A cloud computing-based exhibition information synchronization and real-time update management method and platform

Abstract

The application discloses a kind of based on cloud computing's exhibition information synchronization and real-time update management method and platform, comprising the following steps: real-time collection multi-source heterogeneous exhibition information by edge computing node;Collected heterogeneous data is transmitted to cloud computing center, carries out format standardization, and carries out semantic extraction to unstructured text, generates uniform exhibition information data pool;Periodic synchronization and instant synchronization based on event-driven are carried out;When detecting the multi-version update of same information, call the historical operation record of blockchain storage, in combination with preset priority rule and time stamp check, generate final effective version;According to the display capacity and interactive authority of target terminal type, generate the information push package of adaptation.The present application is used to solve the technical problems of multi-source heterogeneous data integration difficulty in traditional exhibition information management, real-time update response deficiency, unstructured text semantic processing ability weakness, multi-source update conflict resolution untrustworthy and multi-terminal information push adaptability poor.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of cloud computing application, and particularly relates to a conference and exhibition information synchronization and real-time update management method and platform based on cloud computing. BACKGROUND

[0002] In traditional conference and exhibition activities, information management faces multiple technical challenges, mainly reflected in the following aspects:

[0003] Firstly, multi-source heterogeneous data integration is difficult. The data sources on the conference and exhibition site are complex and diverse, covering exhibitor terminals, organizer management systems, and Internet of Things sensors, and many other types of equipment. The transmission protocols (such as RESTful API, SOAP, MQTT, etc.) and data formats (structured fields, unstructured text) used by different data sources differ significantly, resulting in a prominent problem of "structural fragmentation" of data. For example, the "exhibition number" of the exhibitor terminal and the "exhibition unique identifier" field of the organizer system are not uniformly named, and the original perception data of the Internet of Things sensor and the structured data of the business system cannot be directly integrated, forming an information island.

[0004] Secondly, the real-time requirement is high but the response is insufficient. In conference and exhibition activities, dynamic information such as exhibition state changes, preferential activity effectiveness, environmental abnormalities (such as temperature and humidity exceeding the standard, human flow density being too high, etc.) needs to be captured and synchronized in a timely manner. However, the traditional centralized data collection method relies on long-distance communication transmission, which is prone to data loss or update lag due to network delay or interruption, and cannot meet the real-time information needs.

[0005] Thirdly, the semantic processing capability of unstructured text is weak. Unstructured text such as exhibitor promotional copy and product description contains a large amount of conference and exhibition field terminology (such as "limited-time discount" and "smart terminal"). However, general language models have limited understanding of the conference and exhibition scene, often resulting in ambiguous identification of field terminology, difficult disambiguation of polysemous words (such as "smart terminal" may refer to consumer electronics or industrial equipment), and loss of focus in extracting key elements (such as "core selling point" and "application scenario"), making it difficult to integrate discrete text information with structured data and form a unified information data pool.

[0006] In addition, multi-source update conflicts are difficult to resolve. When the exhibitor terminal, the organizer system, or the Internet of Things device simultaneously modifies the same information (such as the exhibition contact information), the traditional method lacks a reliable historical operation record tracing mechanism, and has no clear priority rules (such as the authority level of the organizer and the exhibitor updates) and timestamp verification mechanism, which easily leads to the "data fighting" phenomenon when multiple information is synchronized, and cannot generate a unique valid version.

[0007] Finally, the multi-terminal information pushing is not adaptive enough. The traditional "one-size-fits-all" pushing mode does not consider the display capabilities (such as screen resolution, supported media format) and interaction permissions (such as editing permission, browsing permission) of different terminals (exhibitor APP, audience applet, on-site electronic screen, cooperative media interface), resulting in problems such as mismatch of information granularity (such as audience receiving sensitive business information), incompatible display format (such as electronic screen unable to adapt to small font text), imbalance between real-time performance and resource consumption (such as repeated transmission of regular information), and prominent security risks (such as sensitive data leakage).

[0008] In summary, the traditional exhibition information management has significant defects in data integration, real-time response, semantic processing, conflict resolution, and multi-terminal adaptation, and an integrated solution based on cloud computing is urgently needed to improve the efficiency and reliability of information synchronization and real-time updating. SUMMARY

[0009] In order to overcome the deficiencies of the prior art, the purpose of the present application is to provide a cloud computing-based exhibition information synchronization and real-time updating management method and platform, which solves the technical problems of difficulty in integrating multi-source heterogeneous data, insufficient real-time updating response, weak semantic processing capability of unstructured text, untrusted multi-source update conflict resolution, and poor adaptability of multi-terminal information pushing in traditional exhibition information management.

[0010] To solve the above problems, the technical solution adopted by the present application is as follows:

[0011] A cloud computing-based exhibition information synchronization and real-time updating management method, comprising the following steps:

[0012] Through the deployment of edge computing nodes in the exhibition site, real-time collection of multi-source heterogeneous exhibition information from exhibitor terminals, organizer management systems, and Internet of Things sensors is performed;

[0013] The collected heterogeneous data is transmitted to the cloud computing center, standardized by pre-configured metadata mapping rules, and the semantic extraction of unstructured text is performed using an exhibition language processing model to generate a unified exhibition information data pool;

[0014] A double-trigger mechanism is set up in the cloud computing center for periodic synchronization based on time threshold and instant synchronization based on event-driven;

[0015] When multiple versions of the same information are detected, the historical operation records of blockchain notarization are called, combined with the preset priority rules and timestamp verification, to generate the final effective version;

[0016] According to the display capabilities and interaction permissions of the target terminal type, the API gateway of the cloud computing center dynamically generates an adaptive information pushing package.

[0017] Preferably, when collecting multi-source heterogeneous exhibition information in real time, it includes:

[0018] The HTTP / HTTPS interface adapter connects to the exhibitor's terminal's RESTful API interface and the organizer's management system's SOAP protocol interface to achieve standardized access to business system data.

[0019] The raw sensor data from the booth environment sensors and binocular vision people flow counter are analyzed using an IoT protocol parser.

[0020] The local caching module uses the LRU algorithm to temporarily store the collected heterogeneous data. When communication with the cloud computing center is interrupted, the collection records are retained and automatically re-uploaded to the cloud after communication is restored.

[0021] Preferably, when performing format standardization and semantic extraction, the following are included:

[0022] The metadata mapping rules are constructed based on an exhibition industry ontology library containing multi-dimensional ontology concepts in the metadata mapping unit.

[0023] For exhibitors' RESTful API interface data, organizers' SOAP protocol data, and IoT sensor perception data, establish field-level mapping relationships to map non-uniform fields in heterogeneous data to the standard data model defined in the exhibition industry ontology library.

[0024] The exhibition language processing model built in the NLP semantic extraction unit is used to identify domain terms and disambiguate semantics in unstructured text, extract key semantic elements and associate them with corresponding nodes in the exhibition industry ontology library, and finally generate a unified exhibition information data pool containing structured data and semantic tags.

[0025] Preferably, when mapping to a standard data model, it includes:

[0026] By formally defining the core concepts, attributes, and relationships between concepts in the domain through multi-dimensional ontology concepts, a unified semantic interpretation framework is provided for the standard data model, and semantic constraints are imposed on the definition, value range, and association rules of each data element in the standard data model.

[0027] In this context, the abstract semantics of multi-dimensional ontology concepts need to be made operational through the specific fields, relationships, and constraints of standard data models.

[0028] Preferably, when extracting key semantic elements and associating them with corresponding nodes, the process includes:

[0029] A domain dictionary is constructed based on multi-dimensional ontology concepts to identify domain terms in unstructured text;

[0030] By combining the hierarchical relationships and attribute constraints of multi-dimensional ontology concepts, the ambiguity of polysemous words in unstructured text can be eliminated.

[0031] Based on the semantic boundary of the multi-dimensional ontology concept, key semantic elements strongly related to the exhibition scenario are extracted and mapped to the corresponding structured fields in the standard data model, realizing the semantic and structured transformation of unstructured text.

[0032] Preferably, when constructing a language processing model for an exhibition, the following are included:

[0033] Based on the BERT pre-trained model, we utilize its bidirectional Transformer architecture to capture the strong semantics of long text context and understand the complex semantic relationships in unstructured text.

[0034] By customizing and fine-tuning the BERT pre-trained model using a corpus of the exhibition industry domain, the model learns the specific vocabulary, sentence structure, and domain logic of the exhibition scenario, thereby generating the exhibition language processing model.

[0035] Preferably, when setting up a dual-trigger mechanism in a cloud computing center, it includes:

[0036] The synchronization cycle can be flexibly configured using Cron expressions through time triggers to periodically synchronize routine exhibition information that has not undergone significant changes;

[0037] The event listener component of the rule engine in the event trigger monitors the status changes of the exhibition information data pool in real time. When any of the following events is detected, such as exhibit status change event, promotional activity activation event, and environmental related event, synchronization is triggered immediately.

[0038] The dual-trigger mechanism works in collaboration with the synchronization control module, cloud processing engine module, and conflict management module to perform periodic synchronization and event-driven synchronization.

[0039] Preferably, when generating the final effective version, the following are included:

[0040] Through the blockchain evidence storage submodule of the conflict management module, all historical operation records of the same information can be retrieved based on the consortium blockchain architecture. Each record contains the operation subject, operation timestamp, and hash value of the modified content.

[0041] Based on the priority rule library built into the version arbitration submodule, the operation subjects of multiple version updates are classified and determined.

[0042] For multiple version updates with the same priority, extract the operation timestamp of each version and select the version with the latest timestamp as the candidate version;

[0043] If there are update versions with different priorities, the version corresponding to the subject with the highest priority will be selected, and the changes in the version with the latest timestamp will be synchronized to the unified exhibition information data pool.

[0044] Preferably, when generating an adapted push notification package, the following steps are included:

[0045] The API gateway dynamically parses the display capabilities and interaction permissions of the target terminal by receiving the terminal's request header information and pre-stored terminal type tags.

[0046] Based on terminal type, deployment location, and authorization level, the rule engine is invoked to execute differentiated pruning strategies;

[0047] By calling display templates that match the terminal type through the template engine, the content of the unified format exhibition information data pool is converted into a presentation format adapted to the terminal.

[0048] Based on the triggering mechanism of the synchronization control module, event-driven information is pushed in real time using a WebSocket long connection; regular information that is periodically synchronized is pushed using an HTTP short connection with cache verification.

[0049] Push packets containing sensitive information are encrypted with AES and a terminal-specific dynamic token is attached; watermarks are added to push data from partner media interfaces.

[0050] A cloud-based platform for synchronizing and updating exhibition information in real time, comprising:

[0051] Multi-source data acquisition module: Deployed on edge computing nodes at the exhibition site, including HTTP / HTTPS interface adapters, IoT protocol parsers and local caching modules, used to collect multi-source heterogeneous exhibition information in real time;

[0052] Cloud processing engine module: running in the cloud computing center, it includes a metadata mapping unit with a built-in exhibition industry ontology library and an NLP semantic extraction unit with a built-in exhibition language processing model and an exhibition industry domain corpus, which is used to convert heterogeneous data into a unified exhibition information data pool.

[0053] Synchronization control module: integrates time triggers and event triggers to drive periodic synchronization and event-driven synchronization of exhibition information;

[0054] Conflict Management Module: Includes a blockchain evidence storage submodule and a version arbitration submodule, used to resolve multi-source update conflicts;

[0055] Multi-device adaptation push module: includes API gateway and template engine, used to generate differentiated synchronous information push packages according to terminal type.

[0056] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0057] I. Efficiency and Continuity of Multi-Source Heterogeneous Data Acquisition

[0058] By integrating HTTP / HTTPS interface adapters (interfacing with RESTful APIs and SOAP protocols), IoT protocol parsers (supporting sensor protocols such as MQTT and LoRaWAN), and local caching modules (using the LRU algorithm to retain data from the most recent 24 hours) into edge computing nodes deployed at the exhibition site, real-time acquisition of multi-source heterogeneous data and data continuity assurance during communication interruptions were achieved. This design reduces the computing pressure on the cloud and avoids data loss caused by transmission delays or interruptions in traditional centralized acquisition, providing a complete and real-time data source foundation for subsequent processing.

[0059] II. Structured and Semantic Integration of Heterogeneous Data

[0060] Based on an ontology library for the exhibition industry, metadata mapping rules are constructed to map heterogeneous data (such as "booth number" and "unique booth identifier") from exhibitors' terminals, organizers' systems, and IoT sensors to a unified standard data model, solving the "fragmentation" problem in traditional methods. Simultaneously, a customized, fine-tuned BERT pre-trained model (combined with an exhibition domain corpus) performs domain terminology recognition, semantic disambiguation, and key element extraction (such as "core selling points" and "limited-time discounts") on unstructured text (such as exhibitor promotional copy), and associates it with corresponding nodes in the ontology library, achieving "semantic and structural transformation" of unstructured text. The synergistic effect of these two methods forms a unified exhibition information data pool containing structured data and semantic tags, significantly improving the efficiency of multi-source data fusion.

[0061] III. Balancing the Timeliness and Coverage of Information Synchronization

[0062] By employing a dual-trigger mechanism (periodic synchronization + event-driven synchronization), the system balances the coverage of routine information with the real-time nature of emergency events: periodic synchronization (configurable from 10 seconds to 5 minutes) ensures the regular updating of routine information that has not undergone significant changes (such as basic booth information); event-driven synchronization (real-time monitoring of events such as changes in exhibit status, activation of promotional activities, and environmental anomalies) ensures the immediate delivery of emergency information (such as warnings of crowd density exceeding thresholds). This design avoids the problems of traditional single-trigger mechanisms that are either "incomplete in coverage or wasteful of resources," thus improving the overall efficiency of information synchronization.

[0063] IV. Reliable Resolution of Multi-Source Update Conflicts

[0064] The credibility of version traceability is achieved through a blockchain-based evidence storage submodule (which stores the operation entity, timestamp, and hash value of the modified content based on a consortium blockchain). Combined with the priority rules of the version arbitration submodule (organizer system > exhibitor terminal > IoT device) and timestamp verification, the update authority levels of different entities are clearly defined, and conflicts of the same priority are handled. This mechanism effectively solves the "data discrepancy" problem in traditional methods, ensuring that a unique and valid version is generated when multiple devices are synchronized, thus guaranteeing the consistency and accuracy of information.

[0065] V. Adaptability and Security of Multi-Terminal Scenario-Based Push Notifications

[0066] By dynamically parsing terminal display capabilities and interaction permissions through an API gateway, executing differentiated tailoring strategies in conjunction with a rules engine, and adapting display formats using a template engine, precise push notifications are achieved for exhibitors to view business information, visitors to view offers, electronic screens to view guidance, and media to view the structure. Simultaneously, WebSocket long connections ensure the timeliness of urgent information, HTTP short connections with cache verification reduce the amount of regular information transmitted, and AES encryption (for sensitive data), dynamic tokens (terminal-specific), and watermarking (for media data) enhance security. This design solves the information mismatch, display incompatibility, and security risks of the traditional "one-size-fits-all" push mode, significantly improving the information access experience and data security for multi-terminal users.

[0067] In summary, this invention, through the deep integration of technologies such as cloud computing, edge computing, NLP semantic processing, and blockchain, constructs a comprehensive exhibition information management system covering the entire process of data collection, processing, synchronization, conflict resolution, and push, providing efficient, real-time, reliable, and adaptable information synchronization and update services for exhibition activities.

[0068] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. Attached Figure Description

[0069] Figure 1 This is a flowchart illustrating the steps of the exhibition information synchronization and real-time update management method according to an embodiment of the present invention;

[0070] Figure 2 This is a flowchart illustrating the process of generating the final effective version according to an embodiment of the present invention;

[0071] Figure 3 This is a flowchart illustrating the process of generating an adapted information push package according to an embodiment of the present invention.

[0072] Figure 4 This is a schematic diagram of the exhibition information synchronization and real-time update management platform according to an embodiment of the present invention.

[0073] The following are the icon numbers: 200, Exhibition Information Synchronization and Real-time Update Management Platform; 201, Multi-Source Data Acquisition Module; 202, Cloud Processing Engine Module; 203, Synchronization Control Module; 204, Conflict Management Module; 205, Multi-Terminal Adaptation Push Module. Detailed Implementation

[0074] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0075] At the same time, it should be understood that, for ease of description, the dimensions of the various parts shown in the accompanying drawings are not drawn according to actual scale.

[0076] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the scope of this application and its application or use.

[0077] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.

[0078] Example 1, see Figure 1 The present invention provides a step-by-step diagram of a method for synchronizing and updating exhibition information in real time. Figure 1 The method for synchronizing and updating exhibition information based on cloud computing, as shown, includes the following steps:

[0079] S101. Multi-source heterogeneous data acquisition: Through edge computing nodes deployed at the exhibition site, multi-source heterogeneous exhibition information is collected in real time from exhibitor terminals, organizer management system, and IoT sensors (including booth environment sensors and people flow counters). The multi-source heterogeneous exhibition information includes: basic booth information (including exhibitor qualifications and exhibit parameters), real-time dynamic information (including exhibit demonstration status and changes in promotional activities), and environmental related information (including booth temperature and humidity and people flow density).

[0080] S102, Cloud Platform Standardization Processing: The collected heterogeneous data is transmitted to the cloud computing center, and the format is standardized through pre-configured metadata mapping rules. The exhibition language processing model is used to extract semantics from unstructured text (such as promotional copy updated in real time by exhibitors) to generate a unified exhibition information data pool.

[0081] S103. Dynamic Synchronization Trigger: Set up a dual trigger mechanism in the cloud computing center to perform periodic synchronization based on time thresholds (the period can be configured from 10 seconds to 5 minutes) and instant synchronization based on event-driven events (events include changes in exhibit status, the activation of promotional activities, and the exceeding of the threshold for crowd density).

[0082] S104. Conflict Detection and Resolution: When multiple versions of the same information are detected (such as exhibitors' terminals and organizers' systems modifying booth contact information at the same time), the historical operation records stored on the blockchain are called, and the final valid version is generated by combining the preset priority rules (organizer's system > exhibitors' terminals > IoT devices) and timestamp verification.

[0083] S105, Multi-terminal Differentiated Push: Based on the display capabilities and interaction permissions of the target terminal type (including exhibitor APP, visitor mini-program, on-site electronic screen, and partner media interface), the system dynamically generates adapted information push packages through the API gateway of the cloud computing center. Sensitive business information is filtered on the visitor terminal, and crowd guidance information is prioritized on the on-site electronic screen, achieving contextualized adaptation of synchronous information.

[0084] Background Description: In traditional exhibition and convention activities, information management faces challenges such as the difficulty of integrating multi-source heterogeneous data and high real-time requirements. On the one hand, the data sources at exhibition sites are complex and diverse, with significant differences in transmission protocols and data formats used by different data sources, leading to prominent information silos. On the other hand, exhibitions and conventions have high requirements for the timeliness of information updates; dynamic information such as changes in booth status, the activation of promotional activities, and environmental anomalies need to be captured and synchronized in a timely manner. Traditional centralized data collection methods are prone to data loss or delayed updates due to transmission delays or communication interruptions. Based on this:

[0085] In one possible embodiment, during step S101 above, the real-time acquisition of multi-source heterogeneous exhibition information includes:

[0086] The HTTP / HTTPS interface adapter connects to the exhibitor's terminal's RESTful API interface and the organizer's management system's SOAP protocol interface to achieve standardized access to business system data.

[0087] The raw sensor data from the booth environment sensors (including temperature and humidity sensors, CO2 concentration sensors, and light intensity sensors) and the binocular vision people flow counter are analyzed using an IoT protocol parser.

[0088] The local caching module uses the LRU algorithm to temporarily store the collected heterogeneous data. When communication with the cloud computing center is interrupted, the collection records of the most recent 24 hours can be retained and automatically re-uploaded to the cloud after communication is restored.

[0089] The edge computing nodes are equipped with HTTP / HTTPS interface adapters, IoT protocol parsers, and local caching modules; the IoT protocol parser supports MQTT, LoRaWAN, and ZigBee protocols.

[0090] In this embodiment of the invention, it is necessary to further explain that the embodiment utilizes edge computing nodes deployed at the exhibition site, integrating HTTP / HTTPS interface adapters (for interfacing with business systems), IoT protocol parsers (for adapting sensor protocols), and local caching modules (using the LRU algorithm to temporarily store data), to achieve real-time collection of multi-source heterogeneous exhibition information. This design can both process heterogeneous data of different protocols and formats locally, reducing the computing pressure on the cloud, and ensure data continuity during communication interruptions through local caching (retaining records from the most recent 24 hours), providing a complete and real-time data source foundation for subsequent cloud platform standardization processing and dynamic synchronization.

[0091] Background Description: In traditional exhibition information management, the inconsistent formats and semantic ambiguity of multi-source heterogeneous data are the core pain points restricting information integration and utilization. On the one hand, raw data from exhibitor terminals (such as RESTful API data), organizer systems (such as SOAP protocol data), and IoT sensors (such as MQTT sensing data) are fragmented due to different field naming (such as "booth number" and "unique booth identifier"), protocol standards (such as RESTful and SOAP), and data types (structured fields and unstructured text), making direct integration impossible. On the other hand, unstructured text (such as exhibitor promotional materials and product descriptions) contains a large number of domain-specific terms (such as "limited-time discount" and "smart terminal"), but due to the insufficient understanding of exhibition scenarios by general language models, ambiguity often occurs (such as "smart terminal" referring to consumer electronics or industrial equipment) or omission of key elements (such as "core selling points" not being accurately extracted), resulting in semantic ambiguity. Based on this:

[0092] In one possible embodiment, the format standardization and semantic extraction performed in step S102 above include:

[0093] Metadata mapping rules are constructed based on the exhibition industry ontology library in the metadata mapping unit. The exhibition industry ontology library contains multi-dimensional ontology concepts such as exhibit classification system, exhibitor qualification level, and event type tags.

[0094] For exhibitors' RESTful API interface data, organizers' SOAP protocol data, and IoT sensor perception data, field-level mapping relationships are established to map non-uniform fields such as "booth number," "exhibition model," and "promotional activity description" in heterogeneous data to standard data models such as "unique booth identifier," "exhibition technical parameters," and "marketing activity information" defined in the exhibition industry ontology library.

[0095] The NLP semantic extraction unit uses the exhibition language processing model to identify domain terms and disambiguate unstructured text (such as exhibitor promotional copy), extracts key semantic elements such as "core selling points" and "applicable scenarios" and associates them with corresponding nodes in the exhibition industry ontology library, and finally generates a unified exhibition information data pool containing structured data and semantic tags.

[0096] In this embodiment of the invention, it is necessary to further explain that the structural inconsistency problem is solved by using metadata mapping rules (building field-level mapping relationships based on the exhibition industry ontology) to map heterogeneous data to standard data models such as "unique booth identifier" and "exhibit technical parameters". At the same time, a customized and finely tuned BERT pre-trained model (combined with an exhibition domain corpus) is used for semantic extraction to identify domain terms, dissolve ambiguities, and extract key elements such as "core selling points" and "applicable scenarios", ultimately generating a unified exhibition information data pool. This process realizes the transformation of heterogeneous data from "fragmented" to "structured" and from "fuzzy semantics" to "explicit semantics", providing a standardized data source foundation for subsequent dynamic synchronization, conflict resolution, and multi-terminal adaptation and push.

[0097] Background Description: In traditional exhibition information processing, semantic extraction from unstructured text (such as exhibitor promotional materials and product descriptions) faces the following challenges: ambiguity in domain terminology recognition, difficulty in resolving polysemous words, and loss of focus in key element extraction. Furthermore, if the extracted semantic elements are not associated with a standard data model, structured storage cannot be achieved, making it difficult to integrate discrete textual information with IoT sensor data and structured data from business systems to form a unified information data pool. Based on this:

[0098] In one possible embodiment, extracting key semantic elements and associating them with corresponding nodes includes:

[0099] Based on the multi-dimensional ontology concepts in the exhibition industry ontology library (including exhibit classification system, event type tags, exhibitor qualification level), a domain dictionary is constructed to identify domain terms (such as "limited-time discount", "screen size" and "industry certification") in unstructured text.

[0100] By combining the hierarchical relationship of multi-dimensional ontological concepts (such as the classification hierarchy of "consumer electronics-smart terminals-smartphones") and attribute constraints (such as "AAA-level exhibitors must meet the requirement of annual revenue ≥ 100 million"), the ambiguity of polysemous words in unstructured text can be eliminated (such as clarifying "smart terminal" as "consumer electronics-smart terminal" rather than "smart terminal" in other fields).

[0101] Based on the semantic boundaries of multi-dimensional ontology concepts (such as "promotional activities" corresponding to elements such as "limited-time discounts" and "first 100 customers", and "smartphones" corresponding to attributes such as "screen size" and "battery life"), key semantic elements strongly related to the exhibition scenario (including core selling points, applicable scenarios, and discount levels) are extracted and mapped to the corresponding structured fields in the standard data model (such as the "marketing activity information - activity type" field and the "exhibition product technical parameters - core selling points" field), realizing the semantic and structured transformation of unstructured text;

[0102] The exhibition language processing model is based on the BERT (Bidirectional Encoder Representation Transformer) pre-trained model. It leverages the bidirectional Transformer architecture's strong ability to capture the contextual semantics of long texts to understand the complex semantic relationships in unstructured texts (such as exhibitor promotional copy and product descriptions). It is also customized and fine-tuned through a corpus of exhibition industry domains (containing historical exhibitor promotional copy, product technical descriptions, event announcements, audience feedback, and other text data). This allows the model to learn specialized vocabulary in exhibition scenarios (such as "limited-time discount", "product battery life", "industry certification"), sentence structures (such as the promotional expression "order before 18:00 today and enjoy 50% off"), and domain logic (such as the hierarchical relationship between "smart terminal" and "consumer electronics"), thus completing the adaptation from a general language model to a model specific to the exhibition domain.

[0103] In this embodiment of the invention, it is necessary to further explain that the embodiment constructs an ontology library for the exhibition industry (including exhibit classification systems, activity type tags, and other multi-dimensional concepts) to provide the model with a "domain dictionary" and "semantic filtering rules." On the one hand, the hierarchical relationships and attribute constraints of the ontology library are used to resolve the ambiguity of polysemous words and clarify the exhibition scenario to which the terms refer. On the other hand, based on the semantic boundaries of the ontology library, the model is guided to accurately extract key elements such as "core selling points" and "applicable scenarios" and map them to the structured fields of the standard data model. This process realizes the transformation of unstructured text from "fuzzy semantics" to "scenario-based and structured semantics," providing an operable semantic foundation for subsequent information synchronization, conflict detection, and multi-terminal adaptation and push.

[0104] In one possible implementation, multi-dimensional ontology concepts (such as exhibit classification systems, exhibitor qualification levels, and event type tags) formally define core concepts, attributes, and relationships between concepts within the domain (e.g., the hierarchical relationship in the "exhibitor classification system" where "consumer electronics" includes "smart terminals," and "smart terminals" includes "smartphones"; the "AAA level" in the "exhibitor qualification level" must meet attribute constraints such as "annual revenue ≥ 100 million" and "complete industry certifications"). This provides a unified semantic interpretation framework for the standard data model. The definition, value range, and association rules of each data element in the standard data model (such as the "exhibitor classification" field and the "exhibitor qualification level" field) are directly derived from the semantic constraints of the multi-dimensional ontology concepts. For example, the value of the "exhibitor classification" field in the standard data model must strictly correspond to the hierarchical path of the "exhibitor classification system" in the ontology library (e.g., "consumer electronics - smart terminals - smartphones"), rather than arbitrarily defined classification descriptions.

[0105] The abstract semantics of multi-dimensional ontology concepts need to be operationalized through the specific fields, relationships, and constraints of standard data models. For example, concepts such as "Promotional Activity - Limited-Time Discount" and "Interactive Activity - Audience Experience" defined in the "Activity Type Label" of the exhibition industry ontology will be transformed into enumerated values ​​of the "Activity Type" field in the standard data model (e.g., value range: ["Promotional Activity - Limited-Time Discount", "Interactive Activity - Audience Experience"]). The hierarchical relationship of "AAA Level → AA Level → A Level" defined in the "Exhibitor Qualification Level" of the ontology will be transformed into the sorting rules of the "Exhibitor Qualification Level" field in the standard data model (e.g., numerical priority identifiers: AAA Level = 3, AA Level = 2, A Level = 1). Through this structured mapping, ontology concepts are transformed from abstract domain knowledge into storable and computable concrete data models.

[0106] In this embodiment of the invention, it is necessary to further explain that the multi-dimensional ontology concept solves the "semantic inconsistency" problem of heterogeneous data (for example, unifying the interpretation of "3C products" on exhibitor terminals and "electronic devices" in organizer systems as "consumer electronics-smart terminals"), while the standard data model solves the "structural inconsistency" problem of heterogeneous data (for example, unifying the mapping of "qualification level A" on exhibitor terminals and "level 1 certification" in organizer systems as "AA level" field values). The synergistic effect of the two enables heterogeneous data from different sources (exhibitor terminals, organizer systems, IoT sensors) and different formats (structured fields, unstructured text) to be ultimately converted into a standard data model that follows the same semantic and structural rules, providing a unified data source foundation for subsequent information synchronization, conflict detection, and multi-terminal push.

[0107] The extraction of key semantic elements (such as "core selling points," "applicable scenarios," and "limited-time discounts") is not aimless text extraction, but rather a domain-oriented extraction based on semantic constraints of multi-dimensional ontological concepts. For example:

[0108] The "activity type tags" (such as "promotional activities" and "interactive activities") in the ontology library provide semantic boundaries for the exhibition language processing model, guiding the model to identify key elements belonging to "promotional activities" such as "limited-time discount" and "first 100" from exhibitors' promotional copy, rather than irrelevant information (such as "booth location").

[0109] The "exhibit classification system" in the ontology (such as "consumer electronics - smart terminals - smartphones") provides the model with the core concepts of the exhibit domain, guiding the model to extract "core selling points" elements belonging to "smartphones" such as "screen size" and "battery life" from unstructured descriptions, rather than attributes of other types of exhibits (such as "material" of "furniture exhibits").

[0110] In short, the multi-dimensional ontology concept provides a "domain dictionary" and "semantic filtering rules" for the extraction of key semantic elements, ensuring that the extracted elements are strongly related to the exhibition scenario and conform to domain specifications;

[0111] Multi-dimensional ontological concepts (such as "activity type labels" and "exhibit classification systems") are abstract domain knowledge frameworks, while key semantic elements are the instantiation of these frameworks in specific texts. For example:

[0112] The "Activity Type Tag" in the ontology library defines the concept of "Promotional Activity - Limited Time Discount". When exhibitors' copy includes the phrase "Enjoy 50% off when you place an order before 18:00 today", the "Limited Time Discount" and "50% off" extracted by the NLP model are specific instances of this ontology concept.

[0113] The "Exhibit Classification System" in the ontology defines the attributes of "Smart Terminal" (such as "Interaction Method" and "Operating System"). When the exhibit description includes "Supports Touch Interaction" or "Runs on Android System", the extracted "Touch Interaction" and "Android System" are the instantiated elements of the "Smart Terminal" attribute.

[0114] Through this instantiation, the multidimensional ontology concept is transformed from an abstract semantic framework into a concrete information unit that can be processed by a computer;

[0115] Key semantic elements need to be structured and stored using a standard data model in order to form a unified exhibition information data pool together with other structured data (such as temperature and humidity data from IoT sensors and booth numbers from exhibitors' terminals). For example:

[0116] The extracted key elements such as "limited-time discount" and "50% off" will be mapped to subfields under the "marketing campaign information" field in the standard data model (e.g., "campaign type" = "promotional campaign - limited-time discount", "discount level" = "50% off").

[0117] The extracted key elements, such as "touch interaction" and "Android system", will be mapped to subfields under the "Exhibition Technical Parameters" field in the standard data model (e.g., "Interaction Method" = "Touch", "Operating System" = "Android").

[0118] Standard data models organize discrete key semantic elements into searchable and synchronizeable structured data by defining field hierarchies, data types, and relationships (such as the relationship between the "marketing campaign information" and "exhibit category" fields), ultimately supporting subsequent information synchronization and real-time updates.

[0119] Background Description: In the traditional process of synchronizing exhibition information, a single synchronization trigger mechanism is insufficient to balance the comprehensiveness and real-time nature of information updates. Therefore:

[0120] In one possible embodiment, when setting up the dual-trigger mechanism in the cloud computing center in step S103 above, the following is included:

[0121] Periodic synchronization mechanism: The synchronization period can be flexibly configured using Cron expressions through time triggers (covering a configurable range of 10 seconds to 5 minutes), such as setting it to every 30 seconds, every 2 minutes, etc., to periodically synchronize routine exhibition information that has not undergone significant changes (such as basic booth information and stable exhibit parameters).

[0122] Event-driven synchronization mechanism: The event listener component of the rule engine in the event trigger monitors the state changes of the exhibition information data pool in real time, and immediately triggers synchronization when the following events are detected:

[0123] Exhibit status change events (such as exhibit demonstration status changing from "inactive" to "demonstration" or "paused");

[0124] Promotional activities take effect upon the event (such as the event period reaching the set start / end time, or the event details being modified).

[0125] Environmentally related events (such as crowd density exceeding a preset threshold, such as 20 people / m²) 2 (or any of the following exceeding the set values: booth temperature and humidity, CO2 concentration, and light intensity).

[0126] The dual-trigger mechanism works in collaboration with the synchronization control module 203, cloud processing engine module 202, and conflict management module 204 to ensure the coverage of information updates through periodic synchronization and the real-time processing of emergency changes through event-driven synchronization, thus ensuring the timeliness and accuracy of the exhibition information across multiple terminals.

[0127] In this embodiment of the invention, it is necessary to further explain that the implementation ensures the coverage of routine information (such as basic booth information) that has not changed significantly through periodic synchronization (the period can be configured from 10 seconds to 5 minutes); and handles the immediacy of emergency changes through event-driven synchronization (real-time monitoring of events such as changes in exhibit status, activation of promotional activities, and environmental anomalies), ensuring that sudden information quickly reaches multiple terminals. The two work together to ensure both information coverage efficiency and solve the problem of real-time response to emergency events, providing a foundation for timely and accurate information updates for subsequent conflict detection and multi-terminal push.

[0128] Background Description: Traditional exhibition information management struggles to resolve multi-source update conflicts primarily due to: First, unreliable version tracing. When exhibitor terminals, organizer systems, or IoT devices simultaneously modify the same information (such as booth contact details), a lack of reliable historical operation records makes it difficult to confirm the subject, time, and content of each version's modifications. Second, the absence of priority rules. Updates from different sources (such as organizers and exhibitors) lack a clear hierarchy of authority, making it impossible to objectively determine which version should be adopted. Third, the absence of timestamp verification. Updates submitted simultaneously from the same source (such as two exhibitor terminals) may incorrectly retain older versions while overwriting the latest modifications due to the lack of a timestamp comparison mechanism. These issues easily lead to "data discrepancies" during multi-terminal information synchronization. Based on this:

[0129] In one possible embodiment, see [reference] Figure 2 The flowchart for generating the final valid version, in step S104 above, includes:

[0130] S41, Blockchain Evidence Storage and Traceability: Through the blockchain evidence storage sub-module of the conflict management module 204, all historical operation records of the same information are retrieved based on the consortium blockchain architecture. Each record contains the operation subject (exhibitor terminal / organizer system / IoT device), operation timestamp, and hash value of the modified content.

[0131] S42. Priority grading verification: Based on the priority rule library built into the version arbitration submodule, the operation subjects of multiple version updates are graded and judged, with the priority order being: organizer system > exhibitor terminal > IoT device.

[0132] S43. Timestamp-assisted arbitration: For multiple version updates with the same priority (such as two exhibitors' terminals modifying the same information at the same time), extract the operation timestamps of each version and select the version with the latest timestamp as the candidate version.

[0133] S44. Valid version generation: If there are updated versions with different priorities, the version corresponding to the highest priority subject is selected directly; if there are multiple versions of the highest priority subject, the version with the latest timestamp is selected; finally, the modified content of this version is synchronized to the unified exhibition information data pool to complete the conflict resolution.

[0134] In this embodiment of the invention, it is necessary to further explain that the untrustworthy version issue is resolved through blockchain evidence storage and traceability (based on the storage of complete operation logs on a consortium blockchain), ensuring that each update record is verifiable; an authority level is established through priority-based verification (organizer system > exhibitor terminal > IoT device), clarifying the update priority of different entities; and timestamp-assisted arbitration (selecting the latest timestamp version) handles conflicts of the same priority, ensuring that the latest modification is adopted first. The synergistic effect of these three elements ultimately generates a unique and valid version, resolving the conflict problem of multi-source updates and providing a unified and reliable data source foundation for subsequent multi-terminal synchronization.

[0135] Background Description: In the traditional process of pushing exhibition information, the "one-size-fits-all" push mode leads to problems in multi-terminal application scenarios, such as: mismatch of information granularity, incompatible display formats, imbalance between real-time performance and resource consumption, and significant security risks. Based on this:

[0136] In one possible embodiment, see [reference] Figure 3 The flowchart for generating an adapted push notification package includes the following steps in step S105:

[0137] S51. Terminal Type Identification and Capability Analysis: The API gateway receives the terminal's request header information (including the User-Agent field, screen resolution parameters, and supported media formats) and pre-stored terminal type tags (exhibitor APP / visitor mini-program / on-site electronic screen / cooperative media interface), and dynamically analyzes the target terminal's display capabilities (such as maximum supported image resolution, whether it supports animation effects, and text length limits) and interaction permissions (such as exhibitor's editing permissions, visitor's browsing permissions, and media's reprint authorization scope).

[0138] S52. Contextualized Content Tailoring: Based on terminal type, deployment location, and authorization level, the rule engine is invoked to execute differentiated tailoring strategies.

[0139] For exhibitors' apps:

[0140] Core business information retention: Sensitive business information of exhibitors' own booths (such as company contact information, unpublished summaries of letters of intent for cooperation, and unreleased technical parameters of exhibits) will be fully retained. Only business information of non-affiliated exhibitors (such as details of promotional activities of competitors' booths) will be anonymized (displayed as "[Other exhibitor activities]").

[0141] Prioritize business management data: Focus on pushing dynamic information directly related to exhibitor operations, including:

[0142] Real-time booth status data (such as exhibit demonstration status change records, temperature and humidity abnormality alarms, and the current number of visitors displayed by the visitor counter);

[0143] Interaction effectiveness statistics (such as the number of inquiries initiated by visitors through the mini-program, the number of information forms submitted, and the distribution of page view duration for exhibits);

[0144] Instructions from the organizer (such as "Cooperation with media interviews is required from 15:00 to 16:00" or "Booth setup exceeds the time limit and rectification is required").

[0145] Editable fields are enabled: For exhibitors' self-management permissions, editable field modules (such as "description of promotional activities" and "supplementary explanation of exhibit parameters") are embedded in the push package. The field format matches the RESTful API interface requirements of the exhibitor's terminal, and supports direct modification and subsequent transmission back to the cloud computing center to update the data pool.

[0146] Irrelevant information filtering: Filter out public information that is not related to the exhibitor's own business (such as crowd guidance prompts in other exhibition areas, and gift collection rules for visitors), and retain only information with global impact (such as "12:00 closing security check reminder" and "sudden weather warning").

[0147] For the visitor mini-program: filter exhibitor contact information and unpublished business cooperation information, and prioritize retaining the core selling points of exhibits, real-time promotional activities (such as "the first 50 visitors can receive a gift") and crowd guidance information for the current area (such as "the current queue at Gate 3 in Area B is 5 minutes").

[0148] For on-site electronic screens: dynamically adjust information priority based on deployment location (entrance / exhibition area / rest area). Entrance screens highlight the overall exhibition overview and recommended popular booths, while exhibition area screens strengthen the warning of current crowd density (e.g., "Current density 25 people / m²"). 2 (It is recommended to visit during off-peak hours), and the rest area screen displays catering services and event schedules;

[0149] For media partners: adjust the granularity of information according to the authorization level (basic level only provides exhibit name and category; advanced level additionally provides exhibit technical parameters and exhibitor qualification level), and convert it into the structured format required by the media (such as the "title-abstract-content" level of JSON).

[0150] S53. Dynamic Format Adaptation: By calling display templates that match the terminal type through the template engine (such as card-style layout templates for the audience and large-font carousel templates for electronic screens), the content of the unified format exhibition information data pool is converted into a terminal-adapted presentation format.

[0151] Text-based information: Intelligent summarization based on terminal text length limits (e.g., ≤500 characters per screen for mini-programs), retaining "core selling points + key data" (e.g., "Smartwatch: 48-hour battery life, 20% off for a limited time today from 10:00 to 12:00").

[0152] Multimedia information: Based on the media support capabilities of the terminal (e.g., electronic screens support video, while mini programs only support images), match materials of the corresponding format from the cloud media resource library (e.g., transcode 4K video into 720P images);

[0153] S54. Real-time optimized push: Combined with the triggering mechanism of the synchronization control module 203, event-driven information (such as sudden promotional activities and warnings of excessive crowds) is pushed in real time using WebSocket long connections to ensure timely arrival at the terminal; regular information that is periodically synchronized (such as basic booth information) is pushed using HTTP short connections + cache verification (ETag / Last-Modified) to reduce the amount of duplicate data transmission.

[0154] S55. Enhanced Security Processing: Push packets containing sensitive information (such as exhibitors' contact information) are encrypted with AES and a terminal-specific dynamic token (generated based on terminal ID and timestamp) is attached to ensure that only authorized terminals can decrypt and access them; watermarks (such as "Data Source: XX Exhibition Cloud Platform") are added to push data from partner media interfaces to prevent unauthorized reproduction.

[0155] In this embodiment of the invention, it is necessary to further explain that the embodiment clarifies the demand boundaries through terminal type identification and capability analysis, and achieves precise push notifications for "exhibitors viewing business, visitors viewing discounts, electronic screens viewing guidance, and media viewing structure" through scenario-based content tailoring; ensures the display friendliness of information on different terminals through dynamic format adaptation; balances the immediacy of urgent information with the transmission efficiency of regular information through real-time optimized push notifications; and ensures the controllability of access to sensitive data and the traceability of media data sources through enhanced security processing. This process realizes the transformation of exhibition information from "unified push" to "scenario-based and adapted push," providing exhibitors, visitors, management, and partner media with efficient, secure, and user-friendly information synchronization services.

[0156] Example 2, see Figure 4 This invention provides a schematic diagram of the structure of a conference and exhibition information synchronization and real-time update management platform. Figure 4 The exhibition information synchronization and real-time update management platform 200 shown includes: a multi-source data acquisition module 201, a cloud processing engine module 202, a synchronization control module 203, a conflict management module 204, and a multi-terminal adaptation and push module 205.

[0157] Multi-source data acquisition module 201: An edge computing node deployed at the exhibition site, including an HTTP / HTTPS interface adapter, an IoT protocol parser and a local cache module, used to collect multi-source heterogeneous exhibition information in real time;

[0158] Cloud processing engine module 202: running in the cloud computing center, it includes a metadata mapping unit with a built-in exhibition industry ontology library and an NLP semantic extraction unit with a built-in exhibition language processing model and an exhibition industry domain corpus, which is used to convert heterogeneous data into a unified exhibition information data pool.

[0159] Synchronization control module 203: integrates time triggers and event triggers to drive periodic synchronization and event-driven synchronization of exhibition information;

[0160] Conflict Management Module 204: Includes a blockchain evidence storage submodule and a version arbitration submodule, used to resolve multi-source update conflicts;

[0161] Multi-terminal adaptation push module 205: Includes API gateway and template engine, used to generate differentiated synchronous information push packages according to terminal type.

[0162] In one possible embodiment, the HTTP / HTTPS interface adapter is used to interface with the exhibitor's terminal's RESTful API interface and the organizer's management system's SOAP protocol interface to enable standardized access to business system data.

[0163] In one possible embodiment, the IoT protocol parser supports MQTT, LoRaWAN, and ZigBee protocols to parse raw perception data from booth environment sensors (including temperature and humidity sensors, CO2 concentration sensors, and light intensity sensors) and binocular vision people flow counters.

[0164] In one possible embodiment, the local cache module uses the LRU algorithm to temporarily store the collected heterogeneous data, so that the collection records of the most recent 24 hours can be retained when communication with the cloud computing center is interrupted, and automatically re-uploaded to the cloud after communication is restored.

[0165] In one possible embodiment, the exhibition industry ontology library built into the metadata mapping unit is used to construct metadata mapping rules. The exhibition industry ontology library contains multi-dimensional ontology concepts such as exhibit classification system, exhibitor qualification level, and event type tags.

[0166] In one possible embodiment, the metadata mapping unit is used to establish field-level mapping relationships for exhibitor terminal RESTful API interface data, organizer system SOAP protocol data, and IoT sensor perception data, respectively, mapping non-uniform fields in heterogeneous data to the standard data model defined in the exhibition industry ontology library.

[0167] In one possible embodiment, the exhibition language processing model built into the NLP semantic extraction unit is used to perform domain terminology recognition and semantic disambiguation on unstructured text, as well as to extract key semantic elements and associate them with corresponding nodes in the exhibition industry ontology library.

[0168] In one possible embodiment, the exhibition language processing model is based on the BERT pre-trained model, which utilizes its bidirectional Transformer architecture to capture the strong semantics of long text context, solving the problem of understanding complex semantic relationships in unstructured text; and is customized and fine-tuned through a corpus of the exhibition industry domain, enabling the model to learn the special vocabulary, sentence structure and domain logic of the exhibition scenario.

[0169] In one possible implementation, the time trigger uses a Cron expression to flexibly configure the synchronization period for periodically synchronizing routine event information that has not undergone significant changes.

[0170] In one possible embodiment, the event trigger includes a rule engine-based event listening component that monitors the state changes of the exhibition information data pool in real time and immediately triggers synchronization when a specific event is detected.

[0171] In one possible embodiment, the blockchain evidence storage submodule uses a consortium blockchain architecture to retrieve all historical operation records of the same information. Each record contains the operation subject, operation timestamp, and hash value of the modified content.

[0172] In one possible embodiment, the version arbitration submodule has a built-in priority rule base and timestamp verification algorithm for classifying and determining the operation subjects of multiple version updates.

[0173] In one possible embodiment, the API gateway supports RESTful and gRPC protocols, and is used to dynamically parse the display capabilities and interaction permissions of the target terminal by receiving the terminal's request header information and pre-stored terminal type tags.

[0174] In one possible embodiment, the multi-terminal adaptation push module 205 calls the rule engine to execute a differentiated pruning strategy based on the terminal type, deployment location, and authorization level.

[0175] In one possible embodiment, the template engine has built-in display templates for various types of visitors, exhibitors, electronic screens, etc., which are used to call display templates that match the terminal type and convert the content of the unified format exhibition information data pool into a presentation format adapted to the terminal.

[0176] The above embodiments are merely preferred embodiments of the present invention and should not be construed as limiting the scope of protection of the present invention. Any non-substantial changes and substitutions made by those skilled in the art based on the present invention shall fall within the scope of protection claimed by the present invention.

Claims

1. A cloud computing-based method for synchronizing and real-time updating exhibition information, characterized in that, Includes the following steps: By deploying edge computing nodes at the exhibition site, multi-source heterogeneous exhibition information from exhibitor terminals, organizer management systems, and IoT sensors can be collected in real time. The collected heterogeneous data is transmitted to the cloud computing center, where it is standardized in format using pre-configured metadata mapping rules. The unstructured text is then semantically extracted using an exhibition language processing model to generate a unified exhibition information data pool. A dual-trigger mechanism is set up in the cloud computing center to perform periodic synchronization based on time thresholds and instant synchronization based on events. When multiple versions of the same information are detected, the historical operation records stored on the blockchain are called, and the final valid version is generated by combining the preset priority rules and timestamp verification. Based on the display capabilities and interaction permissions of the target terminal type, the appropriate information push package is dynamically generated through the API gateway of the cloud computing center. This includes real-time collection of multi-source heterogeneous exhibition information, including: The HTTP / HTTPS interface adapter connects to the exhibitor's terminal's RESTful API interface and the organizer's management system's SOAP protocol interface to achieve standardized access to business system data. The raw sensor data from the booth environment sensors and binocular vision people flow counter are analyzed using an IoT protocol parser. The local caching module uses the LRU algorithm to temporarily store the collected heterogeneous data. When communication with the cloud computing center is interrupted, the collection records are retained and automatically re-uploaded to the cloud after communication is restored.

2. The method according to claim 1, characterized in that, When performing format standardization and semantic extraction, the following are included: The metadata mapping rules are constructed based on an exhibition industry ontology library containing multi-dimensional ontology concepts in the metadata mapping unit. For exhibitors' RESTful API interface data, organizers' SOAP protocol data, and IoT sensor perception data, establish field-level mapping relationships to map non-uniform fields in heterogeneous data to the standard data model defined in the exhibition industry ontology library. The exhibition language processing model built in the NLP semantic extraction unit is used to identify domain terms and disambiguate semantics in unstructured text, extract key semantic elements and associate them with corresponding nodes in the exhibition industry ontology library, and finally generate a unified exhibition information data pool containing structured data and semantic tags.

3. The method according to claim 2, characterized in that, When mapping to a standard data model, this includes: By formally defining the core concepts, attributes, and relationships between concepts in the domain through multi-dimensional ontology concepts, a unified semantic interpretation framework is provided for the standard data model, and semantic constraints are imposed on the definition, value range, and association rules of each data element in the standard data model. In this context, the abstract semantics of multi-dimensional ontology concepts need to be made operational through specific fields, relationships, and constraints of standard data models.

4. The method according to claim 2 or 3, characterized in that, When extracting key semantic elements and associating them with corresponding nodes, the following steps are included: A domain dictionary is constructed based on multi-dimensional ontology concepts to identify domain terms in unstructured text; By combining the hierarchical relationships and attribute constraints of multi-dimensional ontology concepts, the ambiguity of polysemous words in unstructured text can be eliminated. Based on the semantic boundary of the multi-dimensional ontology concept, key semantic elements strongly related to the exhibition scenario are extracted and mapped to the corresponding structured fields in the standard data model, realizing the semantic and structured transformation of unstructured text.

5. The method according to claim 4, characterized in that, When constructing a language processing model for exhibitions, the following are included: Based on the BERT pre-trained model, we utilize its bidirectional Transformer architecture to capture the strong semantics of long text context and understand the complex semantic relationships in unstructured text. By customizing and fine-tuning the BERT pre-trained model using a corpus of the exhibition industry domain, the model learns the specific vocabulary, sentence structure, and domain logic of the exhibition scenario, thereby generating the exhibition language processing model.

6. The method according to claim 5, characterized in that, When setting up a dual-trigger mechanism in a cloud computing center, the following should be included: The synchronization cycle can be flexibly configured using Cron expressions through time triggers to periodically synchronize routine exhibition information that has not undergone significant changes; The event listener component of the rule engine in the event trigger monitors the status changes of the exhibition information data pool in real time. When any of the following events is detected, such as exhibit status change event, promotional activity activation event, and environmental related event, synchronization is triggered immediately. The dual-trigger mechanism works in collaboration with the synchronization control module, cloud processing engine module, and conflict management module to perform periodic synchronization and event-driven synchronization.

7. The method according to claim 6, characterized in that, When generating the final valid version, the following are included: Through the blockchain evidence storage submodule of the conflict management module, all historical operation records of the same information can be retrieved based on the consortium blockchain architecture. Each record contains the operation subject, operation timestamp, and hash value of the modified content. Based on the priority rule library built into the version arbitration submodule, the operation subjects of multiple version updates are classified and determined. For multiple version updates with the same priority, extract the operation timestamp of each version and select the version with the latest timestamp as the candidate version; If there are update versions with different priorities, the version corresponding to the subject with the highest priority will be selected, and the changes in the version with the latest timestamp will be synchronized to the unified exhibition information data pool.

8. The method according to claim 7, characterized in that, When generating the adapted push notification package, the following are included: The API gateway dynamically parses the display capabilities and interaction permissions of the target terminal by receiving the terminal's request header information and pre-stored terminal type tags. Based on terminal type, deployment location, and authorization level, the rule engine is invoked to execute differentiated pruning strategies; By calling display templates that match the terminal type through the template engine, the content of the unified format exhibition information data pool is converted into a presentation format adapted to the terminal. Based on the triggering mechanism of the synchronization control module, event-driven information is pushed in real time using a WebSocket long connection; regular information that is periodically synchronized is pushed using an HTTP short connection with cache verification. Push packets containing sensitive information are encrypted with AES and a terminal-specific dynamic token is attached; watermarks are added to push data from partner media interfaces.

9. A cloud computing-based platform for synchronizing and real-time updating exhibition information, characterized in that: include: Multi-source data acquisition module: Deployed on edge computing nodes at the exhibition site, including HTTP / HTTPS interface adapters, IoT protocol parsers and local caching modules, used to collect multi-source heterogeneous exhibition information in real time; Cloud processing engine module: running in the cloud computing center, it includes a metadata mapping unit with a built-in exhibition industry ontology library and an NLP semantic extraction unit with a built-in exhibition language processing model and an exhibition industry domain corpus, which is used to convert heterogeneous data into a unified exhibition information data pool. Synchronization control module: integrates time triggers and event triggers to drive periodic synchronization and event-driven synchronization of exhibition information; Conflict Management Module: Includes a blockchain evidence storage submodule and a version arbitration submodule, used to resolve multi-source update conflicts; Multi-terminal adaptation push module: includes API gateway and template engine, used to generate differentiated synchronous information push packages according to terminal type; Among them, the HTTP / HTTPS interface adapter is used to connect to the exhibitor's terminal RESTful API interface and the organizer's management system SOAP protocol interface to enable standardized access to business system data. The IoT protocol parser is used to parse the raw sensing data from booth environment sensors and binocular vision people flow counters; The local caching module uses the LRU algorithm to temporarily store the collected heterogeneous data. This is used to retain the collection records when communication with the cloud computing center is interrupted, and to automatically re-upload them to the cloud after communication is restored.

Images