An information service method and system based on agent causal relationship reasoning

Abstract

The application discloses an information service method and system based on agent causal relationship reasoning, and belongs to the technical field of artificial intelligence. The method comprises the following steps: constructing a relationship extraction agent, performing entity extraction on information files in a field, extracting entity attributes, and constructing an entity library; performing basic relationship extraction according to the information of the entity library, and establishing a basic relationship library; constructing a causal reasoning agent, performing reasoning and mining according to the entity attributes and the relationship information in the basic relationship library, and establishing a causal relationship library; using a model training reasoning acceleration technology based on matrix low-rank decomposition to accelerate the processing of the agent base model; providing periodic business inspection services; and providing accurate question and answer feedback services. The application realizes causal reasoning and information mining by using intelligent means, and provides personalized services for different business scenarios.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, specifically to an information service method and system based on agent-based causal reasoning. Background Technology

[0002] With the deepening of digital transformation of business systems, various business systems have shifted from paper-based, heavily manual office models to computer- and internet-based digital office models. Information carriers have changed from paper documents to various types of data. How to extract useful information from massive amounts of data for business operations has been a long-term goal actively explored by various business systems. Especially in recent years, with the explosive development of artificial intelligence technology, how to utilize intelligent methods to mine data information and improve the efficiency of business systems has become a key focus of technological research in the information service field. Summary of the Invention

[0003] The technical objective of this invention is to address the above-mentioned shortcomings by providing an information service method and system based on intelligent agent causal reasoning, enabling the use of intelligent means to complete causal reasoning and information mining, and providing personalized services for different business scenarios.

[0004] The technical solution adopted by this invention to solve its technical problem is:

[0005] An information service method based on agent-based causal reasoning, the implementation of which includes the following steps:

[0006] Step 1: Construct a relation extraction agent to extract entities from information files within the domain, extract entity attributes, and build an entity library;

[0007] Step 2: Based on the entity database information, extract basic relationships and establish a basic relationship database;

[0008] Step 3: Construct a causal reasoning agent, perform reasoning and mining based on entity attributes and relation information in the basic relation database, and establish a causal relation database;

[0009] Step 4: Accelerate the intelligent agent base model by using model training and inference acceleration technology based on matrix low-rank decomposition to reduce the reliance on computing resources during model inference.

[0010] Step 5: Provide periodic business inspection services, set inspection items and indicators according to the routine inspection requirements of the business domain, build an inspection intelligent agent, and conduct periodic inspections of the causal relationship database;

[0011] Step 6: Provide precise question-and-answer feedback services, build a question-and-answer intelligent agent, receive targeted needs input by users, query the causal relationship database, and perform reasoning to generate conclusions.

[0012] This method constructs intelligent agents with customized functions to extract basic relationships and causal reasoning from information files within the domain, builds an information database that meets the causal relationships within the business domain, and provides periodic business inspection services and precise question-and-answer feedback services. It realizes the use of intelligent means to complete causal reasoning and information mining, and provides personalized services for different business scenarios by constructing intelligent agents with business-specific service capabilities.

[0013] Furthermore, in step 1, before entity extraction, the entity identification rules in the domain are input to the intelligent agent by inputting prompt words, and the scope of entity attribute extraction is standardized to ensure the accuracy of entity identification and business matching.

[0014] Furthermore, in step 2, during the basic relationship extraction process, full-scale relationship extraction is performed based on entities and entity attributes in the entity library, or targeted relationship extraction is performed according to business needs.

[0015] Furthermore, in step 3, the types of causal relationships include explicit causal relationships and implicit causal relationships.

[0016] Explicit causal relationships represent the associations that can be inferred from explicit relationships in the existing basic relational database. The mining of explicit causal relationships is obtained through mining multiple layers of direct associations or through statistical mining of quantitative indicators. This kind of capability can be achieved by building a training corpus set with a certain number of business cases and fine-tuning the base model that drives the causal reasoning agent.

[0017] Implicit causal relationships refer to the overlapping relationships that are further explored beyond the consideration of multiple direct relationships, including time and space dimensions. In addition to fine-tuning the base model of the agent driving causal reasoning through case studies, it is also necessary to formulate a workflow to standardize the direction and focus of the agent's information mining in order to meet business needs.

[0018] Furthermore, in step 4, the agent base model is accelerated using a model training and inference acceleration technique based on low-rank matrix decomposition. The specific steps are as follows:

[0019] Step 4.1: From the large-dimensional matrix in the base model Selecting evenly Column Formation Matrix ;

[0020] Step 4.2: From Before the election Row forming matrix ;

[0021] Step 4.3: For Perform singular value decomposition and determine the truncation rank of the matrix based on the truncation error. ,get ,if ,and ,Will Add to Repeat steps 4.1 and 4.2.

[0022] Step 4.4: From Selecting evenly Row forming matrix ;

[0023] Step 4.5: The formed matrix and ,satisfy ,in , .

[0024] Furthermore, in step 5, the periodic business inspection service specifically includes:

[0025] The inspection subjects and indicators are input into the inspection agent through prompt words. The inspection agent performs periodic inspections of the causal relationship database according to the inspection rules. If target information is found, alarm prompts and visual displays are provided.

[0026] The inspection process and the causal relationship database construction process are parallel processes. During the inspection process, the content of the causal relationship database can be updated in real time according to changes in the information files.

[0027] Furthermore, in step 6, the precise question-and-answer feedback service specifically includes:

[0028] The question-answering agent queries a causal relationship database based on the user's specific input, and generates conclusions and visualizations through reasoning.

[0029] This invention also claims an information service system based on agent-based causal reasoning, comprising:

[0030] The entity library is used to complete the entity extraction function and collect and store the extracted entities and entity attributes.

[0031] The basic relation library is used to extract the basic relations of entities and store the extracted basic relations;

[0032] The causal relationship database is used to perform inference and mining based on entity attributes and relationship information in the basic relationship database to obtain causal relationships.

[0033] The intelligent agent kit is used to build intelligent agents with customized functions according to the needs of the system, including relation extraction intelligent agents, causal reasoning intelligent agents, inspection intelligent agents, question answering intelligent agents, etc.

[0034] The model acceleration module uses a model training and inference acceleration technology based on matrix low-rank decomposition to accelerate the processing of the agent base model, thereby compressing and decomposing large-dimensional matrices in the model to reduce the reliance on computing resources during the model inference process.

[0035] The application service module is used to provide periodic business inspection services and precise question-and-answer feedback services according to the set requirements;

[0036] The system uses the methods described above to achieve causal reasoning and information services.

[0037] The present invention also claims an information service device based on agent-based causal reasoning, characterized in that it comprises: at least one memory and at least one processor;

[0038] The at least one memory is used to store a machine-readable program;

[0039] The at least one processor is used to call the machine-readable program to implement the above method.

[0040] The present invention also claims a computer-readable medium, characterized in that the computer-readable medium stores computer instructions that, when executed by a processor, enable the implementation of the above-described method.

[0041] Compared with the prior art, the information service method and system based on agent-based causal reasoning of the present invention have the following advantages:

[0042] This invention constructs customized intelligent agents with business attributes, such as relation extraction agents and causal reasoning agents, to extract basic relationships and infer causal relationships from information files within a specific domain. This process sequentially builds a basic relation database and a causal relationship database, forming an information repository that satisfies the causal logic relationships within the business domain. Furthermore, it constructs inspection agents and question-answering agents to provide periodic business inspection services and precise question-and-answer feedback services, automating result output and visualization. This creates an information service system based on agent-driven causal reasoning, enabling the use of intelligent methods to complete causal reasoning and information mining, and providing personalized services to meet the needs of different business scenarios. Attached Figure Description

[0043] Figure 1 This is a flowchart illustrating an information service method based on agent-based causal reasoning, provided in one embodiment of the present invention.

[0044] Figure 2 This is a schematic diagram of a periodic business inspection service provided in one embodiment of the present invention;

[0045] Figure 3This is a schematic diagram of a precise question-and-answer feedback service provided by an embodiment of the present invention. Detailed Implementation

[0046] The present invention will be further described below with reference to specific embodiments.

[0047] This invention provides an information service method based on intelligent agent causal reasoning. By constructing customized intelligent agents with business attributes, such as relation extraction agents and causal reasoning agents, basic relation extraction and causal reasoning extraction are performed on information files within a specific domain. This sequentially builds a basic relation database and a causal relation database, forming an information database that satisfies causal relationships within the business domain. Furthermore, an inspection agent and a question-answering agent are constructed to provide periodic business inspection services and precise question-and-answer feedback services. The method automates result output and visualization, enabling causal reasoning and information mining using intelligent means, and providing personalized services to meet the needs of different business scenarios.

[0048] like Figure 1 As shown, firstly, a relation extraction agent is constructed to build an entity database and extract basic relations, establishing a basic relation database; then, a causal reasoning agent is constructed to establish a causal relation database; next, inspection subjects and indicators are set according to the routine inspection requirements of the business domain, and an inspection agent is constructed to conduct periodic inspections of the causal relation database, providing periodic business inspection services; finally, a question-answering agent is constructed to receive targeted needs input by users, query the causal relation database, and perform reasoning to generate conclusions, providing accurate question-and-answer feedback services.

[0049] The specific steps for implementing this method are as follows:

[0050] Step 1: Construct a relation extraction agent to extract entities from information files within the domain, extract entity attributes, and build an entity library. Before entity extraction, the agent needs to be prompted with domain-specific entity identification rules and the scope of entity attribute extraction should be standardized to ensure the accuracy of entity identification and business relevance. For example, prompt the relation extraction agent with the following prompt: Please extract entities based on "person (name)," "account (card number)," "property (address)," "vehicle (license plate number)," "telephone (number)," and "company (name)," and extract the corresponding entity attributes to build an entity library.

[0051] Step 2: The relation extraction agent extracts basic relationships based on the entity database information to build a basic relation database. During the relationship extraction process, full-scale relation extraction can be performed based on entities and entity attributes in the entity database, or targeted relation extraction can be performed according to business needs. For example, inputting the prompt "extract relations centered on the entity 'person'" into the relation extraction agent will extract relationships such as "person-to-person," "person-to-account," "person-to-real estate," "person-to-vehicle," "person-to-phone," and "person-to-company" from the basic relation database to build the basic relation database.

[0052] Step 3: Construct a causal reasoning agent, and perform reasoning and mining based on entity attributes and relationship information in the basic relation database to establish a causal relation database. When extracting causal relations, it is necessary to extract two types separately: explicit causal relations and implicit causal relations.

[0053] Explicit causal relationships refer to the associations that can be inferred from explicit relationships in the existing basic relational database. This type of relationship can be mined through multi-layer direct association mining or through quantitative index statistical mining. This capability can be achieved simply by building a training corpus based on a certain number of business cases and fine-tuning the base model that drives the causal reasoning agent.

[0054] Implicit causal relationships refer to considering not only multiple direct relationships but also overlapping relationships in time and space. Establishing this capability requires not only fine-tuning the foundational model driving the causal reasoning agent through case studies but also defining workflows to standardize the direction and focus of the agent's information mining to meet business needs. For example, to determine whether two people have met at a certain location, a workflow needs to be established: "Determine if the two people appeared or arrived in a certain city within the same time period - Query the two people's accommodation and consumption records in that city to determine if the two people appeared on the same street at the same time - Query whether the two people had telephone or other contact during this overlapping time - Analyze the mined information and output conclusions, while simultaneously visualizing all the mined relationship information through a graph."

[0055] Step 4: Accelerate the agent base model using a model training and inference acceleration technique based on low-rank matrix decomposition. This involves compressing and decomposing large-dimensional matrices in the model to reduce the computational resource dependence during inference. The specific steps are as follows:

[0056] Step 4.1: From the large-dimensional matrix in the base model Selecting evenly Column Formation Matrix .

[0057] Step 4.2: From Before the election Row forming matrix .

[0058] Step 4.3: For Perform singular value decomposition and determine the truncation rank of the matrix based on the truncation error. ,get ,if ,and ,Will Add to Repeat steps 4.1 and 4.2.

[0059] Step 4.4: From Selecting evenly Row forming matrix .

[0060] Step 4.5: The formed matrix and ,satisfy ,in , .

[0061] Step 5: Provide periodic business inspection services. Based on the routine inspection requirements of the business domain, set inspection items and indicators, construct an inspection intelligence agent, and conduct periodic inspections of the causal relationship database. For example... Figure 2 As shown, the inspection items and indicators are input into the inspection agent via prompts. The inspection agent performs periodic inspections of the causal relationship database according to the inspection rules. If target information is found, an alarm is triggered and a visual display is provided. The inspection process and the causal relationship database construction process are parallel. During the inspection, the content of the causal relationship database can be updated in real time according to changes in the information files.

[0062] Step 6: Provide precise question-and-answer feedback services, build a question-and-answer intelligent agent, receive targeted user input, query a causal relationship database, and perform reasoning to generate conclusions. For example... Figure 3 As shown, the question-answering agent queries a causal relationship database and generates conclusions and visualizations based on the user's specific input requirements.

[0063] This invention also provides an information service system based on agent-based causal reasoning, comprising:

[0064] The entity library is used to complete the entity extraction function and collect and store the extracted entities and entity attributes.

[0065] The basic relation library is used to extract the basic relations of entities and store the extracted basic relations;

[0066] The causal relationship database is used to perform inference and mining based on entity attributes and relationship information in the basic relationship database to obtain causal relationships.

[0067] The intelligent agent kit is used to build intelligent agents with customized functions according to the needs of the system, including relation extraction intelligent agents, causal reasoning intelligent agents, inspection intelligent agents, question answering intelligent agents, etc.

[0068] The model acceleration module uses a model training and inference acceleration technology based on matrix low-rank decomposition to accelerate the processing of the agent base model, thereby compressing and decomposing large-dimensional matrices in the model to reduce the reliance on computing resources during the model inference process.

[0069] The application service module is used to provide periodic business inspection services and precise question-and-answer feedback services according to the set requirements;

[0070] The system implements causal reasoning and information services through the agent-based causal reasoning information service method described in the above embodiments.

[0071] First, a relation extraction agent is constructed to extract entities from information files within the domain and extract entity attributes to build an entity database. Then, based on the entity database information, basic relations are extracted to establish a basic relation database. Next, a causal reasoning agent is constructed to perform reasoning and mining based on entity attributes and relation information in the basic relation database, extract causal relationships, integrate basic relation information, and establish a causal relation database. Information services are provided in two ways: periodic business inspection service and precise question-and-answer feedback service. The periodic business inspection service is used by business systems to set inspection subjects and indicators according to routine inspection requirements of the business domain. The inspection agent drives periodic inspections of the causal relation database, and if target information is found, alarm prompts and visual displays are provided. The precise question-and-answer feedback service is used in business scenarios where, when receiving targeted demand input, the question-and-answer agent queries the causal relation database and generates conclusions and visual displays through reasoning.

[0072] The specific steps for this system to provide information services are as follows:

[0073] Step 1: Construct a relation extraction agent to extract entities from information files within the domain, extract entity attributes, and build an entity library. Before entity extraction, the agent needs to be fed domain-specific entity identification rules via input prompts, and the scope of entity attribute extraction should be standardized to ensure the accuracy of entity identification and business relevance.

[0074] Step 2: The relation extraction agent extracts basic relationships based on the entity database information to build a basic relation database. During the relationship extraction process, full-scale relation extraction can be performed based on entities and entity attributes in the entity database, or targeted relation extraction can be performed according to business needs. For example, inputting the prompt "extract relations centered on the entity 'person'" into the relation extraction agent will extract relationships such as "person-to-person," "person-to-account," "person-to-real estate," "person-to-vehicle," "person-to-phone," and "person-to-company" from the basic relation database to build the basic relation database.

[0075] Step 3: Construct a causal reasoning agent, and perform reasoning and mining based on entity attributes and relationship information in the basic relation database to establish a causal relation database. When extracting causal relations, it is necessary to extract two types separately: explicit causal relations and implicit causal relations.

[0076] Explicit causal relationships refer to the associations that can be inferred from explicit relationships in the existing basic relational database. This type of relationship can be mined through multi-layer direct association mining or through quantitative index statistical mining. This capability can be achieved simply by building a training corpus based on a certain number of business cases and fine-tuning the base model that drives the causal reasoning agent.

[0077] Implicit causal relationships mean that in addition to considering multiple layers of direct relationships, it is also necessary to consider the overlapping relationships in the time and space dimensions. In addition to fine-tuning the base model of the agent driving causal reasoning through case studies, it is also necessary to formulate a workflow to standardize the direction and focus of the agent in mining information in order to meet business needs.

[0078] Step 4: Accelerate the agent base model using a model training and inference acceleration technique based on low-rank matrix decomposition. This involves compressing and decomposing large-dimensional matrices in the model to reduce the computational resource dependence during inference. The specific steps are as follows:

[0079] Step 4.1: From the large-dimensional matrix in the base model Selecting evenly Column Formation Matrix .

[0080] Step 4.2: From Before the election Row forming matrix .

[0081] Step 4.3: For Perform singular value decomposition and determine the truncation rank of the matrix based on the truncation error. ,get ,if ,and ,Will Add to Repeat steps 4.1 and 4.2.

[0082] Step 4.4: From Selecting evenly Row forming matrix .

[0083] Step 4.5: The formed matrix and ,satisfy ,in , .

[0084] Step 5: Provide periodic business inspection services. Based on the routine inspection requirements of the business domain, set inspection items and indicators, construct an inspection intelligence agent, and conduct periodic inspections of the causal relationship database. For example... Figure 2 As shown, the inspection items and indicators are input into the inspection agent via prompts. The inspection agent performs periodic inspections of the causal relationship database according to the inspection rules. If target information is found, an alarm is triggered and a visual display is provided. The inspection process and the causal relationship database construction process are parallel. During the inspection, the content of the causal relationship database can be updated in real time according to changes in the information files.

[0085] Step 6: Provide precise question-and-answer feedback services, build a question-and-answer intelligent agent, receive targeted user input, query a causal relationship database, and perform reasoning to generate conclusions. For example... Figure 3 As shown, the question-answering agent queries a causal relationship database and generates conclusions and visualizations based on the user's specific input requirements.

[0086] This invention also provides an information service device based on agent-based causal reasoning, characterized in that it includes: at least one memory and at least one processor;

[0087] The at least one memory is used to store a machine-readable program;

[0088] The at least one processor is used to invoke the machine-readable program to implement the information service method based on agent-based causal reasoning described in the above embodiments.

[0089] This invention also provides a computer-readable medium storing computer instructions. When executed by a processor, these computer instructions implement the information service method based on agent-based causal reasoning described in the above embodiments. Specifically, a system or apparatus equipped with a storage medium storing software program code that implements the functions of any of the embodiments described above can be provided, enabling the computer (or CPU or MPU) of the system or apparatus to read and execute the program code stored in the storage medium.

[0090] In this case, the program code read from the storage medium can itself implement the function of any of the above embodiments, and therefore the program code and the storage medium storing the program code constitute part of the present invention.

[0091] Storage media embodiments for providing program code include floppy disks, hard disks, magneto-optical disks, optical disks (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD+RW), magnetic tapes, non-volatile memory cards, and ROMs. Alternatively, program code can be downloaded from a server computer via a communication network.

[0092] Furthermore, it should be clear that not only can the program code read by the computer be executed, but also the operating system or other components operating on the computer can be instructed based on the program code to perform some or all of the actual operations, thereby realizing the function of any of the embodiments described above.

[0093] Furthermore, it is understood that the program code read from the storage medium is written to the memory set in the expansion board inserted into the computer or to the memory set in the expansion unit connected to the computer. Then, based on the instructions of the program code, the CPU or other components installed on the expansion board or expansion unit execute some and all of the actual operations, thereby realizing the function of any of the embodiments described above.

[0094] The present invention has been shown and described in detail above with reference to the accompanying drawings and preferred embodiments. However, the present invention is not limited to these disclosed embodiments. Based on the above embodiments, those skilled in the art will know that more embodiments of the present invention can be obtained by combining the code review methods in the different embodiments. These embodiments are also within the protection scope of the present invention.

Claims

1. An information service method based on agent-based causal reasoning, characterized in that, The implementation of this method includes the following steps: Step 1: Construct a relation extraction agent to extract entities from information files within the domain, extract entity attributes, and build an entity library; Step 2: Based on the entity database information, extract basic relationships and establish a basic relationship database; Step 3: Construct a causal reasoning agent, perform reasoning and mining based on entity attributes and relation information in the basic relation database, and establish a causal relation database; Step 4: Accelerate the intelligent agent base model by using model training and inference acceleration technology based on matrix low-rank decomposition to reduce the reliance on computing resources during model inference. Step 5: Provide periodic business inspection services, set inspection items and indicators according to the routine inspection requirements of the business domain, build an inspection intelligent agent, and conduct periodic inspections of the causal relationship database; Step 6: Provide precise question-and-answer feedback services, build a question-and-answer intelligent agent, receive targeted needs input by users, query the causal relationship database, and perform reasoning to generate conclusions; In step 4, the agent base model is accelerated using a model training and inference acceleration technique based on low-rank matrix decomposition. The specific steps are as follows: Step 4.1: From the large-dimensional matrix Z in the base model m×n A matrix is ​​formed by uniformly selecting r0 columns. Step 4.2: Select the first r0 rows from U0 to form a matrix. Step 4.3: Perform singular value decomposition on W0, determine the truncation rank r of the matrix based on the truncation error, and obtain... If r ≤ r0 / 3, and Increase r0 to 2r0, and repeat steps 4.1 and 4.2; Step 4.4: From Z m×n A matrix is ​​formed by uniformly selecting r0 rows from the middle. Step 4.5: The resulting matrix U m×r and V r×n Satisfying Z m×n ≈U m×r ·V r×n ,in 2. The information service method based on agent-based causal reasoning according to claim 1, characterized in that, In step 1, before entity extraction, the entity identification rules in the domain are input to the intelligent agent by inputting prompt words, and the scope of entity attribute extraction is standardized to ensure the accuracy of entity identification and business matching.

3. The information service method based on agent-based causal reasoning according to claim 1, characterized in that, In step 2, during the basic relationship extraction process, full-scale relationship extraction is performed based on entities and entity attributes in the entity database, or targeted relationship extraction is performed according to business needs.

4. The information service method based on agent-based causal reasoning according to claim 1, characterized in that, Step 3, the types of causal relationships include explicit causal relationships and implicit causal relationships. Explicit causal relationships represent the associations that can be inferred from explicit relationships in the existing basic relational database. Explicit causal relationships are mined through multi-level direct association mining or through statistical mining of quantitative indicators. Implicit causal relationships refer to overlapping relationships that include time and space dimensions, which are further explored beyond multiple layers of direct correlation.

5. The information service method based on agent-based causal reasoning according to claim 1, characterized in that, In step 5, the periodic business inspection service specifically includes: The inspection subjects and indicators are input into the inspection agent through prompt words. The inspection agent performs periodic inspections of the causal relationship database according to the inspection rules. If target information is found, alarm prompts and visual displays are provided. The inspection process and the causal relationship database construction process are parallel processes. During the inspection process, the content of the causal relationship database is updated in real time according to the changes in the information files.

6. The information service method based on agent-based causal reasoning according to claim 1, characterized in that, In step 6, the precise question-and-answer feedback service specifically includes: The question-answering agent queries a causal relationship database based on the user's specific input, and generates conclusions and visualizations through reasoning.

7. An information service system based on agent-based causal reasoning, characterized in that, include: The entity library is used to complete the entity extraction function and collect and store the extracted entities and entity attributes. The basic relation library is used to extract the basic relations of entities and store the extracted basic relations; The causal relationship database is used to perform inference and mining based on entity attributes and relationship information in the basic relationship database to obtain causal relationships. The intelligent agent kit is used to build intelligent agents with customized functions as needed for the system, including relation extraction intelligent agents, causal reasoning intelligent agents, inspection intelligent agents, and question answering intelligent agents; The model acceleration module uses a model training and inference acceleration technology based on matrix low-rank decomposition to accelerate the processing of the agent base model, thereby compressing and decomposing large-dimensional matrices in the model to reduce the reliance on computing resources during the model inference process. The application service module is used to provide periodic business inspection services and precise question-and-answer feedback services according to the set requirements; The system achieves causal reasoning and information services through the method described in any one of claims 1 to 6.

8. An information service device based on agent-based causal reasoning, characterized in that, include: At least one memory and at least one processor; The at least one memory is used to store a machine-readable program; The at least one processor is configured to invoke the machine-readable program to implement the method according to any one of claims 1 to 6.

9. A computer-readable medium, characterized in that, The computer-readable medium stores computer instructions that, when executed by a processor, enable the implementation of the method described in any one of claims 1 to 6.

Images