Pathology multi-information comprehensive diagnosis device based on multi-agent cooperation

Abstract

The invention provides a pathological multi-information comprehensive diagnosis device based on multi-agent cooperation, which comprises the following steps: acquiring update information corresponding to each specified information channel, distributing each update information to a corresponding agent for information storage in combination with a corresponding information attribute, generating an information map corresponding to each agent in a specified period, the information atlas is modified according to management content uploaded by doctors, a multi-agent platform of each doctor is constructed, when the doctors submit the patient information, relevant diagnosis information of the patient information is searched in the multi-agent platform, simulation diagnosis is conducted on the patient information through the relevant diagnosis information, and a plurality of physical conditions are obtained; meanwhile, the diagnosis effectiveness corresponding to each physical condition is determined and displayed, efficient integration of pathological information, personalized intelligent agent management, accurate query and diagnosis effectiveness quantification are achieved, and scientific and reliable pathological diagnosis auxiliary support is provided for doctors.

Description

Technical Field

[0001] This invention relates to the field of comprehensive pathological diagnosis technology, and in particular to a comprehensive pathological diagnosis device based on multi-agent collaboration. Background Technology

[0002] In the medical field, besides the pathological images themselves, patient history, biochemical indicators, sequencing, and radiological imaging all play a crucial role in the pathologist's final diagnosis. Pathological diagnosis is the core step in confirming a disease, playing an irreplaceable role, especially in tumor classification and the differentiation of complex cases. With the development of medical technology, the information dimensions required for pathological diagnosis are constantly expanding, encompassing pathological slide images, gene testing data, clinical history, imaging reports, and previous diagnostic records, among other types of information. Furthermore, this information is scattered across different hospital systems and physicians' personal experience databases, leading to the following key challenges in pathological diagnosis:

[0003] 1. Information integration is difficult. There is a lack of a unified mechanism for collecting existing pathology diagnosis-related information, and information retrieval and integration are time-consuming and labor-intensive.

[0004] 2. Lack of personalization in intelligent agent management: Although some auxiliary diagnostic systems have introduced intelligent agent technology, the intelligent agents can only process general information according to fixed rules and cannot be personalized according to the physician's professional direction, diagnostic habits and clinical needs.

[0005] 3. The accuracy of diagnostic information retrieval is insufficient. Existing query systems are mostly based on single keywords for retrieval, making it difficult to achieve accurate matching by combining multi-dimensional information about patients.

[0006] 4. The auxiliary diagnostic results lack validity assessment, directly outputting a single diagnostic conclusion without quantitatively assessing the reliability of the diagnostic results, making it impossible for physicians to determine whether the diagnostic conclusion is based on sufficient information.

[0007] Therefore, the present invention provides a pathological multi-information integrated diagnostic device based on multi-agent collaboration. Summary of the Invention

[0008] This invention is a pathological multi-information integrated diagnostic device based on multi-agent collaboration, which realizes efficient integration of pathological information, personalized intelligent agent management, accurate query and effective quantification of diagnosis, providing physicians with scientific and reliable pathological diagnostic support.

[0009] This invention provides a pathological multi-information integrated diagnostic device based on multi-agent collaboration, comprising:

[0010] The information aggregation module is used to acquire the updated information corresponding to each specified information channel, and allocate each updated information to the corresponding intelligent agent for information storage according to the information attributes corresponding to each updated information.

[0011] The intelligent agent management module is used to generate an information graph corresponding to each intelligent agent within a specified period, and modify the information graph according to the management content uploaded by the doctor, so as to build a multi-intelligent agent platform for each doctor.

[0012] The intelligent query module is used to search for relevant diagnostic information of the patient in the multi-agent platform when the doctor submits patient information;

[0013] The comprehensive auxiliary module is used to perform a simulated diagnosis on the patient information using the relevant diagnostic information, obtain several physical conditions of the patient, and determine and display the diagnostic validity corresponding to each physical condition.

[0014] In one feasible approach

[0015] The information aggregation module includes:

[0016] The channel review unit is used to obtain the channel characteristics and information sources corresponding to each specified information channel, construct the channel factors corresponding to each specified information channel, and set the corresponding trust level for each specified information channel based on the doctor's evaluation results of each channel factor.

[0017] The information filtering unit is used to construct corresponding channel generation information based on the real-time data corresponding to each specified channel, and to verify the information of each channel generation information using the corresponding trust level and the storage information corresponding to each intelligent agent, so as to obtain the information fusion features between each channel generation information and different intelligent agents.

[0018] An information update unit is used to filter channel-generated information whose information fusion features meet the filtering criteria, and to perform information compensation on the channel-generated information in combination with the trust level of the specified information channel in which each channel-generated information is located, so as to obtain the updated information corresponding to each specified channel.

[0019] A storage unit is allocated for determining several pre-matching agents corresponding to each update information based on several information fusion features corresponding to each update information, and for matching and verifying the information attributes corresponding to each update information with each pre-matching agent to determine and store the agent corresponding to each update information.

[0020] In one feasible approach

[0021] The process of compensating for information generated by the aforementioned channels includes:

[0022] Historical update information corresponding to each specified channel is obtained, an information structure tree corresponding to the specified channel is constructed, branch identification is performed on each information structure tree to obtain several information subtrees corresponding to each specified information channel, and the hierarchical relationship between different information subtrees is determined.

[0023] The channel-generated information is mapped to the corresponding information structure tree to obtain several related information subtrees corresponding to the channel-generated information. The information influence of each channel-generated information on the information structure tree is determined by combining the corresponding hierarchical relationship.

[0024] Information generated by target channels with an impact value higher than a specified impact value is filtered out, and the scope of the impact of each target channel's generated information on the corresponding information structure tree is determined by combining the trust level corresponding to each specified information channel.

[0025] Based on the screening criteria, each target channel generated information is converted into several information conditions within the scope of the information's influence, and the information semantics corresponding to each information condition are determined.

[0026] Semantic compensation is performed on invalid information conditions with missing semantic information to generate alternative information conditions for each invalid information condition;

[0027] Replace the corresponding invalid information condition with the alternative information condition that has the least impact on the structure of the information structure tree.

[0028] In one feasible approach

[0029] The intelligent agent management module includes:

[0030] The logic management unit is used to acquire the storage information corresponding to each of the intelligent agents within a specified period, determine the information coverage area and information coverage range of the storage information in combination with the intelligent agent attributes corresponding to each of the intelligent agents, and deduce the information logic relationship between several storage information corresponding to each of the intelligent agents.

[0031] The dynamic tracking unit is used to acquire the auxiliary operation data corresponding to each of the intelligent agents within a specified period, construct the extraction information dynamic and update information dynamic of each of the intelligent agents within the specified period, and identify the first dynamic logic corresponding to the extraction information dynamic and the second dynamic logic corresponding to the update information dynamic in the information logic relationship.

[0032] The graph construction unit is used to optimize the first dynamic logic and the second dynamic logic in the information logic relationship, construct the information structure corresponding to the intelligent agent according to the optimization result, and map each of the stored information into the information structure to obtain the information graph corresponding to each intelligent agent.

[0033] The manual optimization unit is used to identify several management items contained in the management content uploaded by the doctor and the management purpose corresponding to each management item, find the corresponding management agent according to the management item, optimize the management agent to the management purpose, and generate the doctor's multi-agent platform.

[0034] In one feasible approach

[0035] The intelligent query module includes:

[0036] An identity determination unit is used to identify the patient identity corresponding to the patient information submitted by the doctor in the multi-agent platform, search for several pieces of intelligent information related to the patient identity in each of the intelligent agents, and determine the source of the intelligent agent corresponding to each piece of intelligent information.

[0037] The semantic analysis unit is used to convert the corresponding intelligent information into corresponding medical semantics according to the intelligent agent function corresponding to each intelligent agent source, generate medical description text of the patient identity, and perform semantic optimization on the medical description text in the multi-agent platform to obtain several auxiliary diagnostic semantics of the patient identity.

[0038] The information integration unit is used to utilize the auxiliary diagnostic semantic recognition to identify the missing semantics corresponding to the intelligent information, perform multi-dimensional semantic expansion on the intelligent information to obtain semantically complete and effective intelligent agent information, construct relevant diagnosis and treatment information of the patient's identity based on the effective intelligent agent information, and display it.

[0039] In one feasible approach

[0040] The intelligent information is subjected to multi-dimensional semantic expansion to obtain semantically complete and effective intelligent agent information, including:

[0041] Based on the intelligent information, several preliminary diagnostic and treatment records for the patient's identity are constructed;

[0042] Based on the missing semantics, the information defects between different preliminary diagnostic information are determined, and at the same time, the information logical relationship between different preliminary diagnostic information is constructed according to the intelligent agent function corresponding to each intelligent agent source.

[0043] In a multi-agent platform, a diagnostic knowledge base for the doctor is constructed, and alternative information corresponding to each information defect is searched in the diagnostic knowledge base according to the information logical relationship.

[0044] Each of the aforementioned alternative information is used to perform alternative optimization on the information defect, and the target optimized semantics with complete semantics are selected based on the optimized semantics corresponding to each optimization result.

[0045] Effective agent information is constructed based on the optimization process corresponding to the target optimization semantics.

[0046] In one feasible approach

[0047] Also includes:

[0048] The platform display module is used to display the extracted information in the multi-agent platform.

[0049] In one feasible approach

[0050] The integrated auxiliary module includes:

[0051] The feature generation unit is used to obtain the patient information uploaded by the doctor, combine it with the corresponding relevant diagnostic information to construct several health status features for the patient, and search for the health standard corresponding to each health status feature in the multi-agent platform.

[0052] The integrated simulation unit is used to evaluate the health status characteristics using the health standards, obtain several non-healthy information and several healthy information of the patient, and construct a virtual health model of the patient based on the non-healthy information and the healthy information;

[0053] The simulation diagnostic unit is used to map the virtual health model to each specified scenario, control the virtual health model to perform several specified situations in each specified scenario, obtain several physical conditions of the patient, and determine the health value corresponding to each physical condition according to the health standard.

[0054] The diagnostic integration unit is used to use the current patient information to make disease judgments for each of the health values, determine the probability of each suspected disease, determine the diagnostic validity of each physical condition, generate a health diagnosis probability report, and display it.

[0055] In one feasible approach

[0056] Also includes:

[0057] An auxiliary tracing unit is used to filter several target physical conditions in the health diagnosis probability report whose diagnostic validity is higher than the specified validity.

[0058] For each of the target physical conditions, the source of the target-related diagnostic information corresponding to the target physical condition is determined by the intelligent agent.

[0059] Search for the relevant diagnostic information of the target in the information source agent, generate auxiliary reference information and display it.

[0060] The beneficial effects of this invention are as follows: By automatically acquiring updated information through multiple designated information channels and allocating it to corresponding intelligent agents based on information attributes, it solves the problem of traditionally scattered information, reduces the time physicians spend manually integrating information, and achieves classified information storage. Subsequent retrieval allows for quick location of the corresponding intelligent agent. Simultaneously, encrypted storage of intelligent agents ensures data security, version management facilitates information traceability, and intelligent agent information graphs are generated periodically to intuitively present information relationships. Furthermore, physicians can modify the graphs according to management content and build their own multi-agent platforms, ensuring the information graphs align with physicians' professional directions and diagnostic habits. The platform also allows for the customization of intelligent agents. The platform displays and pushes data, and allows for flexible adjustments to the update cycle. When doctors use the platform, it analyzes query requirements, performs multi-dimensional weighted matching to find relevant diagnostic information, and then categorizes, sorts, and deduplicates the results to reduce interference from irrelevant information. This quickly provides doctors with valuable diagnostic references such as patient information and similar cases. Finally, using the relevant diagnostic information, the platform employs a multi-model fusion algorithm to obtain several aspects of the patient's physical condition, avoiding the limitations of a single model diagnosis. It also combines the sufficiency of diagnostic evidence, the reliability of information, and the confidence level of the model to determine the validity of the diagnosis and visualize it, helping doctors judge the reliability of the results, improving diagnostic accuracy, and reducing the risk of missed or misdiagnosed diagnoses.

[0061] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0062] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0063] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0064] Figure 1 This is a schematic diagram of the composition of the pathological multi-information integrated diagnostic device based on multi-agent collaboration in an embodiment of the present invention;

[0065] Figure 2This is a schematic diagram of the composition of the comprehensive auxiliary module in the pathological multi-information integrated diagnostic device based on multi-agent collaboration in an embodiment of the present invention. Detailed Implementation

[0066] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0067] Example 1:

[0068] This embodiment provides a pathological multi-information integrated diagnostic device based on multi-agent collaboration, such as... Figure 1 As shown, it includes:

[0069] The information aggregation module is used to acquire the updated information corresponding to each specified information channel, and allocate each updated information to the corresponding intelligent agent for information storage according to the information attributes corresponding to each updated information.

[0070] The intelligent agent management module is used to generate an information graph corresponding to each intelligent agent within a specified period, and modify the information graph according to the management content uploaded by the doctor, so as to build a multi-intelligent agent platform for each doctor.

[0071] The intelligent query module is used to search for relevant diagnostic information of the patient in the multi-agent platform when the doctor submits patient information;

[0072] The comprehensive auxiliary module is used to perform a simulated diagnosis on the patient information using the relevant diagnostic information, obtain several physical conditions of the patient, and determine and display the diagnostic validity corresponding to each physical condition.

[0073] In this example, the specified information channels include: hospital network data, big data, medical knowledge websites, and doctors' own uploads, etc.

[0074] In this example, the updated information refers to the first appearance of the updated information in the specified information channel;

[0075] In this example, the intelligent agents include: In this solution, the system gathers various types of information and connects with multiple intelligent agents, including diagnostic planning intelligent agents, arbitration intelligent agents, critical intelligent agents, pathological slide analysis intelligent agents, in-hospital information aggregation intelligent agents, information search intelligent agents, knowledge base search intelligent agents, etc. Each intelligent agent is responsible for different types of work to assist in diagnosis;

[0076] In this example, the information graph represents the graph structure of relationships and connections between information already stored in an agent;

[0077] In this example, the managed content refers to the content uploaded by the doctor to correct the infographic.

[0078] In this example, the multi-agent platform consists of several agents;

[0079] In this example, the specified period can be 7 days;

[0080] In this example, the relevant diagnostic information represents information related to the diagnosis and treatment of the disease in relation to the patient's current condition;

[0081] In this example, simulated diagnosis refers to the process of diagnosing a patient through simulation in a virtual space;

[0082] In this example, diagnostic validity refers to the reliability of the physical condition obtained through simulated diagnosis.

[0083] The working principle and beneficial effects of the above technical solution are as follows: By automatically acquiring updated information through multiple designated information channels and allocating it to corresponding intelligent agents based on information attributes, it solves the problem of traditionally scattered information, reduces the time physicians spend manually integrating information, and achieves classified information storage. Subsequent retrieval allows for quick location of the corresponding intelligent agent. Simultaneously, encrypted storage of intelligent agents ensures data security, version management facilitates information traceability, and intelligent agent information graphs are generated periodically to intuitively present information relationships. Furthermore, physicians can modify the graphs according to management content and build dedicated multi-agent platforms, ensuring the information graphs align with physicians' professional directions and diagnostic habits. The platform allows for customization of intelligent agents. The platform can display and push data rules, and can flexibly adjust the update cycle. When doctors use the platform, they can analyze query requirements, use multi-dimensional weighted matching to find relevant diagnostic information, and then classify, sort and deduplicate the results to reduce interference from irrelevant information. It can quickly provide doctors with valuable diagnostic references such as patient information and similar cases. Finally, using relevant diagnostic information, the platform can obtain several physical conditions of the patient through multi-model fusion algorithms, avoiding the one-sidedness of diagnosis by a single model. At the same time, it can determine the validity of the diagnosis by combining the sufficiency of diagnostic evidence, the reliability of information and the confidence of the model, and display the results visually. This helps doctors judge the reliability of the results, improve diagnostic accuracy and reduce the risk of missed diagnosis and misdiagnosis.

[0084] Example 2:

[0085] Based on Example 1, the information aggregation module of the pathological multi-information integrated diagnostic device based on multi-agent collaboration includes:

[0086] The channel review unit is used to obtain the channel characteristics and information sources corresponding to each specified information channel, construct the channel factors corresponding to each specified information channel, and set the corresponding trust level for each specified information channel based on the doctor's evaluation results of each channel factor.

[0087] The information filtering unit is used to construct corresponding channel generation information based on the real-time data corresponding to each specified channel, and to verify the information of each channel generation information using the corresponding trust level and the storage information corresponding to each intelligent agent, so as to obtain the information fusion features between each channel generation information and different intelligent agents.

[0088] An information update unit is used to filter channel-generated information whose information fusion features meet the filtering criteria, and to perform information compensation on the channel-generated information in combination with the trust level of the specified information channel in which each channel-generated information is located, so as to obtain the updated information corresponding to each specified channel.

[0089] A storage unit is allocated for determining several pre-matching agents corresponding to each update information based on several information fusion features corresponding to each update information, and for matching and verifying the information attributes corresponding to each update information with each pre-matching agent to determine and store the agent corresponding to each update information.

[0090] In this example, channel characteristics refer to the unique features of a specified information channel, which are used to distinguish different specified information channels;

[0091] In this example, the information source refers to the source from which the information is provided through the specified information channel;

[0092] In this example, channel factors refer to the interfering and supporting factors that define the channel when updating information;

[0093] In this example, real-time data refers to data generated in real time within the specified channel;

[0094] In this example, information verification refers to the process of determining whether the information generated by the channel conflicts with the intelligent agent, and the authenticity of the information generated by the channel;

[0095] In this example, the pre-matching agent represents the agent that matches the updated information.

[0096] The working principle and beneficial effects of the above technical solution are as follows: By acquiring the channel characteristics and information sources of specified information channels to construct channel factors, and combining this with doctor evaluations to set trust levels, authoritative and reliable information channels can be screened out. This reduces the probability of low-quality and unreliable information entering the system from the source, laying the foundation for the accuracy of subsequent diagnostic information. Then, channel-generated information is constructed based on real-time channel data. This information is verified by combining channel trust levels with intelligent agent storage information to obtain information fusion characteristics. This effectively identifies the authenticity and relevance of channel-generated information, avoids interference from invalid and conflicting information, and ensures a high degree of matching between the information entering the system and the intelligent agent storage information. Further screening of channel-generated information that meets the standards and information compensation based on channel trust levels yields updated information. This not only retains high-quality information but also improves the completeness and accuracy of updated information by compensating for missing or insufficient information, providing more comprehensive information support for diagnosis. Finally, pre-matched intelligent agents are determined based on information fusion characteristics, and the final stored intelligent agent is determined through information attribute matching verification. This achieves precise matching and storage of updated information and intelligent agents, avoiding incorrect or random information storage. During subsequent retrieval, the corresponding intelligent agent can be quickly located, improving information retrieval efficiency.

[0097] Example 3:

[0098] Based on Example 2, the process of information compensation for the channel-generated information in the pathological multi-information integrated diagnostic device based on multi-agent collaboration includes:

[0099] Historical update information corresponding to each specified channel is obtained, an information structure tree corresponding to the specified channel is constructed, branch identification is performed on each information structure tree to obtain several information subtrees corresponding to each specified information channel, and the hierarchical relationship between different information subtrees is determined.

[0100] The channel-generated information is mapped to the corresponding information structure tree to obtain several related information subtrees corresponding to the channel-generated information. The information influence of each channel-generated information on the information structure tree is determined by combining the corresponding hierarchical relationship.

[0101] Information generated by target channels with an impact value higher than a specified impact value is filtered out, and the scope of the impact of each target channel's generated information on the corresponding information structure tree is determined by combining the trust level corresponding to each specified information channel.

[0102] Based on the screening criteria, each target channel generated information is converted into several information conditions within the scope of the information's influence, and the information semantics corresponding to each information condition are determined.

[0103] Semantic compensation is performed on invalid information conditions with missing semantic information to generate alternative information conditions for each invalid information condition;

[0104] Replace the corresponding invalid information condition with the alternative information condition that has the least impact on the structure of the information structure tree.

[0105] In this example, historical update information represents information that has been updated through a designated channel;

[0106] In this example, the information structure tree represents the result of sorting through the historical update information of a specified channel;

[0107] In this example, branch identification represents the result of dividing the branches in the information structure tree;

[0108] In this example, the hierarchical relationship represents the hierarchical logical relationship between different information sub-items;

[0109] In this example, the information impact represents the degree of influence that channel-generated information has on existing information;

[0110] In this example, the scope of information influence represents the impact of information generated by the channel on the information structure tree.

[0111] In this example, alternative information conditions represent information conditions generated after compensating invalid information conditions in different dimensions.

[0112] The working principle and beneficial effects of the above technical solution are as follows: To effectively and accurately compensate for channel-generated information, an information structure tree is first constructed and information subtrees and hierarchical relationships are identified. This systematically organizes the historical information logic of the specified channels, providing a structured framework for the mapping and analysis of channel-generated information, avoiding disorder in information analysis, and making information relationships clearer. Then, channel-generated information is mapped to the information structure tree and the information impact is determined. This allows for accurate assessment of the degree to which channel-generated information affects the historical information system, facilitating the selection of target channel-generated information that has a significant impact on diagnosis, reducing interference from worthless information. Finally, the impact range of target channel-generated information is determined by combining channel trust levels, and information can be defined based on channel reliability. By defining the boundaries of action, subsequent information processing is ensured to focus on a credible and critical scope, enhancing the relevance and reliability of information processing. Furthermore, information generated from target channels is transformed into information conditions and semantics are defined, making the information more standardized and easier to analyze. This facilitates subsequent processing of invalid information conditions lacking semantics, providing a clear direction for information compensation. Finally, semantic compensation is performed on invalid information conditions, generating alternative conditions to supplement missing semantic parts of the information, improve information content, and avoid affecting diagnosis due to incomplete information. Simultaneously, selecting alternative conditions with the least impact on the information tree structure maximizes the stability of the historical information structure, ensuring the coherence and reliability of the information system, and providing complete and structurally stable information support for subsequent diagnosis.

[0113] Example 4:

[0114] Based on Example 1, the pathological multi-information integrated diagnostic device based on multi-agent collaboration, wherein the agent management module includes:

[0115] The logic management unit is used to acquire the storage information corresponding to each of the intelligent agents within a specified period, determine the information coverage area and information coverage range of the storage information in combination with the intelligent agent attributes corresponding to each of the intelligent agents, and deduce the information logic relationship between several storage information corresponding to each of the intelligent agents.

[0116] The dynamic tracking unit is used to acquire the auxiliary operation data corresponding to each of the intelligent agents within a specified period, construct the extraction information dynamic and update information dynamic of each of the intelligent agents within the specified period, and identify the first dynamic logic corresponding to the extraction information dynamic and the second dynamic logic corresponding to the update information dynamic in the information logic relationship.

[0117] The graph construction unit is used to optimize the first dynamic logic and the second dynamic logic in the information logic relationship, construct the information structure corresponding to the intelligent agent according to the optimization result, and map each of the stored information into the information structure to obtain the information graph corresponding to each intelligent agent.

[0118] The manual optimization unit is used to identify several management items contained in the management content uploaded by the doctor and the management purpose corresponding to each management item, find the corresponding management agent according to the management item, optimize the management agent to the management purpose, and generate the doctor's multi-agent platform.

[0119] In this example, the agent attribute represents the type of information stored in the agent;

[0120] In this example, the information coverage domain represents the domain to which the stored information in the intelligent agent belongs;

[0121] In this example, information coverage refers to the extent to which stored information in an agent is covered within its information coverage area;

[0122] In this example, the information logical relationship represents the relationship between different stored information;

[0123] In this example, the auxiliary operation data represents the data generated by the relevant agents when the multi-agent platform assists the doctor.

[0124] In this example, the dynamics of extracting information dynamics refer to the dynamics generated when extracting information from the agent, while the dynamics of updating information dynamics refer to the dynamics generated when storing updated information into the agent.

[0125] In this example, the first dynamic logic represents the logic embodied in the dynamic extraction of information, and the second dynamic logic represents the logic embodied in the dynamic updating of information.

[0126] In this example, the information structure represents the information structure presented by the extraction and updating of information in the agent.

[0127] The working principle and beneficial effects of the above technical solution are as follows: By acquiring the stored information of the intelligent agent, combining it with the agent's attributes to determine the information coverage area and scope, and deriving the logical relationships between the stored information, the boundaries and connections of the intelligent agent's information can be clearly sorted out, avoiding information chaos and providing an orderly logical foundation for subsequent information graph construction. This allows physicians to quickly understand the organizational logic of the intelligent agent's information, then acquire the intelligent agent's assisted operation data, construct the dynamic extraction and updating of information, identify the corresponding dynamic logic in the information logical relationships, and monitor the flow and changes of the intelligent agent's information in real time. This enables timely detection of abnormal dynamics in information extraction or updating, ensuring the timeliness of the intelligent agent's information. To improve accuracy and reliability, the system further optimizes dynamic logic and constructs an intelligent agent information structure. It maps stored information to generate an information graph, presenting the information relationships between intelligent agents in a visual format. This intuitively displays information hierarchy and logic, allowing physicians to quickly locate the required information and improve information retrieval efficiency. Simultaneously, dynamic logic optimization makes the graph more aligned with actual information flow needs. Finally, it identifies management items and objectives within the managed content, finds the corresponding management intelligent agents, and optimizes them to meet the objectives, generating a physician-specific multi-agent platform. This platform accurately matches physicians' professional directions and diagnostic habits, making its functions more tailored to physician needs, improving ease of use, and ultimately increasing diagnostic efficiency.

[0128] Example 5:

[0129] Based on Example 1, the intelligent query module of the pathological multi-information integrated diagnostic device based on multi-agent collaboration includes:

[0130] An identity determination unit is used to identify the patient identity corresponding to the patient information submitted by the doctor in the multi-agent platform, search for several pieces of intelligent information related to the patient identity in each of the intelligent agents, and determine the source of the intelligent agent corresponding to each piece of intelligent information.

[0131] The semantic analysis unit is used to convert the corresponding intelligent information into corresponding medical semantics according to the intelligent agent function corresponding to each intelligent agent source, generate medical description text of the patient identity, and perform semantic optimization on the medical description text in the multi-agent platform to obtain several auxiliary diagnostic semantics of the patient identity.

[0132] The information integration unit is used to utilize the auxiliary diagnostic semantic recognition to identify the missing semantics corresponding to the intelligent information, perform multi-dimensional semantic expansion on the intelligent information to obtain semantically complete and effective intelligent agent information, construct relevant diagnosis and treatment information of the patient's identity based on the effective intelligent agent information, and display it.

[0133] In this example, intelligent information refers to patient-related information stored in the intelligent agent;

[0134] In this example, the source of the intelligent agent refers to the intelligent agent in which the intelligent information resides;

[0135] In this example, medical semantic representation uses semantics to present intelligent information.

[0136] In this example, auxiliary diagnostic semantics represent the relevant semantics when performing auxiliary diagnosis;

[0137] In this example, multidimensional semantic extension refers to the process of semantically extending intelligent information in different dimensions;

[0138] In this example, the relevant medical information refers to the medical information related to the patient.

[0139] The working principle and beneficial effects of the above technical solution are as follows: To screen accurate and effective diagnostic and treatment information, the system first accurately identifies the patient's identity and locates relevant intelligent information from various intelligent agents. Simultaneously, it clarifies the source of the information from the intelligent agent, avoiding information mismatch caused by confusion regarding patient identity and allowing physicians to clearly understand the source of the information to assess its reliability. This provides a precise and traceable information foundation for subsequent diagnosis. Then, combining the functions of the intelligent agents, the intelligent information is converted into medical semantics, generating medical descriptive text and optimizing auxiliary diagnostic semantics. This eliminates comprehension barriers caused by differences in information expression, unifying the information from different intelligent agents into standardized medical language. This facilitates physicians' rapid understanding of the core information and improves information interpretation efficiency. Finally, by identifying missing semantics and performing multi-dimensional semantic expansion, the intelligent information is completed to form complete and effective intelligent agent information, avoiding diagnostic impact due to incomplete information. Simultaneously, relevant diagnostic and treatment information is constructed and displayed, presenting physicians with comprehensive and systematic patient diagnostic and treatment data, reducing the workload of physicians manually integrating information and helping them efficiently obtain the complete information needed for diagnosis.

[0140] Example 6:

[0141] Based on Example 5, the pathological multi-information integrated diagnostic device based on multi-agent collaboration performs multi-dimensional semantic expansion on the intelligent information to obtain semantically complete and effective intelligent agent information, including:

[0142] Based on the intelligent information, several preliminary diagnostic and treatment records for the patient's identity are constructed;

[0143] Based on the missing semantics, the information defects between different preliminary diagnostic information are determined, and at the same time, the information logical relationship between different preliminary diagnostic information is constructed according to the intelligent agent function corresponding to each intelligent agent source.

[0144] In a multi-agent platform, a diagnostic knowledge base for the doctor is constructed, and alternative information corresponding to each information defect is searched in the diagnostic knowledge base according to the information logical relationship.

[0145] Each of the aforementioned alternative information is used to perform alternative optimization on the information defect, and the target optimized semantics with complete semantics are selected based on the optimized semantics corresponding to each optimization result.

[0146] Effective agent information is constructed based on the optimization process corresponding to the target optimization semantics.

[0147] In this example, preliminary diagnostic information represents the result of converting intelligent information into diagnostic information;

[0148] In this example, agent function refers to the functions that an agent can perform; different agents can perform different functions.

[0149] In this example, the diagnostic knowledge base represents diagnostic and treatment-related knowledge obtained by organizing the relevant knowledge in the multi-agent platform;

[0150] In this example, the substitute information refers to information derived from the diagnostic knowledge base used to optimize information deficiencies.

[0151] The working principle and beneficial effects of the above technical solution are as follows: Preliminary diagnostic information is constructed using intelligent information, providing a basic framework for subsequent semantic expansion. This ensures that the expansion process revolves around the patient's core diagnostic data, avoiding deviation from key information directions. Then, missing semantics are identified to clarify information deficiencies. Simultaneously, information logic relationships are constructed based on intelligent agent functions, accurately locating information deficiencies while ensuring that semantic expansion follows reasonable medical information association logic, guaranteeing the relevance and rationality of the expanded content. Furthermore, a doctor's diagnostic knowledge base is built to find substitute information based on information logic relationships. High-quality information that aligns with doctors' diagnostic habits and professional experience can be invoked, ensuring that substitute information meets actual clinical needs and improving the reliability and applicability of information supplementation. Then, substitute information is used to optimize information deficiencies, and semantically complete target optimization semantics is selected. This effectively fills the semantic gaps in intelligent information, generating complete and standardized diagnostic information fragments, avoiding interference from incomplete information in diagnostic judgment. Finally, effective intelligent agent information is constructed based on the optimization process of target semantics, preserving the logical trajectory of semantic expansion, making the final information both complete and traceable. This facilitates doctors' understanding of information sources and supplementary logic, while providing high-quality information support with complete semantics and clear logic for subsequent diagnosis.

[0152] Example 7:

[0153] Based on Embodiment 1, the pathological multi-information integrated diagnostic device based on multi-agent collaboration further includes:

[0154] The platform display module is used to display the extracted information in the multi-agent platform.

[0155] The working principle and beneficial effects of the above technical solution are as follows: By setting up a platform display module, relevant knowledge is presented to doctors, which facilitates their effective diagnosis and treatment.

[0156] Example 8:

[0157] Based on Example 1, the pathological multi-information integrated diagnostic device based on multi-agent collaboration, such as Figure 2 As shown, the integrated auxiliary module includes:

[0158] The feature generation unit is used to obtain the patient information uploaded by the doctor, combine it with the corresponding relevant diagnostic information to construct several health status features for the patient, and search for the health standard corresponding to each health status feature in the multi-agent platform.

[0159] The integrated simulation unit is used to evaluate the health status characteristics using the health standards, obtain several non-healthy information and several healthy information of the patient, and construct a virtual health model of the patient based on the non-healthy information and the healthy information;

[0160] The simulation diagnostic unit is used to map the virtual health model to each specified scenario, control the virtual health model to perform several specified situations in each specified scenario, obtain several physical conditions of the patient, and determine the health value corresponding to each physical condition according to the health standard.

[0161] The diagnostic integration unit is used to use the current patient information to make disease judgments for each of the health values, determine the probability of each suspected disease, determine the diagnostic validity of each physical condition, generate a health diagnosis probability report, and display it.

[0162] In this example, health status features represent characteristics about the patient's health condition;

[0163] In this example, the health standard refers to the standard corresponding to a health status characteristic when it exhibits a healthy characteristic;

[0164] In this example, the virtual health model represents a model of the patient's health status constructed in a virtual space;

[0165] In this example, the specified scenarios include: daily scenarios, exercise scenarios, sleep scenarios, temperature scenarios at various stages, and humidity scenarios at various stages;

[0166] In this example, the specified scenarios include: performing several exercise scenarios with varying levels of exercise intensity from low to high;

[0167] In this example, the health diagnosis probability report represents a report on the diagnostic validity corresponding to each of the patient's physical conditions.

[0168] The working principle and beneficial effects of the above technical solution are as follows: It avoids evaluation bias caused by ambiguous standards, ensuring the validity and reliability of diagnostic results. First, it constructs health status characteristics using patient information and related diagnostic information to avoid the bias of a single information source. Then, it builds a virtual health model to intuitively display the correlation of health status among various body systems, helping doctors quickly grasp the overall health of the patient and significantly reducing the cost of information comprehension. Furthermore, it controls the model to simulate the patient performing different actions in different scenarios, obtaining the patient's physical condition and related health values. Finally, it generates a health diagnosis probability report to provide doctors with a tiered and reliable diagnostic reference. Results with high validity can be used as the core diagnostic basis, while results with low validity can be used as supplementary references, avoiding misleading diagnoses due to unreliable data.

[0169] Example 9:

[0170] Based on Example 8, the pathological multi-information integrated diagnostic device based on multi-agent collaboration further includes:

[0171] An auxiliary tracing unit is used to filter several target physical conditions in the health diagnosis probability report whose diagnostic validity is higher than the specified validity.

[0172] For each of the target physical conditions, the source of the target-related diagnostic information corresponding to the target physical condition is determined by the intelligent agent.

[0173] Search for the relevant diagnostic information of the target in the information source agent, generate auxiliary reference information and display it.

[0174] In this example, the validity rate is specified as 80%.

[0175] In this example, derivation and tracing refers to the process of tracing the source of the target's physical condition. The derivation and tracing process makes the generation path of diagnostic information completely transparent: doctors can clearly know which agent's data analysis a certain physical condition conclusion comes from, which makes it easier to verify whether the agent's information collection logic and analysis algorithm comply with medical standards, thus ensuring the reliability of the diagnostic results from the root and providing a clear basis for possible subsequent diagnostic review.

[0176] The working principle and beneficial effects of the above technical solution are as follows: First, target physical conditions with diagnostic validity higher than the prescribed validity are screened out to avoid the time cost for doctors to sift through massive amounts of diagnostic data for low-validity information. Then, the target physical conditions are traced to ensure that the traceability results are strongly correlated with the core diagnostic needs, reducing interference from irrelevant information, helping doctors to more accurately judge the condition based on high-validity information, reducing the risk of diagnostic deviation caused by low-quality information, and further displaying the relevant information to doctors, providing them with technical reference.

[0177] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A pathological multi-information comprehensive diagnosis device based on multi-agent cooperation, characterized in that, The application relates to a medical information management system, which comprises the following parts: an information collection module, which is used for collecting updated information corresponding to each specified information channel, and distributing each updated information to a corresponding intelligent agent for information storage according to the information attribute of each updated information; an intelligent agent management module, which is used for generating an information graph corresponding to each intelligent agent in a specified period, and modifying the information graph according to the management content uploaded by a doctor, so as to build a multi-intelligent agent platform for each doctor; an intelligent query module, which is used for searching relevant diagnosis information of patient information in the multi-intelligent agent platform when the doctor submits the patient information; a comprehensive auxiliary module, which is used for simulating diagnosis of the patient information by using the relevant diagnosis information, obtaining several body conditions of the patient, and determining and displaying the diagnosis validity of each body condition.

2. The pathological multi-information comprehensive diagnosis apparatus based on multi-agent cooperation according to claim 1, characterized in that, The information collection module comprises the following parts: a channel auditing unit, which is used for collecting channel characteristics and information sources corresponding to each specified information channel, building a channel factor corresponding to each specified information channel, and setting a corresponding trust degree for each specified information channel according to the evaluation result of each channel factor by the doctor; an information screening unit, which is used for building corresponding channel generated information according to real-time data corresponding to each specified channel, verifying each channel generated information by using the corresponding trust degree and the storage information corresponding to each intelligent agent, and obtaining information fusion characteristics between each channel generated information and different intelligent agents; an information updating unit, which is used for screening channel generated information meeting a screening standard according to the information fusion characteristics, compensating the channel generated information according to the trust degree corresponding to the specified information channel where each channel generated information is located, and obtaining updated information corresponding to each specified channel; a distribution and storage unit, which is used for determining a plurality of pre-matching intelligent agents corresponding to each updated information according to a plurality of information fusion characteristics corresponding to each updated information, respectively matching and verifying the information attribute of each updated information and each pre-matching intelligent agent, determining the intelligent agent corresponding to each updated information, and storing the intelligent agent.

3. The pathological multi-information comprehensive diagnosis apparatus based on multi-agent cooperation according to claim 2, characterized in that, The process of compensating the channel generated information comprises the following steps: collecting historical updated information corresponding to each specified channel, building an information structure tree corresponding to the specified channel, respectively identifying each information structure tree, obtaining a plurality of information sub-trees corresponding to each specified information channel, and determining the hierarchical relationship between different information sub-trees; mapping the channel generated information into the corresponding information structure tree, obtaining a plurality of relevant information sub-trees corresponding to the channel generated information, determining the information influence amount of each channel generated information on the information structure tree according to the hierarchical relationship; screening target channel generated information with an information influence amount higher than a specified influence amount, and determining the information influence range of each target channel generated information on the corresponding information structure tree according to the trust degree corresponding to each specified information channel. According to the screening standard, each target channel generated information is converted into several information conditions within the information influence range, and the information semantics corresponding to each information condition is determined; The invalid information conditions with missing information semantics are compensated for semantics, and the alternative information conditions of each invalid information condition are generated; The alternative information condition with the minimum structural influence on the information structure tree is used to replace the corresponding invalid information condition.

4. The pathological multi-information comprehensive diagnosis apparatus based on multi-agent cooperation according to claim 1, characterized in that, The intelligent agent management module comprises: A logic management unit is configured to acquire the storage information corresponding to each intelligent agent in a specified period, determine the information coverage field and information coverage range of the corresponding storage information in combination with the intelligent agent attribute corresponding to each intelligent agent, and deduce the information logic relationship between several storage information corresponding to each intelligent agent; A dynamic tracking unit is configured to acquire the auxiliary operation data corresponding to each intelligent agent in a specified period, construct the extraction information dynamics and update information dynamics of each intelligent agent in the specified period, and identify the first dynamic logic corresponding to the extraction information dynamics and the second dynamic logic corresponding to the update information dynamics in the information logic relationship; A graph construction unit is configured to optimize the first dynamic logic and the second dynamic logic in the information logic relationship, construct the information structure corresponding to the intelligent agent according to the optimization result, and map each storage information in the information structure to obtain the information graph corresponding to each intelligent agent; A manual optimization unit is configured to identify several management items contained in the management content uploaded by the doctor and the management purpose corresponding to each management item, find the corresponding management intelligent agent according to the management item, and optimize the management intelligent agent to the management purpose to generate the multi-intelligent agent platform of the doctor.

5. The pathological multi-information comprehensive diagnosis apparatus based on multi-agent cooperation according to claim 1, characterized in that, The intelligent query module comprises: An identity determination unit is configured to identify the patient identity corresponding to the patient information in the multi-intelligent agent platform after the patient information submitted by the doctor, find several intelligent information related to the patient identity in each intelligent agent, and determine the intelligent agent source corresponding to each intelligent information; A semantic analysis unit is configured to convert the corresponding intelligent information into the corresponding medical semantics according to the intelligent agent function corresponding to each intelligent agent source, generate the medical description text of the patient identity, and perform semantic optimization on the medical description text in the multi-intelligent agent platform to obtain several auxiliary diagnosis semantics of the patient identity; An information integration unit is configured to identify the missing semantics corresponding to the intelligent information by using the auxiliary diagnosis semantics, perform multi-dimensional semantic expansion on the intelligent information to obtain the effective intelligent agent information with complete semantics, construct the related diagnosis and treatment information of the patient identity according to the effective intelligent agent information, and display the related diagnosis and treatment information.

6. The pathological multi-information comprehensive diagnosis apparatus based on multi-agent cooperation according to claim 5, wherein, The multi-dimensional semantic expansion on the intelligent information to obtain the effective intelligent agent information with complete semantics comprises: Constructing several preliminary diagnosis and treatment information of the patient identity according to the intelligent information; According to the missing semantics, information defects between different preliminary diagnosis information are determined, and according to the agent function corresponding to each agent source, information logic relationships between different preliminary diagnosis information are constructed; A doctor's diagnosis knowledge base is constructed in a multi-agent platform, and according to the information logic relationship, substitute information corresponding to each information defect is found in the diagnosis knowledge base; Each information defect is replaced and optimized by using each substitute information, and target optimization semantics with complete semantics are selected according to the optimization semantics corresponding to each optimization result; According to the optimization process corresponding to the target optimization semantics, effective agent information is constructed.

7. The pathological multi-information comprehensive diagnosis apparatus based on multi-agent cooperation according to claim 1, characterized in that, Further comprising: A platform display module for displaying the extracted information in the multi-agent platform.

8. The pathological multi-information comprehensive diagnosis apparatus based on multi-agent cooperation according to claim 1, characterized in that, The comprehensive auxiliary module comprises: A feature generation unit for obtaining the patient information uploaded by the doctor, combining the corresponding related diagnosis information to construct a plurality of health state features of the corresponding patient, and finding the health standard corresponding to each health state feature in the multi-agent platform; A comprehensive simulation unit for health evaluation of the health state features by using the health standard, obtaining a plurality of unhealthy information and a plurality of healthy information of the patient, and constructing a virtual health model of the patient according to the unhealthy information and the healthy information; A simulation diagnosis unit for mapping the virtual health model into each specified scene, controlling the virtual health model to execute a plurality of specified situations in each specified scene, obtaining a plurality of body conditions of the patient, and determining the health value corresponding to each body condition according to the health standard; A diagnosis integration unit for disease judgment of each health value by using the patient information, determination of the suspected probability corresponding to each suspected disease, determination of the diagnosis effective degree corresponding to each body condition, generation of a health diagnosis probability report, and display.

9. The pathological multi-information comprehensive diagnosis apparatus based on multi-agent cooperation according to claim 8, characterized in that, Further comprising: An auxiliary tracing unit for screening a plurality of target body conditions with a diagnosis effective degree higher than a specified effective degree in the health diagnosis probability report; Each target body condition is deduced and traced to determine the information source agent of the target related diagnosis information corresponding to the target body condition; In the information source agent, the associated information of the target related diagnosis information is found, and auxiliary reference information is generated and displayed.

Images