system should be adaptable, scalable, transparent, and designed to encourage community-driven growth, ensuring it remains relevant and expansive as the digital landscape continues to evolve. SUMMARY

[0009] In one embodiment, a system for providing artificial intelligence services is presented. For instance, the system includes a user agent corresponding to a user of the system; a plurality of worker agents, each worker agent of the plurality of worker agents being configured to communicate with a different artificial intelligence system; and at least one core node, the at least one core node being configured to discover and rank worker agents. In such a case, the system is configured to perform the following steps: responsive to a request initiated by the user agent, the at least one core node discovers worker agents and provides details of the worker agents to the user agent, the details including rankings; based on the details of the worker agents provided by the core node, the user agent requests task execution from the worker agents based on the details provided by the core node; and the user agent receives results of the task execution and displays the results to the user.

[0010] In another embodiment, the core node creates the rankings of the plurality of worker agents based on historical performance, user feedback and latency statistics.

[0011] In another embodiment, each worker agent of the plurality of worker agents comprises a membership handler to enable discovery by the core node and a request engine to enable the task execution.

[0012] In another embodiment, the core node comprises a language model for processing natural language, and the system is further configured to determine the request for the task execution by converting a natural language query from the user agent.

[0013] In another embodiment, the request initiated by the user agent leads to multiple requests for task execution by each of the plurality of worker agents, wherein each of the multiple requests comprises an artificial intelligence query by the worker agent.

[0014] The above embodiments are exemplary' only. Other embodiments as described herein are within the scope of the disclosed subject matter.BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

[0015] So that the manner in which the features of the disclosure can be understood, a detailed description may be had by reference to certain embodiments, some of which are illustrated in the accompanying drawings. It is to be noted, however, that the drawings illustrate only certain embodiments and are therefore not to be considered limiting of its scope, for the scope of the disclosed subject matter encompasses other embodiments as well. The drawings are not necessarily to scale, emphasis generally being placed upon illustrating the features of certain embodiments. In the drawings, like numerals are used to indicate like parts throughout the various views, in which:

[0016] Figure 1 illustrates an exemplary realization of the network protocol architecture showcasing six network entities, in accordance with one or more aspects set forth herein;

[0017] Figure 2 illustrates a sequence diagram showcasing the flow of a user request through its lifecycle, in accordance with one or more aspects set forth herein; and

[0018] Figure 3 illustrates a secure enrollment mechanism flowchart, in accordance with one or more aspects set forth herein.

[0019] Corresponding reference characters indicate corresponding parts throughout several views. The examples set out herein illustrate several embodiments, but should not be construed as limiting in scope in any manner.DETAILED DESCRIPTION

[0020] The present disclosure presents a novel network system and protocol designed to unify interactions between specialized Al agents. The protocol employs a triad communication model, allowing (1) User Agents to decompose user requests, and then (2) through Core Nodes locate, and (3) select Worker Agents for task execution. This facilitates the execution of complex tasks across multiple platforms with a single user request. The system leverages a distributed asynchronous communication framework, ensuring efficient and scalable inter-agent communication. Additionally, the disclosure introduces a dynamic agent discovery’ mechanism and a ranking system, ensuring users receive best fitting services. This disclosure represents a significant advancement in collaborative Al systems, bridging user requests with actionable solutions in a structured, dynamic environment.

[0021] In addition, this disclosure introduces a decentralized network protocol comprising interconnected Al agents designed to facilitate seamless interactions among multiple onlinesen-ices and platforms, without navigating through disparate interfaces, reducing the cognitive load, inefficiencies, and limitations associated with multi-platform use.

[0022] Seamless Interoperability: One aspect of the disclosure is to provide a unified platform that enables data exchange, smooth interaction, collaboration, and task execution between various specialized Al agents catering to specific online products and services.

[0023] Dynamic Agent Discovery and Ranking Mechanism: One aspect of the disclosure is to implement a system that allows user agents to discover available worker agents, ensuring that users can always access the latest and better ranked services and functionalities as the network grows.

[0024] Holistic User-Centric Personalization: One aspect of the disclosure is to offer a mechanism that continually adapts to individual preferences and historical behaviors, ensuring that the outcomes are tailored to each user's unique requirements.

[0025] Scalability and Open Integration: One aspect of the disclosure is to establish a scalable network protocol, capable of integrating an increasing number of specialized services. Moreover, the system aims to be open, fostering community contributions and continuous improvement.

[0026] Network Defense and Resilience: One aspect of the disclosure is is to introduce safeguards that effectively deter and handle potential spamming, flooding, and other malicious activities, ensuring the integrity and availability of the network.

[0027] The present disclosure introduces a dynamic and scalable network protocol designed to facilitate seamless interactions between independent specialized Al agents, each executing tasks associated with certain online products and services. One aspect of this disclosure lies the vision of unifying the vast and fragmented digital Al agents landscape into an integrated ecosystem, wherein Al agents can communicate, collaborate, and share resources efficiently. The protocol standardizes the communication methods, data formats, and interaction patterns (agent enrollment, discovery, and ranking), ensuring that Al agents, regardless of their underlying architecture or primary' function, can discover each other for collaboration, understand and respond to each other's requests.

[0028] Central to the disclosure are three distinct network entities, defined by this disclosure:(1) Worker Agents'. Specialized Al entities tailored to interface with specific online products or sendees, serving as dedicated proxies for their respective external platforms.(2) User Agents'. Personalized Al entities that interact directly with users, collating user requests, and navigating the network to identify the most suitable Worker Agents for task execution.(3) Core Nodes'. Fundamental latent technical entities that underpin the entire system's architecture, responsible for crucial functions including the enrollment of new agents, the discovery of Worker Agents, and the ranking of these agents based on diverse metrics.

[0029] The inherent strength of the disclosure lies in its triad communication model between independent network entities, wherein User Agents, acting on user inputs, employ the Core Nodes to (1) decompose the request into atomic tasks. (2) locate, and (3) select Worker Agents for task executions. As a result, complex tasks that would traditionally necessitate maneuvering through multiple platforms can now be distilled into a singular user request executed by multiple Worker Agents, epitomizing the "One Request Does It All" principle.

[0030] A cornerstone of this disclosure is its distributed asy nchronous communication framework, leveraging state-of-the-art message queue systems. This ensures resilient, efficient, and scalable inter-agent communication. Agents are incorporated into the network through a structured and secure enrollment process, ensuring consistency and accountability.

[0031] Notably, the disclosure presents a robust discovery' mechanism where User Agents can swiftly locate Worker Agents apt for particular tasks. Central to this process is the Core Node's dynamically updated agent directory, which classifies and organizes agents based on capabilities, specializations, and more.

[0032] Furthermore, this disclosure pioneers an intelligent Worker Agent ranking system. Beyond mere discovery', the network meticulously ranks Worker Agents based on multiple parameters like historical performance, user feedback, latency statistics, and more. This ensures that users consistently receive the best-suited agent for their request, thereby improving sendee quality.

[0033] In conclusion, this disclosure is a paradigm shift in the realm of collaborative Al systems, providing a structured yet dynamic architecture that seamlessly bridges requests with actionable solutions, facilitating Artificial General Intelligence (AGI) user experience.

[0034] NETWORK PROTOCOL ENTITIES

[0035] One aspect of the disclosure is a unified network protocol that connects three distinct types of Al entities, as depicted in Figure 1 : Worker and User Agents (visible), and Core Nodes (latent).

[0036] Worker Agents: These are specialized agents tailored to interact with specific online products or services. Their primary’ role is to serve as a proxy for the services they are designed for and execute tasks or retrieve data from them. Every Worker Agent implements two components: (1) Membership Handler and (2) Request Engine. The former component governs the agent's enrollment, authentication, and ongoing membership within the unified network. It ensures a trusted relationship between the agent and the core network. The latter component is responsible for executing tasks, interfacing with the designated online service, and retrieving or sending data within the network.

[0037] User Agents: These agents interact directly with human users, acting as a personalized interface. They gather user requests, identify the right Worker Agents to fulfill these requests, and relay necessary’ instructions, all while preserving user data. Similar to Worker Agents, every User Agent implements tw o components: (1) Membership Handler and (2) Query Engine. The membership handler serves the same purpose - to govern the agent's enrollment, authentication, and ongoing membership within the unified netyvork. The query engine handles user input decomposition, orchestrates the distributed task execution, and aggregates the final result to present it to the user.

[0038] Core Nodes: These technical entities in the network serve three purposes, with one component for each: Worker Agent enrollment (Membership Handler), discovery’ (Discovery’ Handler) and ranking (Ranking Engine). Core Nodes rely on a Large Language Model (LLM) as the reasoning engine to handle incoming natural language request and convert it into executable action items based on available Worker Agents in the directory.

[0039] Figure 1 illustrates an exemplary’ realization of the netyvork protocol architecture showcasing six network entities. Three separate net vork entity types are depicted: Worker Agents, User Agents, and a Core Node. The Worker Agents act as proxies for the products and services within the network. User Agents provide an interface to the netyvork for human users. Core Node facilitates netyvork membership, discovery, and ranking mechanisms, connecting both User and Worker Agents. The diagram differentiates betyveen tyvo ty pes of connections - (1) Any AgentoCore Node connection (solid line) and (2) UserAgento Worker Agent connection (dotted line). The former connection serves as a gateway to the network, enabling agent discovery’ to establish a direct link. The latter connection represents a direct communication channel betyveen User and Worker Agents for requests and responses.

[0040] By way of further explanation, in the embodiment of Figure 1, a network system 100 includes users 101 controlling user agents 110, worker agents 120 and core node(s) 130. For instance, each user agent 110 includes a query engine 112 and a membership handler 114. Continuing, each worker agent 120 includes a request engine 122 and a membership handler 124. And each core node(s) 130 includes a discovery handler 132, a ranking engine 134 and a membership handler 136.

[0041] NETWORK PROTOCOL MECHANICS

[0042] The architecture facilitates a triad communication model depicted in Figure 2. Users channel their requests through their designated User Agent. This User Agent, leveraging the Core Node, evaluates the requirement, identifies, and selects relevant Worker Agents based on their ranking. Post this evaluation, the User Agent collaborates with the chosen Worker Agent(s) to fulfill the request, all while leveraging stored user data for maximum personalization. With this robust infrastructure, a single user command can set in motion a series of tasks that would traditionally require navigation through multiple platforms or interfaces, following the "One Request Solves It All" principle.

[0043] Figure 2 illustrates a sequence diagram showcasing the flow of a user request through its lifecycle. Beginning with the initiation from the User to the User Agent, the diagram traces the discovery' process via the Core Node, followed by parallel task execution by two Worker Agents, and culminates in the aggregation of results by the User Agent. This multi-agent execution underscores the network's ability to distribute tasks and collate diverse results for optimized user outcomes. In summary, the diagram delineates between (1) User AgentoCore Node interactions for Worker discovery, (2) User Agento Worker Agent direct communications for task assignments and result relay, and (3) the final aggregation process within the User Agent to synthesize and present consolidated results to the User.

[0044] Communication Protocol

[0045] This communication mechanism is achieved through four distinct network mechanics: Agent Enrollment, Discovery', Inter-Agent Communication, and Ranking.

[0046] The protocol leverages a distributed message queue system, which facilitates communication among the different agents in an asynchronous manner. Each network entity acts as both a message producer and consumer with the following key aspects of the communication protocol:

[0047] Message Publishing: Agents publish messages to specific topics or channels designated for distinct tasks or services. This ensures that only the relevant agents consume these messages.

[0048] Message Consumption: Agents subscribe to specific topics or channels relevant to their functions. Once a message is available in a subscribed topic, the agent retrieves, processes it, and can subsequently publish a response to a different topic if necessary.

[0049] Message Format: Every message transmitted in the system adheres to a standardized format, encapsulating meta-information like sender ID, intended recipient, timestamp, message ty pe, and payload. The payload contains the main content of the message.

[0050] Message Persistence & Retention: The message queue system retains messages until they are successfully consumed by the target agent. This ensures that no information is lost in the case of temporary agent downtimes.

[0051] Error Handling & Acknowledgments: Once a message is consumed and processed, agents send an acknowledgment back to the system. In the event of processing failures, error messages are generated and can be acted upon by either the sender, recipient, or administrative nodes.

[0052] Choice of System: The architecture utilizes message-queueing systems for the queue mechanism, given their robustness, reliability, and extensive scalability7options.

[0053] By adhering to this decoupled, asynchronous approach, the system ensures a scalable, efficient, and resilient mode of communication, which is essential for the diverse and dynamic interactions envisioned among the various Al agents.

[0054] Agent Enrollment

[0055] Enrollment into the network is a foundational step that ensures the system's orderliness, traceability7, and security. The process varies slightly depending on the entity type.

[0056] As integral components of the architecture. Core Nodes are commissioned and maintained directly by the protocol's originators. They serve as the bedrock upon which the network's stability and reliability are built. Each Core Node is uniquely identified and has predefined functionality: agent registration, discovery, and ranking. Every other Agent knows how to reach a Core Node by network design as Core Nodes reserve special message queues.

[0057] By way of further explanation, Figure 2 depicts an embodiment of a method 200 corresponding to the flow of a user request through its lifecycle. For instance, at block 202 a user 101 by way of user agent 110 initiates a request. In such a case, at block 204 the useragent 110 discovers worker agents using the core node 130. Next, at block 206, the core node 130 provides worker agent details to the user agent 110. Next, at block 208 the user agent 110 requests task execution from the worker agent 130-1 and at block 210 receives a task result. Similarly, at block 212 the user agent 110 requests task execution from the worker agent 130-2 and at block 214 receives a task result. Subsequently, at block 216, the user agent 110 displays the aggregated results to the user 101.

[0058] The protocol encourages an open ecosystem wherein external developers can create and introduce their own Worker and User Agents. The enrollment process for these agents is standardized to ensure network compatibility and security, as depicted in Figure 3.

[0059] User Agent(1) Initialization-. The User Agent generates a unique public-private key pair. The User Agent identifies the secure channel topic of the Core Node.(2) Connection Request. The User Agent sends a connection request to the Core Node on the secure channel topic. This request includes its public key and a unique identifier (UID).(3) Validation & Challenge. The Core Node generates a unique cryptographic challenge for the User Agent using the provided public key. The challenge is sent back to the User Agent.(4) Challenge Response'. The User Agent decrypts the challenge using its private key. It processes the challenge (e.g.. computes a cryptographic hash) and sends the result back to the Core Node.(5) Verification & Acceptance'. The Core Node verifies the response. If the response is correct, the Core Node adds the User Agent's UID and public key to its list of approved agents. An acknowledgment message, along with a unique token, is sent back to the User Agent to complete the enrollment.(6) Finalization'. The User Agent stores the received token, which will be used for subsequent authentications and interactions within the network.

[0060] Worker Agent(1) Initialization'. The Worker Agent generates its public-private key pair. The Worker Agent identifies the secure channel topic of the Core Node.(2) Service Declaration & Connection Request Besides its public key and UID, the Worker Agent sends details about the services it offers.(3) Validation & Service-specific Challenge'. The Core Node, based on the declared sendees, generates a unique service-specific cryptographic challenge for the Worker Agent. This challenge is not just about agent identity but also about its capabilities. The challenge is sent back.(4) Challenge Response'. The Worker Agent decry pts and processes the challenge. This might involve demonstrating a particular service capability. It then sends the result back to the Core Node.(5) Verification Provisional Acceptance'. The Core Node verifies the response. If correct, the Worker Agent is provisionally accepted. A provisional acknowledgment message, along with a unique token, is sent to the Worker Agent.(6) Trial Phase (Optional) '. For a certain period, the Worker Agent's interactions are monitored, ensuring its services align with the declared capabilities and meet network quality standards.(7) Final Approval'. After the trial phase, if the Worker Agent's performance is satisfactory, it gets final approval. An updated acknowledgment is sent, confirming its permanent status in the network.(8) Finalization'. The Worker Agent stores the received token for subsequent authentications.

[0061] Figure 3 illustrates a Secure Enrollment Mechanism Flowchart: Steps for User and Worker Agent Registration with the Core Node. For the Worker Agent enrollment, the enrollment challenge can be extended to test for claimed functionality'.

[0062] By maintaining a streamlined and secure enrollment process, the system strives to allow for all participating agents, whether originating from the protocol's creators or external developers, to coexist in a harmonized, accountable, and efficient environment.

[0063] Worker Agent Discovery via Core Nodes

[0064] A fundamental aspect of the interconnected agent ecosystem is the ability' for agents to dynamically discover one another, ensuring fluidity7in service delivery7. The mechanism hinges on the Core Nodes, which act as the nexus of information and the primary facilitators of agent discovery.(1) Agent Directory & Categorization '. Core Nodes maintain a dynamic directory' of all enrolled agents. This directory' classifies agents based on their capabilities, specialization, and other meta-attributes, for instance, based on the online services orproducts they cater to. Table 1 demonstrates an exemplary specification of a Core Node agents directory.(2) Discovery Request Initiation'. A User Agent, upon receiving a user request, formulates a discovery query. This query delineates the specifics of what kind of Worker Agent(s) it needs to interact with.(3) Query Processing'. The Core Node, upon receiving the discovery query, scours its directory to identify Worker Agents that align with the criteria specified.(4) Responsive Communication'. The Core Node then communicates the relevant Worker Agent details back to the query ing User Agent. This data might include the Worker Agent's UID, capabilities, communication channels, and any other pertinent attributes.(5) Discovery Caching (Optional)'. To expedite repetitive queries and reduce load, the discovery mechanism supports caching. User Agents can locally cache frequently interacted Worker Agents to minimize lookup times in subsequent interactions.(6) Dynamic Updates'. The Core Node's agent directory is not static. It dynamically updates to account for new agent enrollments, agent retirements, or capabilitychanges. This ensures that the discovery process remains current, relevant, and comprehensive.

[0065] Table 1 The specification for the Agents' directory in a Core Node.

[0066] By centralizing the discovery mechanism around Core Nodes and maintaining a dynamic and comprehensive agent directory, the system guarantees swift, accurate, and efficient agent interactions, optimizing user request fulfilment.

[0067] Inter- Agent User- Worker Data Flow

[0068] The communication between User Agents and Worker Agents is paramount. This interaction is defined by an efficient, standardized, and secure protocol ensuring streamlined communication.(!) Message Initiation'. Once a User Agent, through the Core Node's discovery7mechanism, identifies the appropriate Worker Agent(s) to cater to a user's request, it crafts a standardized request message. This message contains the details of the task, any specific parameters, and other relevant metadata.(2) Asynchronous Communication Protocol'. Leveraging a message queue-based approach, the User Agent places the request in the designated queue for the Worker Agent. This asynchronous methodology ensures that neither agent needs to be continuously active, awaiting responses, thereby conserving computational resources.(3) Task Execution'. Upon fetching the message from its queue, the Worker Agent processes the request. It interfaces with the specific online service / product it's tailored for, executes the necessary7operations, and prepares a response.(4) Response Formulation'. The Worker Agent formulates a response, which can range from data retrievals to task status updates. This response is then encapsulated into a standardized message structure, ensuring uniformity7.(5) Response Dispatch'. Utilizing the same asynchronous communication approach, the Worker Agent dispatches the response back to the User Agent through the corresponding message queue.(6) Filial User Interface Rendering'. The User Agent, upon receiving and parsing the Worker Agent's response, integrates this data with locally stored user data for personalized results. It then dynamically crafts a tailored output for the user, either rendering the results or suggesting further actions based on the outcome.

[0069] This data flow design ensures a robust, scalable, and time-efficient communication pipeline between agents. By employing standardized message structures and leveraging asynchronous communication, the protocol minimizes latency and optimizes user experience.

[0070] By way of further explanation, Figure 3 depicts an embodiment of a method 300 for secure enrollment. For instance, at block 302 the agent begins enrollment. Next, at block 304 the agent generates a public-private key pair. Next, at block 306 the agent sends a connection request with the generated public key, a UID and a type. Next at block 308 the core node generates a cryptographic challenge. Next at block 310 the agent decrypts and processes the challenge and sends a response. Next at block 312 the core node determines whether or not to verily the response. If verified, at block 314 the core node sends an acknowledgement and a token (otherwise enrollment fails at block 313). Then at block 316 the agent enrollment is completed.

[0071] Worker Agent Ranking

[0072] The network's strength lies not just in the mere discovery of Worker Agents but in effectively ranking them to present the most suitable choices for a given user request. This intelligent ranking system ensures that the user benefits from the highest quality of service available.(1) Input Parameters for Ranking'. The ranking mechanism considers various factors, with its implementation details being out of scope of this patent. A few examples include, but are not limited to: historical performance data of the Worker Agent, feedback or ratings (if available) from previous interactions, specificity and relevance of the Worker Agent's capabilities concerning the current request, latency or response time statistics from past engagements, and agent's load or current activity levels.(2) Ranking Algorithm: The Core Node employs sophisticated algorithm - subject to a separate white paper - to process these parameters. Over time, the system continually refines its rankings as it learns from more interactions, ensuring an adaptive and evolving ranking system.(3) User Personalization'. Apart from the global ranking, the User Agent can overlay user-specific preferences based on past interactions, stored user data, or user-set priorities. This ensures a fine-tuned match, enhancing user satisfaction.(4) Ranking Output'. For each discovery query, the Core Node provides not just a list of matching Worker Agents but a prioritized one. This list is dynamically generated for every request, ensuring real-time relevance.(5) Feedback Loop: After each interaction, there's an optional feedback mechanism where the User Agent can rate or review' the Worker Agent's performance. This feedback isthen integrated into the ranking system, ensuring a continuously improving and selfregulating system.

[0073] Through this ranking mechanism, the protocol ensures that users consistently receive top-tier service, while also promoting competition and quality improvement among Worker Agents in the network.

[0074] Overall, the tri-pronged mechanism of enrollment, discovery, and ranking ensures that the network remains organized, efficient, and delivers an optimal user experience, irrespective of its scale or the diversity of agents it encompasses.

[0075] Different aspects of different embodiments of the present disclosure may be summarized as the following numbered Aspects:1. A unified network protocol, characterized by its capability to facilitate interactions among distinct Al entities comprising Worker Agents, User Agents, and Core Nodes.2. The Worker Agent as described in Aspect 1, where said agent functions as a service proxy interfacing with online services.3. The User Agent as referenced in Aspect 1, designed to provide a personalized interface for human users and orchestrating interactions with Worker Agents to fulfill user requests.4. The Core Node as referenced in Aspect 1, functioning to oversee Agent enrollment, discovery, and ranking.5. A distributed message queue system integrated within the unified network protocol as specified in Aspect 1, ensuring asynchronous, resilient, and efficient communication among agents.6. An enrollment process as described in Aspect 1, wherein said process verifies the authenticity of agents and ensures network harmony.7. The dynamic agent discovery feature as mentioned in Aspect 1, enabling agents to swiftly identify and interact with appropriate counterparts via Core Nodes.8. A Worker Agent ranking system as depicted in Aspect 1, where agents are ranked based on multifaceted evaluation criteria, ensuring users engage with the highest- quality agents.9. The method by which User Agents interface with Worker Agents, as detailed in Aspect 3, in which interactions are coordinated to fulfill specific user requests.The protocol as mentioned in Aspect 1, where the interaction among Al entities is scalable, allowing for the integration of new agents without significant alterations to the existing infrastructure. The mechanism within the Core Node, as specified in Aspect 4, which validates the credentials and capabilities of Worker Agents during the enrollment process. A security framework incorporated within the unified network protocol as described in Aspect 1, designed to safeguard the communication and data exchange among agents. A feedback loop incorporated within the unified network protocol as mentioned in Aspect 1, allowing for the continuous improvement of agent interactions based on user feedback and system analytics. A fail-safe mechanism within the distributed message queue system as described in Aspect 5, ensuring message delivery even in cases of temporary network disruptions. A unified network protocol, characterized by its capability to facilitate interactions among distinct Al entities comprising Worker Agents, User Agents, and Core Nodes. The Worker Agent as described in Aspect 1, where said agent functions as a service proxy interfacing with online services. The User Agent as referenced in Aspect 1, designed to provide a personalized interface for human users and orchestrating interactions with Worker Agents to fulfill user requests. The Core Node as referenced in Aspect 1, functioning to oversee Agent enrollment, discovery', and ranking. A distributed message queue system integrated within the unified network protocol as specified in Aspect 1, ensuring asynchronous, resilient, and efficient communication among agents. An enrollment process as described in Aspect 1, wherein said process verifies the authenticity' of agents and ensures network harmony. The dynamic agent discovery' feature as mentioned in Aspect 1, enabling agents to swiftly identify and interact with appropriate counterparts via Core Nodes. A Worker Agent ranking system as depicted in Aspect 1, where agents are ranked based on multifaceted evaluation criteria, ensuring users engage with the highest- quality agents.23. The method by which User Agents interface with Worker Agents, as detailed in Aspect 3. in which interactions are coordinated to fulfill specific user requests.24. The protocol as mentioned in Aspect 1, where the interaction among Al entities is scalable, allowing for the integration of new agents without significant alterations to the existing infrastructure.25. The mechanism within the Core Node, as specified in Aspect 4, which validates the credentials and capabilities of Worker Agents during the enrollment process.26. A security framework incorporated within the unified network protocol as described in Aspect 1, designed to safeguard the communication and data exchange among agents.27. A feedback loop incorporated within the unified network protocol as mentioned in Aspect 1, allowing for the continuous improvement of agent interactions based on user feedback and system analytics.28. A fail-safe mechanism within the distributed message queue system as described in Aspect 5, ensuring message delivery even in cases of temporary network disruptions.

[0076] SUMMARY OF NETWORK MESSAGE TYPES

Claims

CLAIMSWhat is claimed is:

1. A system for providing artificial intelligence services, the system comprising: a user agent corresponding to a user of the system; a plurality of worker agents, each worker agent of the plurality of worker agents being configured to communicate with a different artificial intelligence system; at least one core node, the at least one core node being configured to discover and rank worker agents; and wherein the system is configured to perform the following steps: responsive to a request initiated by the user agent, the at least one core node discovers worker agents and provides details of the worker agents to the user agent, the details including rankings; based on the details of the worker agents provided by the core node, the user agent requests task execution from the worker agents based on the details provided by the core node; and the user agent receives results of the task execution and displays the results to the user.

2. The system of claim 1, wherein the core node creates the rankings of the plurality of worker agents based on historical performance, user feedback and latency statistics.

3. The system of claim 1, wherein each worker agent of the plurality of worker agents comprises a membership handler to enable discovery’ by the core node and a request engine to enable the task execution.

4. The system of claim 1, wherein the core node comprises a language model for processing natural language, and the system is further configured to determine the request for the task execution by converting a natural language query from the user agent.

5. The system of claim 1, wherein the request initiated by the user agent leads to multiple requests for task execution by each of the plurality of worker agents, wherein each of the multiple requests comprises an artificial intelligence query by the worker agent.

6. A system for providing artificial intelligence services, the system comprising: a user agent corresponding to a user of the system; a plurality of worker agents, each worker agent of the plurality of worker agents comprising a membership handler to enable discovery by the core node and a request engine to enable task execution and being configured to communicate with a different artificial intelligence system; at least one core node, the at least one core node comprising a language model for processing natural language and being configured to discover and rank worker agents based on historical performance, user feedback and latency statistics; and wherein the system is configured to perform the following steps: responsive to a request initiated by the user agent, the at least one core node determine a request for task execution by converting a natural language query from the user agent, discovers worker agents for task execution, and provides details of the worker agents to the user agent, the details including rankings; based on the details of the worker agents provided by the core node, the user agent requests task execution from the worker agents based on the details provided by the core node; and the user agent receives results of the task execution and displays the results to the user.

Images