Hierarchical knowledge representation method for skills in dynamic task scenario

Abstract

The present invention relates to the technical field of robots and knowledge engineering, and in particular relates to a hierarchical knowledge representation method for skills in a dynamic task scenario, an apparatus, an electronic device, and a storage medium. The method comprises: on the basis of different semantic knowledge types, using a graph database to construct a knowledge hierarchical understanding framework; and, on the basis of a preset task, performing knowledge extraction on the dynamic task scenario and, on the basis of a triple, performing external knowledge integration and dynamically updating a knowledge base to obtain a dynamic operation skill knowledge base. The application further discloses: designing a multi-level knowledge expression system of operations, planning, tasks, and objects, establishing a dynamic action library, a skill library, and a task library, and forming a robot skill operation knowledge library in a dynamic scenario together with a static entity library and a scenario library, thereby achieving knowledge guidance and data-driven operation task reasoning. Thus, a model is designed according to different scenario characteristics and an operation skill knowledge library is constructed after collecting data, thereby improving the interpretability, operation task planning, and skill migration efficiency of robot operation skill learning.

Description

A hierarchical knowledge representation method for skills in dynamic task scenarios Technical Field

[0001] This application relates to the fields of robotics and knowledge engineering technology, and in particular to a hierarchical knowledge representation method, device, electronic device, and storage medium for skills in dynamic task scenarios. Background Technology

[0002] Task understanding and skill refinement are key capabilities for service robots. Decomposing tasks in a manner similar to human cognition can be transformed into a planning problem within a symbolic space. Classical task and motion planning methods heavily rely on predefined planning domains, symbolic rules, and complex policy searches. This prevents them from processing problems in a way that aligns with human cognition and semantics, resulting in high formulation costs. The constraint that robots must ensure all elements are known before starting a task hinders efficient migration and adaptation to constantly changing dynamic task scenarios. To address this, some semantic-based knowledge base construction methods have been incorporated into robot operation representation and task planning. This allows robots to extract task-related content from human knowledge. However, knowledge is a higher-dimensional organizational form than data, inherently discrete and structured, making it difficult to directly describe continuous operational data. Therefore, existing knowledge-based robot operation representation methods focus on characterizing static object information in operations, failing to achieve reasonable decoupling between different elements and dynamic factors, resulting in flat and chaotic descriptions of tasks, actions, and skills. During querying and reasoning, they only consider rule-based symbolic computation.

[0003] In summary, how to design models and build operational skills knowledge bases based on the characteristics of different scenarios is a problem that urgently needs to be solved.

[0004] Summary of the Invention

[0005] This application aims to at least partially address one of the technical problems in the related art.

[0006] Therefore, the first objective of this application is to propose a hierarchical knowledge representation method for skills in dynamic task scenarios, in order to solve the problems that existing knowledge-based robot operation representation methods focus on depicting static object information in the operation and do not achieve reasonable decoupling between different elements and dynamic factors.

[0007] The second objective of this application is to provide an apparatus.

[0008] The third objective of this application is to propose an electronic device.

[0009] The fourth objective of this application is to provide a computer-readable storage medium.

[0010] To achieve the above objectives, the first aspect of this application proposes a hierarchical knowledge representation method for skills in dynamic task scenarios, including:

[0011] Based on different semantic knowledge types, a knowledge hierarchical understanding framework is constructed using graph databases;

[0012] Knowledge is extracted from dynamic task scenarios based on preset tasks, and triples are constructed.

[0013] External knowledge is fused based on the triples, and the knowledge base is dynamically updated to obtain a dynamic operational skills knowledge base.

[0014] Preferably, the construction of a knowledge hierarchical understanding framework based on different semantic knowledge types using a graph database includes:

[0015] A hierarchical knowledge understanding framework is constructed using the graph database Neo4j. Neo4j stores data as a graph structure, where nodes represent entities and relations represent connections between entities, thus constructing a hierarchical knowledge understanding framework.

[0016] Preferably, the knowledge layering understanding framework includes:

[0017] Construct an ontology module, and use the ontology module to dynamically express the initial knowledge of the task to be parsed;

[0018] Construct a template module and use the template module to operate on task templates;

[0019] Build an instance module and use the instance module to perform instance operations.

[0020] Preferably, it further includes:

[0021] The knowledge layering understanding framework is divided into two levels: Level 1 is the environment level, which is used to express the static factors of operations, including objects, agents, and scenarios; Level 2 is the operation level, which is used to express the dynamic factors of operations, including tasks, skills, and actions.

[0022] Preferably, the knowledge extraction from dynamic task scenarios based on preset tasks includes:

[0023] Based on the preset tasks, knowledge extraction is performed on dynamic task scenarios. The text-to-knowledge processing method is divided into four tasks. Task 1 is operation text collection, which includes text descriptions of operation processes. Task 2 is entity extraction, which uses the prior knowledge and reasoning ability of large language models to extract labels from text data and construct triples.

[0024] Preferably, the step of fusing external knowledge based on the triples and dynamically updating the knowledge base to obtain a dynamic operational skills knowledge base includes:

[0025] Based on the triples constructed in Task 2, external knowledge fusion is performed using Task 3, and the context neighbors are retrieved from the general knowledge base using the standard query language to expand the knowledge related to the operation task.

[0026] Based on the knowledge expanded from Task 3, Task 4 is used to dynamically update the knowledge base. A unique identifier is attached to all actions in each template, and the existing knowledge is updated to the knowledge base to obtain a dynamic operation skill knowledge base.

[0027] Preferably, the execution of instance operations using the instance module includes: for each instance execution, adding the operation execution parameters and timestamp to the corresponding template.

[0028] To achieve the above objectives, a second aspect of this application proposes a hierarchical knowledge representation device for skills in dynamic task scenarios, comprising:

[0029] The framework building module utilizes graph databases to construct a hierarchical knowledge understanding framework based on different semantic knowledge types.

[0030] The knowledge extraction module extracts knowledge from dynamic task scenarios based on preset tasks and constructs triples.

[0031] The knowledge base update module performs external knowledge fusion based on the triples and dynamically updates the knowledge base to obtain a dynamic operation skills knowledge base.

[0032] To achieve the above objectives, a third aspect of this application provides an electronic device, including: a processor, and a memory communicatively connected to the processor;

[0033] The memory stores computer-executed instructions;

[0034] The processor executes computer execution instructions stored in the memory to implement the method described in any of the preceding descriptions.

[0035] To achieve the above objectives, a fourth aspect of this application provides a computer-readable storage medium, comprising computer-executable instructions stored therein, which, when executed by a processor, are used to implement the method described in any of the above embodiments.

[0036] This application provides a hierarchical knowledge representation method for skills in dynamic task scenarios. It designs a multi-level knowledge expression system for operations, planning, tasks, and objects, and establishes dynamic action libraries, skill libraries, and task libraries. These, together with static entity libraries and scenario libraries, constitute a robot skill operation knowledge base in dynamic scenarios. Combined with a real-time knowledge update mechanism, it realizes knowledge-guided and data-driven operation task reasoning. After designing models and collecting data according to different scenario characteristics, it constructs an operation skill knowledge base, improves the interpretability of robot operation skill learning, and enhances the efficiency of robot operation task planning and skill transfer.

[0037] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0038] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0039] Figure 1 is a flowchart of a first specific embodiment of a hierarchical knowledge representation method for skills in a dynamic task scenario provided by the present invention;

[0040] Figure 2 is a structural block diagram of the hierarchical knowledge representation method for operational skills in dynamic task scenarios;

[0041] Figure 3 is a case illustration of the model design;

[0042] Figure 4a is a case illustration of the first knowledge extraction;

[0043] Figure 4b is a case illustration of the second knowledge extraction.

[0044] Figure 5 is a structural block diagram of a hierarchical knowledge representation device for skills in a dynamic task scenario provided by an embodiment of the present invention. Detailed Implementation

[0045] The core of this invention is to provide a hierarchical knowledge representation method, device, electronic device, and storage medium for skills in dynamic task scenarios. By designing a multi-level knowledge representation system for operations, planning, tasks, and objects, a dynamic action library, skill library, and task library are established to improve the interpretability of robot operation skill learning and enhance the efficiency of robot operation task planning and skill transfer.

[0046] To enable those skilled in the art to better understand the present invention, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] Please refer to Figure 1, which is a flowchart of a first specific embodiment of a hierarchical knowledge representation method for skills in a dynamic task scenario provided by the present invention; the specific operation steps are as follows:

[0048] Step S101: Based on different semantic knowledge types, construct a knowledge hierarchical understanding framework using graph databases;

[0049] The framework for constructing a hierarchical knowledge understanding system based on different semantic knowledge types and using graph databases includes:

[0050] A hierarchical knowledge understanding framework is constructed using the graph database Neo4j. Neo4j stores data as a graph structure, where nodes represent entities and relations represent connections between entities, thus constructing a hierarchical knowledge understanding framework.

[0051] Step S102: Extract knowledge from the dynamic task scenario based on the preset task and construct triples;

[0052] The knowledge layering understanding framework includes:

[0053] Construct an ontology module, and use the ontology module to dynamically express the initial knowledge of the task to be parsed, expressing abstract conceptual knowledge, that is, the semantic knowledge of the entity itself;

[0054] A template module is constructed, which is used to operate on task templates to express various operation tasks or skill templates. Each task or skill is an operation process with an action sequence as its backbone, involving the scenario in which the operation occurs and the intelligent agent and object corresponding to each action primitive.

[0055] Build an instance module and use the instance module to perform instance operations. Each instance execution adds the operation execution parameters and timestamp to the corresponding template.

[0056] The knowledge layering understanding framework is divided into two levels: Level 1 is the environment level, which is used to express the static factors of operations, including objects, agents, and scenarios; Level 2 is the operation level, which is used to express the dynamic factors of operations, including tasks, skills, and actions.

[0057] The knowledge extraction from dynamic task scenarios based on preset tasks includes:

[0058] Based on the preset tasks, knowledge extraction is performed on dynamic task scenarios. The text-to-knowledge processing method is divided into four tasks. Task 1 is operation text collection, which includes text descriptions of operation processes. Task 2 is entity extraction, which uses the prior knowledge and reasoning ability of large language models to extract labels from text data and construct triples.

[0059] Step S103: Perform external knowledge fusion based on the triples and dynamically update the knowledge base to obtain a dynamic operation skill knowledge base.

[0060] Based on the triples constructed in Task 2, external knowledge fusion is performed using Task 3, and the context neighbors are retrieved from the general knowledge base using the standard query language to expand the knowledge related to the operation task.

[0061] Based on the knowledge expanded from Task 3, Task 4 is used to dynamically update the knowledge base. A unique identifier is attached to all actions in each template, and the existing knowledge is updated to the knowledge base to obtain a dynamic operation skill knowledge base.

[0062] The environment level represents static factors that express operations, including objects, agents, and scenes. Objects represent the objects of operations, typically physical entities; agents represent the subjects of operations, primarily humans and robots; and scenes represent the background space of operations.

[0063] The operation level is used to express the dynamic factors of an operation, including tasks, skills, and actions. Among them, tasks express the purpose of the operation; skills express the condensed prior knowledge in the operation; and actions express the underlying primitives of the operation.

[0064] This embodiment provides a hierarchical knowledge representation method for skills in dynamic task scenarios. After designing models and collecting data according to different scenario characteristics, an operational skill knowledge base is constructed. A multi-level knowledge expression system for operations, planning, tasks, and objects is designed, and dynamic action, skill, and task libraries are established. Together with static entity and scenario libraries, these constitute a robot skill operation knowledge base in dynamic scenarios. Combined with a real-time knowledge update mechanism, knowledge-guided and data-driven operational task reasoning is realized, improving the interpretability of robot operation skill learning and enhancing the efficiency of robot operation task planning and skill transfer.

[0065] Based on the above embodiments, this embodiment describes a hierarchical knowledge representation method for skills in a dynamic task scenario, as shown in Figure 2, as follows:

[0066] Knowledge base construction comprises two parts: model design and knowledge extraction. Model design utilizes the graph database Neo4j to build a hierarchical framework for understanding knowledge. Neo4j stores data as a graph structure, where nodes represent entities and relationships represent connections between entities, thus constructing the hierarchical knowledge understanding framework.

[0067] In terms of model design, operational skills are dynamically and semantically represented from a composite perspective of three modules and two levels. The three modules—Module 1, Module 2, and Module 3—represent different types of semantic knowledge. Module 1 is the ontology module, used to express common-sense knowledge; Module 2 is the template module, used to express operational task templates; and Module 3 is the instance module, used to express operational instance execution. The two levels are Level 1 and Level 2. Level 1 is the environment level, used to express the static factors of the operation, including objects, agents, and scenarios; and Level 2 is the operation level, used to express the dynamic factors of the operation, including tasks, skills, and actions. Knowledge extraction includes Task 1, Task 2, Task 3, and Task 4, representing the processing method from text to knowledge. The first task is to collect operational text, which includes a textual description of the operational process; the second task is entity extraction, which involves extracting labels from text data and constructing triples using the prior knowledge and reasoning capabilities of the Large Language Test (GPT) model; the third task is external knowledge fusion, which involves using the standard query language SparQL to retrieve neighbors from a commonsense knowledge base to expand knowledge; and the fourth task is dynamic knowledge base updating, which involves using SparQL to update the existing knowledge base.

[0068] The details are as follows:

[0069] First, the model design includes L1, L2, M1, M2, and M3, where L1 is level one, L2 is level two, M1 is module one, M2 is module two, and M3 is module three. Second, knowledge extraction includes T1, T2, T3, and T4, where T1 is task one, T2 is task one, T3 is task one, and T4 is task one.

[0070] Through model design, this embodiment hierarchically decouples the knowledge types and operational factors in the form of semantic knowledge. This includes Module 1, Module 2, and Module 3, as well as objects, agents, and scenarios in Level 1, and tasks, skills, and actions in Level 2. Based on this model design, knowledge extraction is performed by sequentially executing Task 1, Task 2, Task 3, and Task 4 to extract sufficient operational knowledge from the text dataset, large language model, and external knowledge base, storing it in the operational skill knowledge base.

[0071] During the operation of the robot, the operation skill knowledge base can provide it with ways to understand and plan the upper-level tasks. It helps the robot to learn from past experience and knowledge to break down complex tasks into simple tasks or even action sequences, and then hand them over to the lower-level motion planning and physical execution to realize the operation. Each successful operation is also stored as a new instance in the operation skill knowledge base, realizing a knowledge-guided robot operation closed loop.

[0072] As shown in Figure 3, a hierarchical knowledge understanding framework is constructed using the graph database Neo4j. Neo4j stores data as a graph structure, where nodes represent entities and relationships represent the connections between entities, thereby constructing a hierarchical knowledge understanding framework.

[0073] Knowledge comes in different types. For example, even though both are objects, they could be a conceptual description or a specific instance in a task. This requires a modular representation based on knowledge types, that is, from ontology to template to instance.

[0074] The ontology module is used to represent abstract conceptual knowledge, much like a dictionary, and similar to a classic commonsense knowledge base. The ontology stores the semantic knowledge of the entities themselves, organizing different types of entities from the perspective of conceptual definition, with strict categorization and hierarchical relationships.

[0075] Template modules are used to represent various operational tasks or skills. Each task or skill is an operational process with a sequence of actions as its backbone, involving the scenario in which the operation occurs and the agent and object corresponding to each action primitive. This means building logical or temporal relationships between nodes at different levels. Each entity in the template is associated with the entity of the same name in the ontology via instanceof, signifying the embodiment of the concept in the operation.

[0076] The instance module represents each execution of an operation, essentially acting as a log. An instance adds the operation's execution parameters and a timestamp to the corresponding template. All entities and relationships within an instance are mapped from those in the corresponding template. Each execution of each operation task generates an instance, so the knowledge within the instance is continuously updated during execution.

[0077] In addition to knowledge types, we can also perform hierarchical decoupling from the perspective of operational factors: the factors in operation are divided into two levels: environment and operation. The environment level refers to concrete and stable factors, including objects, agents and scenarios in the environment; the operation level refers to abstract and variable factors, including tasks, skills and actions in operation.

[0078] Based on the above two levels, the operation process is defined from the perspective of human cognition as follows: In a scenario, the intelligent agent aims to complete the task, invokes a series of skills during the operation process, performs a series of actions on a series of objects, and summarizes similar tasks into skills after the operation is completed.

[0079] As shown in Figures 4a and 4b, Task 1 involves collecting operational text. wikiHow is an open-source life guide, essentially a text dataset. Guides describing physical and simple operations were selected, meaning the objects being operated on undergo state or position changes during the task. This excludes virtual operations and, more importantly, abstract and complex non-operational tasks. We specifically chose guides for physical and simple operational tasks, where the objects undergo state or position changes during the task, such as "how to install a SIM card" and "how to install a flexible flat cable."

[0080] Task 2 is entity extraction. Using text data provided by wikiHow as input, the Large Language Model (GPT) extracts keywords based on prompts, outputting the verb sequence and the corresponding nouns and prepositions for each verb. Large Language Models are converter language models with trillions or more parameters. These models are trained on massive amounts of text data, demonstrating exceptional proficiency in understanding natural language and solving complex problems. Due to their extensive prior knowledge and reasoning capabilities, they facilitate entity extraction and updating. Next, Neo4j is used to construct triples based on the extracted entities and relations. Triples are a fundamental component of knowledge graphs, adopting an "entity-relationship-entity" format. The relation "contain" represents the parent-child relationship between tasks and actions; the relation "next" represents the sequential relationship between actions; the relation "subject" represents the subject (executor) of the action; the relation "object" represents the object of the action; and the relations "from / in / on / beside" represent the target or location of the action.

[0081] Task 3 involves external knowledge fusion. While entity nodes are singular, to obtain more relevant category context information, SparQL (a standard query language used to retrieve and manipulate structured data stored in triples, a format commonly used to describe and link data, especially in the Semantic Web) is used to connect the three types of entities—subject, object, and location—to the commonsense knowledge base DBpedia. In DBpedia, by limiting triple relationships to 'typeof', 'hypernym', and 'ingredient' to retrieve neighboring nodes, the category, context, and composition information of the local entities are obtained and added as new knowledge to the existing knowledge triples.

[0082] Task four involves dynamic updates to the knowledge base. Considering that actions with the same semantics may have different parameters across different tasks during robot operation execution instances, excessive coupling of the template data structure would occur if all these actions pointed to the same node. Therefore, a randomly generated hash suffix is ​​appended to all actions in each template as a unique identifier. Finally, the template data is imported into the triplet knowledge base using SparQL. The meta-nodes of the actions are added to the ontology to aggregate action entities with the same name in the template. Conversely, object entities in the template are also copied into the ontology and joined using `instanceof`.

[0083] This invention provides a hierarchical knowledge representation method for skills in dynamic task scenarios. It designs a multi-level knowledge representation system encompassing operations, planning, tasks, and objects, establishing dynamic action, skill, and task libraries. These, along with static entity and scenario libraries, constitute a robot skill operation knowledge base for dynamic scenarios. Combined with a real-time knowledge update mechanism, it enables knowledge-guided and data-driven operational task reasoning. The model design of the operation skill knowledge base decouples from operational elements and knowledge types, ensuring the stability and flexibility of the knowledge base architecture to support subsequent knowledge extraction. Knowledge extraction dynamically updates and completes the operation skill knowledge base with knowledge from different data sources through four tasks, particularly template-centric operational tasks. This ultimately completes the construction of the operation skill knowledge base. By designing models and collecting data according to different scenario characteristics, the operation skill knowledge base is constructed, improving the interpretability of robot operation skill learning and enhancing the efficiency of robot operation task planning and skill transfer.

[0084] Please refer to Figure 5, which is a structural block diagram of a hierarchical knowledge representation device for skills in a dynamic task scenario provided by an embodiment of the present invention; the specific device may include:

[0085] The framework building module 100 is based on different semantic knowledge types and utilizes graph databases to construct a knowledge layering understanding framework;

[0086] The knowledge extraction module 200 extracts knowledge from dynamic task scenarios based on preset tasks and constructs triples.

[0087] The knowledge base update module 300 performs external knowledge fusion based on the triples and dynamically updates the knowledge base to obtain a dynamic operation skills knowledge base.

[0088] This embodiment provides a hierarchical knowledge representation device for skills in dynamic task scenarios to implement the aforementioned hierarchical knowledge representation method for skills in dynamic task scenarios. Therefore, the specific implementation of the hierarchical knowledge representation device for skills in dynamic task scenarios can be found in the embodiment section of the aforementioned hierarchical knowledge representation method for skills in dynamic task scenarios. For example, the framework construction module 100, the knowledge extraction module 200, and the knowledge base update module 300 are respectively used to implement steps S101, S102, and S103 in the aforementioned hierarchical knowledge representation method for skills in dynamic task scenarios. Therefore, the specific implementation can be referred to the description of the corresponding embodiments, which will not be repeated here.

[0089] To implement the above embodiments, this application also proposes an electronic device, including: a processor and a memory communicatively connected to the processor; the memory stores computer execution instructions; the processor executes the computer execution instructions stored in the memory to implement the method provided in the foregoing embodiments.

[0090] To implement the above embodiments, this application also proposes a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the methods provided in the foregoing embodiments.

[0091] To implement the above embodiments, this application also proposes a computer program product, including a computer program that, when executed by a processor, implements the methods provided in the foregoing embodiments.

[0092] It should be understood that the present invention can be modified, combined, sub-combined, and substituted in various ways for the three modules, two levels, and four tasks, depending on design requirements and other factors. Any modifications or changes made without departing from the concept of the present invention are within the scope of protection of the present invention.

[0093] The collection, storage, use, processing, transmission, provision, and disclosure of user personal information involved in this application all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0094] It should be noted that personal information collected from users should be used for legitimate and reasonable purposes and should not be shared or sold outside of these legitimate uses. Furthermore, such collection / sharing should only be conducted after receiving the user's informed consent, including but not limited to notifying the user to read the user agreement / user notice and sign an agreement / authorization that includes authorization of relevant user information before the user uses the function. In addition, any necessary steps must be taken to protect and safeguard access to such personal information data and ensure that others with access to personal information data comply with their privacy policies and procedures.

[0095] This application is intended to provide an implementation scheme for users to selectively prevent the use or access to their personal information data. Specifically, this disclosure is intended to provide hardware and / or software to prevent or block access to such personal information data. Once personal information data is no longer needed, risks can be minimized by restricting data collection and deleting data. Furthermore, where applicable, such personal information is de-identified to protect user privacy.

[0096] In the foregoing descriptions of the embodiments, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0097] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0098] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0099] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0100] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0101] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0102] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0103] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this application.

Claims

1. A method for hierarchical knowledge representation of skills in dynamic task scenarios, characterized in that, The method comprises the following steps: Based on different semantic knowledge types, a knowledge hierarchical understanding framework is constructed using a graph database; Based on a preset task, knowledge is extracted from a dynamic task scene to construct a triple; Based on the triple, external knowledge is fused, and the knowledge base is dynamically updated to obtain a dynamic operation skill knowledge base.

2. The method of claim 1, wherein, The step of constructing the knowledge hierarchical understanding framework based on different semantic knowledge types comprises the following steps: A knowledge hierarchical understanding framework is constructed using a graph database Neo4j, which stores data as a graph structure, where nodes represent entities and relationships represent connections between entities, to construct the knowledge hierarchical understanding framework.

3. The method of claim 2, wherein, The knowledge hierarchical understanding framework comprises the following steps: An ontology module is constructed, which is used to dynamically express initial knowledge based on a task to be parsed; A template module is constructed, which is used to operate a task template; An instance module is constructed, which is used to execute an instance operation.

4. The method of claim 3, wherein, The method further comprises the following steps: The knowledge hierarchical understanding framework is divided into two levels, level one being an environment level used to express static factors of an operation, including objects, agents and scenes, and level two being an operation level used to express dynamic factors of an operation, including tasks, skills and actions. The step of extracting knowledge based on a preset task from a dynamic task scene comprises the following steps:

5. The method of claim 1, wherein, Based on a preset task, knowledge is extracted from a dynamic task scene, and a method for processing text into knowledge is divided into four tasks, where task one is operation text collection, including text description of an operation process; task two is entity extraction, which uses prior knowledge and reasoning ability of a large language model to extract labels from text data and construct triples. The step of fusing external knowledge based on the triple and dynamically updating the knowledge base to obtain a dynamic operation skill knowledge base comprises the following steps:

6. The method of claim 5, wherein, Based on the triple constructed in task two, task three is used to fuse external knowledge, and a standard query language is used to retrieve context neighbors from a general knowledge base to expand knowledge related to an operation task; Based on the expanded knowledge in task three, task four is used to dynamically update the knowledge base, a unique identifier is attached to all actions in each template, and the existing knowledge is updated to the knowledge base to obtain a dynamic operation skill knowledge base. The step of executing an instance operation using the instance module comprises the following steps: each time an instance is executed, an execution parameter and a timestamp of the operation are added based on the corresponding template.

7. The method of claim 3, wherein, The method comprises the following steps:

8. A device for hierarchical knowledge representation of skills in dynamic task scenarios, characterized in that, A framework construction module is used to construct a knowledge hierarchical understanding framework based on different semantic knowledge types using a graph database; A knowledge extraction module is used to extract knowledge from a dynamic task scene based on a preset task to construct a triple; A knowledge base updating module is used to fuse external knowledge based on the triple and dynamically update the knowledge base to obtain a dynamic operation skill knowledge base. The method comprises the following steps:

9. An electronic device, comprising: A processor and a memory connected in communication with the processor; The memory stores computer execution instructions; The processor executes the computer execution instructions stored in the memory to implement the method according to any one of claims 1-7. ​ 10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer-executable instructions, which, when executed by a processor, implement the method according to any one of claims 1-7.

Images