Enterprise digital diagnosis device, method, equipment and medium

Abstract

The embodiment of the invention provides an enterprise digital diagnosis device and method, equipment and a medium, and belongs to the field of enterprise digital transformation. The device comprises an enterprise data collection and storage module for collecting enterprise data to construct a data set and storing the data set into a memory; the enterprise digital diagnosis agent is used for receiving and jointly embedding structured, semi-structured and unstructured data sets, performing field fine tuning on a general large model based on an industrial digital transformation evaluation standard document, and constructing the general large model for preliminary diagnosis; and the information extraction and diagnosis module is used for extracting feature information in the unstructured data in the data set based on an NLP natural language processing large model and inputting the feature information to a prediction model to predict potential risks. Through application of the large model technology, comprehensive and accurate evaluation of the enterprise digital level is realized, and through real-time diagnosis and optimization, problems can be rapidly identified and targeted suggestions can be provided to assist enterprises in promoting digital transformation.

Description

Technical Field

[0001] The present invention relates to the technical field of enterprise digital transformation, and particularly to an enterprise digital diagnosis device, method, equipment and medium. Background Art

[0002] With the advancement of the global digital wave, enterprise digital transformation has become a key means to enhance competitiveness and adapt to market changes. However, enterprises face many challenges in the process of digital transformation, such as imperfect technical architecture, data island problems, opaque business processes, etc.

[0003] The traditional digital evaluation methods have the following technical problems: 1) Low efficiency: Manual analysis takes a long time and is difficult to meet the needs of enterprises for rapid decision-making.

[0004] 2) Strong subjectivity: The evaluation results are affected by the experience and knowledge level of the evaluators and lack objectivity.

[0005] 3) Limited coverage: Manual analysis is difficult to comprehensively cover all business areas and digital dimensions of an enterprise.

[0006] 4) Lack of dynamics: Traditional methods are difficult to track the progress of enterprise digital transformation in real time and cannot detect problems in time. Summary of the Invention

[0007] The purpose of the embodiments of the present invention is to provide an enterprise digital diagnosis device, method, equipment and medium, which realizes a comprehensive evaluation and diagnosis of enterprise digital transformation through intelligent means, helps enterprises identify problems, formulate improvement strategies, and efficiently promote digital transformation.

[0008] To achieve the above purpose, the embodiments of the present invention provide an enterprise digital diagnosis device, including: An enterprise data collection and storage module, configured to collect enterprise data from the enterprise's business systems and external data sources, construct a data set, and store the data set in a memory; An enterprise digital diagnosis intelligent agent, which is constructed by receiving and jointly embedding structured, semi-structured and unstructured data sets and performing domain fine-tuning on a general large model based on industry digital transformation evaluation standard documents, and is used for the preliminary diagnosis of enterprise digitization; An information extraction and diagnosis module, configured to extract feature information from unstructured data in the data set based on an NLP natural language processing large model, and input the feature information into a prediction model, so that the prediction model predicts potential risks according to the association relationship between the feature information and the structured data in the data set.

[0009] Optionally, the enterprise digital diagnosis intelligent agent includes: The knowledge-driven decision-making core, which includes an industry knowledge graph and a transformation practice library, realizes digital evaluation in four dimensions of technical architecture, business process, data management, and organizational culture through dynamic knowledge retrieval, multi-round dialogue reasoning, and semantic association analysis; The dynamic evolution module continuously optimizes the large model parameters through historical diagnostic feedback data, constructs an industry adaptive learning mechanism to improve cross-domain diagnostic accuracy; The collaborative interaction interface is configured to perform multi-round semantic interaction with enterprise managers, obtain requirements, and visualize the diagnostic process to form a closed-loop diagnostic link.

[0010] Optionally, the process of the enterprise digital diagnosis agent receiving and jointly embedding structured, semi-structured, and unstructured data sets includes: For structured data, through a pre-trained entity alignment model, map the table fields in the E structured data to the diagnostic standard ontology to generate relational vectors with semantic labels; For semi-structured data, construct a tree structure, extract node features, and generate multi-dimensional graph vectors in combination with the multi-head attention mechanism; For unstructured data, adopt a two-stream encoder architecture to extract feature vectors, and perform alignment processing on the extracted feature vectors to eliminate modal differences. Among them, for text streams, perform layout analysis, extract hierarchical features of title-paragraph-table, and generate multi-dimensional semantic vectors. For speech streams, perform ASR transcription and extract prosody features and semantic features; Input the relational vectors, multi-dimensional graph vectors, multi-dimensional semantic vectors, and prosody features and semantic features into a gated fusion network to allocate learning weights and output multi-dimensional joint embedding vectors; According to the industry to which the input data belongs, splice the multi-dimensional joint embedding vectors to form domain-sensitive features.

[0011] Optionally, the process of constructing an enterprise digital diagnosis agent by receiving and jointly embedding structured, semi-structured, and unstructured data sets and fine-tuning the general large model based on the industry digital transformation evaluation standard document includes: Use extraction rules to annotate the corpus of the industry digital transformation evaluation standard document to construct domain corpus; Simulate enterprise digital scenarios based on the knowledge graph and domain corpus to generate dialogue samples with diagnostic labels; Construct a multi-task fine-tuning framework. Among them, for the index classification task, use cross-entropy loss; for the entity annotation task, use the BIOES label system plus a CRF layer; for the inference chain generation task, use the Seq2Seq architecture for training; Freeze the first 24 layers of the general large model, fine-tune the rank decomposition matrix of the last 6 layers, and use the dialogue samples to train the general large model; Among them, an industry-specific prompt is concatenated before the input data to enhance the model's sensitivity to the evaluation criteria.

[0012] Optionally, the information extraction and diagnosis module extracts feature information from the unstructured data in the dataset based on the NLP natural language processing large model, and inputs the feature information into the prediction model, so that the prediction model predicts potential risks according to the association relationship between the feature information and the structured data in the dataset, including: Connect the unstructured features and structured metrics through an ontology relationship to form an inferable association rule, and adjust the association strength coefficient according to the industry to which the enterprise belongs; Construct a dual-time-axis alignment model including a structured axis and an unstructured axis, and predict potential risks by matching the patterns of the structured axis and the unstructured axis; Calculate the attention weights of the unstructured features for each structured metric through the Transformer mechanism and generate a risk heat map; When the association degree between the unstructured feature and the structured metric is detected to exceed the preset threshold, generate a risk chain, perform an intervention simulation on the generated risk chain, verify the effectiveness of the suggestion through causal inference, and output a confidence score; Optimize the prediction result by combining the confidence score and the preliminary diagnosis result.

[0013] Optionally, the enterprise digital diagnosis device further includes: A diagnosis report generation module for generating a diagnosis report, where the diagnosis report includes a digital maturity score, a visual analysis of key bottlenecks, and prioritized transformation suggestions.

[0014] Optionally, the generated diagnosis report includes outline generation, content generation, paragraph division and format adjustment, data visualization and chart citation.

[0015] In a second aspect, an embodiment of the present invention further provides an enterprise digital diagnosis method, including: Collect enterprise data from the enterprise's business system and external data sources, construct a dataset, and store the dataset in a memory; Constructed by receiving and jointly embedding structured, semi-structured, and unstructured datasets and performing domain fine-tuning on a general large model based on an industry digital transformation evaluation standard document for preliminary diagnosis of enterprise digitization; Extract feature information from the unstructured data in the dataset based on the NLP natural language processing large model, and input the feature information into the prediction model, so that the prediction model predicts potential risks according to the association relationship between the feature information and the structured data in the dataset. In a third aspect, an embodiment of the present invention further provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of the enterprise digital diagnosis method described above are implemented.

[0016] In a fourth aspect, an embodiment of the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the enterprise digital diagnosis method described above are implemented.

[0017] Through the above technical solutions, an enterprise digital intelligent diagnosis and analysis function based on a large model is constructed around all aspects of enterprise digital transformation, mainly including an enterprise digital diagnosis intelligent agent, an enterprise data collection and storage module, and an enterprise information extraction and diagnosis module. Through the application of large model technology, a comprehensive and accurate assessment of the enterprise digital level is realized, and the device supports real-time diagnosis and optimization, can quickly identify problems and provide targeted suggestions, helping enterprises efficiently promote digital transformation.

[0018] Other features and advantages of the embodiments of the present invention will be described in detail in the subsequent specific implementation part. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings are used to provide a further understanding of the embodiments of the present invention, and constitute a part of the specification, and are used to explain the embodiments of the present invention together with the following specific implementation manners, but do not constitute a limitation to the embodiments of the present invention. In the drawings: Figure 1 is a schematic structural diagram of an enterprise digital diagnosis device provided by an embodiment of the present invention; Figure 2 is a flowchart of jointly embedding structured, semi-structured, and unstructured data sets provided by an embodiment of the present invention; Figure 3 is a flowchart of domain fine-tuning a general large model provided by an embodiment of the present invention; Figure 4 is an implementation flowchart of an enterprise digital diagnosis method provided by an embodiment of the present invention; Figure 5 is a schematic hardware structure diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] In the following detailed description, various embodiments of the present disclosure will be described more fully. The present disclosure can have various embodiments, and adjustments and changes can be made therein. However, it should be understood that there is no intention to limit the various embodiments of the present disclosure to the specific embodiments disclosed herein, but rather the present disclosure should be understood to cover all adjustments, equivalents, and / or alternative solutions that fall within the spirit and scope of the various embodiments of the present disclosure.

[0021] Hereinafter, the term "comprising" or "may comprise" that can be used in various embodiments of the present disclosure indicates the presence of the disclosed function or operation, and does not limit the addition of one or more functions or operations. In addition, as used in various embodiments of the present disclosure, the terms "comprising", "having" and their cognates are only intended to indicate a specific feature, number, step, operation, or combination of the foregoing items, and should not be construed as first excluding the existence of one or more other features, numbers, steps, operations, or combinations of the foregoing items or the possibility of adding one or more features, numbers, steps, operations, or combinations of the foregoing items.

[0022] In various embodiments of the present disclosure, the expression "or" or "at least one of A or / and B" includes any combination or all combinations of the words listed simultaneously. For example, the expression "A or B" or "at least one of A or / and B" may include A, may include B, or may include both A and B.

[0023] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0024] Refer to Figure 1 As shown, it is a schematic structural diagram of an enterprise digital diagnosis device in a specific embodiment, including: An enterprise data collection and storage module, configured to collect enterprise data from the enterprise's business system and external data sources, construct a data set, and store the data set in a memory; An enterprise digital diagnosis agent, constructed by receiving and jointly embedding structured, semi-structured, and unstructured data sets and performing domain fine-tuning on a general large model based on an industry digital transformation evaluation standard document, for the preliminary diagnosis of enterprise digitization; An information extraction and diagnosis module, configured to extract feature information from unstructured data in the data set based on an NLP natural language processing large model, and input the feature information into a prediction model, so that the prediction model predicts potential risks according to the association relationship between the feature information and the structured data in the data set. Exemplarily, the data collection and storage module: The main function of this module is to collect data from various business systems and external data sources of the enterprise, support multiple data formats and protocols, and ensure high-quality and highly available collected data. Through this module, the enterprise can break data silos and achieve comprehensive integration and efficient utilization of data. The module constructs a unified data lake to implement cleaning and standardized storage of structured data (database tables), semi-structured data (logs / documents), and unstructured data (pictures / videos), and supports automated collection from API interfaces, Internet of Things devices, enterprise ERP / CRM systems, and public data, mainly including data source identification, data collection and preprocessing, and data storage functions. 1) Data source identification: The module first identifies the types, formats, and protocols of data sources to be collected, and based on the identification results, selects appropriate parsers and adapters for subsequent processing. 2) Data collection: Through the parsers and adapters, the module establishes connections with data sources and grabs data according to the predetermined collection strategy. During the collection process, the module performs real-time verification and error handling on the data to ensure the accuracy and integrity of the data. 3) Data preprocessing: For the collected unstructured data, the module performs preprocessing operations such as text tokenization and image recognition. The preprocessed data will be converted into a format suitable for subsequent analysis and processing. 4) Data storage: The collected data will be stored in the specified data storage media, such as databases and file servers. During the storage process, the module encrypts and compresses the data to ensure data security and save storage space.

[0025] In a specific embodiment, the enterprise digital diagnosis intelligent agent includes: A knowledge-driven decision-making core, which includes an industry knowledge graph and a transformation practice library, and through dynamic knowledge retrieval, multi-round dialogue reasoning, and semantic association analysis, realizes digital evaluation in four dimensions: technical architecture, business process, data management, and organizational culture; A dynamic evolution module, which continuously optimizes the large model parameters through historical diagnosis feedback data and constructs an industry adaptive learning mechanism to improve the cross-domain diagnosis accuracy; A collaborative interaction interface, configured to perform multi-round semantic interactions with enterprise managers, obtain requirements, and visualize the diagnosis process to form a closed-loop diagnosis link.

[0026] Exemplarily, the industry adaptive learning mechanism includes feedback data processing, parameter optimization strategy, cross-domain adaptation method, and evaluation mechanism. Among them, feedback data processing includes data cleaning, data normalization processing, and knowledge annotation; parameter optimization includes using the integrated gradient method to locate key parameters and grouping the key parameters into core parameters, domain parameters, and general parameters. Cross-domain adaptation: Construct positive and negative samples, where the positive samples are represented by the patent reports of the same enterprise at different times, and the negative samples are represented by problems in different industries but with similar surface features. Then, based on the positive and negative samples, domain adversarial training is performed to adapt to the domain. Evaluation mechanism: Construct a performance evaluation matrix (using a sliding window to statistically calculate the domain stability index and determine the reduction ratio of the data volume required for fine-tuning in the new industry to determine the cross-domain gain) and keep the performance evaluation matrix within the standard threshold range.

[0027] In some embodiments, referring to Figure 2 as shown, the process of the enterprise digital diagnosis agent receiving and jointly embedding structured, semi-structured, and unstructured data sets includes the following steps: S200: For structured data, through a pre-trained entity alignment model, map the table fields in the structured data to the diagnostic standard ontology to generate relational vectors with semantic labels.

[0028] S201: For semi-structured data, construct a tree structure, extract node features, and generate multi-dimensional graph vectors by combining the multi-head attention mechanism.

[0029] S202: For unstructured data, adopt a two-stream encoder architecture to extract feature vectors and perform alignment processing on the extracted feature vectors to eliminate modality differences. Among them, for the text stream, perform layout analysis, extract hierarchical features of title-paragraph-table, and generate multi-dimensional semantic vectors. For the speech stream, perform ASR transcription and extract prosodic features and semantic features.

[0030] S203: Input the relational vectors, multi-dimensional graph vectors, multi-dimensional semantic vectors, and prosodic features and semantic features into a gated fusion network to allocate learning weights and output multi-dimensional joint embedding vectors.

[0031] S204: According to the industry to which the input data belongs, splice the multi-dimensional joint embedding vectors to form domain-sensitive features.

[0032] Further, referring to Figure 3 as shown, the process of constructing an enterprise digital diagnosis agent by receiving and jointly embedding structured, semi-structured, and unstructured data sets and performing domain fine-tuning on a general large model based on the industry digital transformation evaluation standard document includes the following steps: S300: Perform corpus annotation on the industry digital transformation evaluation standard document using extraction rules to construct domain corpus.

[0033] Exemplarily, based on the "Assessment Criteria for the Digital Transformation of Manufacturing Industry", the defined core dimensions and extraction rules are shown in Table 1.

[0034] Table 1 Core Dimensions and Extraction Rules

[0035] S301: Simulate the enterprise digital scenario based on the knowledge graph and domain corpus to generate dialogue samples with diagnostic labels.

[0036] Exemplarily, the generated dialogue samples are shown in Table 2: Table 2 Summary Table of Dialogue Samples

[0037] S302: Construct a multi-task fine-tuning framework. Among them, for the metric classification task, use the cross-entropy loss; for the entity annotation task, use the BIOES label system plus the CRF layer; for the inference chain generation task, use the Seq2Seq architecture for training.

[0038] S303: Freeze the first 24 layers of the general large model, fine-tune the rank decomposition matrix of the last 6 layers, and use the dialogue samples to train the general large model.

[0039] Among them, splice the industry-specific prompt before the input data to enhance the sensitivity of the model to the assessment criteria.

[0040] Exemplarily, the enterprise digital diagnosis intelligent agent: This module is the core of the entire device, an autonomous decision-making core built based on large models (such as GPT, BERT, etc.), supporting multi-round conversations, dynamic knowledge retrieval, and diagnostic logical reasoning. It has an in-built industry knowledge graph and a digital transformation best practice library, supports semantic data association analysis, and has functions such as large model fine-tuning and adaptation, multi-dimensional evaluation, intelligent analysis, dynamic learning, and interactive diagnosis. The specific functions are as follows: 1) The large model fine-tuning and adaptation function adds a dedicated input channel for enterprise digital diagnosis on the basis of general large models (such as GPT-4, BERT), supports the joint embedded representation of structured data (such as ERP system fields), semi-structured data (JSON / XML configuration files), and unstructured data (documents / conference recordings), injects standard documents for the evaluation of enterprise digital transformation by government departments and industries, and improves the judgment accuracy of the model for "digital transformation" indicators. 2) The multi-dimensional evaluation function supports digital evaluation from multiple dimensions such as technical architecture, business processes, data management, and organizational culture, covering all aspects of enterprise digital transformation. 3) The intelligent analysis function mainly uses natural language processing (NLP) and machine learning (ML) technologies to automatically analyze enterprise data and identify potential digital problems. 4) The dynamic learning function adapts to the digital needs of different industries and enterprises through continuous learning and model updates, improving the accuracy and pertinence of diagnosis. 5) The interactive diagnosis function supports natural language interaction with enterprise managers, obtains enterprise needs through dialogue, and provides real-time diagnosis results and optimization suggestions.

[0041] In a specific embodiment, the information extraction and diagnosis module extracts feature information from unstructured data in the dataset based on the NLP natural language processing large model, and inputs the feature information into the prediction model, so that the prediction model predicts potential risks according to the association relationship between the feature information and the structured data in the dataset, including the following steps: S1: Connect the unstructured features and structured indicators through ontological relationships to form inferable association rules, and adjust the association strength coefficient according to the industry to which the enterprise belongs.

[0042] S2: Construct a dual-time-axis alignment model containing a structured axis and an unstructured axis, and predict potential risks by matching the patterns of the structured axis and the unstructured axis.

[0043] Specifically, the construction process of the structured axis S includes: collecting 9 types of core time series indicators in the enterprise system (S = [order delivery cycle, equipment OEE, inventory turnover rate, CRM system login rate, production plan compliance rate, data interface call success rate, cross-departmental collaboration time consumption, training coverage rate, fault work order resolution duration]), implementing cycle alignment, resampling according to the enterprise production cycle to generate an equally spaced time series, calculating the month-on-month change rate of each cycle's indicators and the multi-cycle moving standard deviation, and constructing a multi-dimensional structured feature vector. The construction process of the unstructured axis U: aggregating the time series features of four types of unstructured data sources, U = [customer service work order sentiment index, abnormal word frequency in meeting recordings, knowledge gap rate in training documents, keyword density in equipment maintenance logs], setting a sliding window to align with the S axis, extracting weekly granularity features for work order-based, recorded speech transcription, etc., applying attention weights to the unstructured data within the window to highlight high-risk periods, and mapping the features of each window to the corresponding risk level. Applying sine position encoding to the S axis and U axis: PE(t, 2k) = sin(t / (10000 2k / d )) PE(t, 2k + 1) = cos(t / (10000 2k / d )) In the formula, d represents the dimension, and t represents the timestamp.

[0044] S3: Calculate the attention weights of the unstructured features for each structured indicator through the Transformer mechanism and generate a risk heat map.

[0045] Specifically, for dynamic time warping alignment, define the industry-sensitive distance metric:

[0046] In the formula, represents the proportion weight of the first industry in the enterprise's various industries, represents the proportion weight of the second industry in the enterprise's various industries, is the comprehensive weight, is the included angle of the feature vectors. Construct a cumulative distance matrix, find the optimal alignment path through backtracking, and identify the lag pattern.

[0047] Design a multi-head time attention mechanism, and use the following formula for the attention weights of each structured indicator:

[0048] In the formula, Q is the feature of the S axis, and K / V is the feature of the U axis.

[0049] S4: When the correlation degree between the unstructured features and the structured indicators is detected to exceed the preset threshold, generate a risk chain, conduct an intervention simulation on the generated risk chain, verify the effectiveness of the suggestions through causal inference, and output a confidence score.

[0050] Exemplarily, the information extraction and diagnosis module: Based on large model technology, this module deeply analyzes and diagnoses enterprise data, and identifies bottleneck problems in digital transformation, mainly including four functions: information extraction function, intelligent diagnosis function, risk prediction, and optimization suggestions. 1) Information extraction: Using natural language processing (NLP) technology, extract key information from unstructured data (such as documents, emails, reports), and conduct correlation analysis with structured data; 2) Intelligent diagnosis: Based on the deep learning algorithm of the large model, conduct multi-dimensional analysis of enterprise data, and identify bottleneck problems in digital transformation, such as backward technical architecture, data islands, redundant business processes, etc.; 3) Risk prediction: Through technologies such as time series analysis and regression analysis, predict the possible risks faced by enterprises in the process of digital transformation, such as supply chain interruption, market demand fluctuation, etc. 4) Optimization suggestions: According to the diagnosis results, combined with industry best practices, generate targeted optimization suggestions to help enterprises formulate digital transformation strategies.

[0051] S5: Combine the confidence score with the preliminary diagnosis result to optimize the prediction result.

[0052] In a specific embodiment, refer to Figure 1 as shown, the enterprise digital diagnosis device further includes: A diagnosis report generation module, used to generate a diagnosis report, wherein the diagnosis report includes a digital maturity score, a visual analysis of key bottlenecks, and prioritized transformation suggestions.

[0053] Specifically, the generated diagnosis report includes outline generation, content generation, paragraph division and format adjustment, data visualization and chart citation.

[0054] Exemplary, diagnostic report generation module: This module automatically outputs diagnostic reports based on NLG (natural language generation) technology, including digital maturity scores (0-100 points), visual analysis of key bottlenecks (such as heat maps), prioritized transformation recommendations (short-term / long-term ROI forecasts), etc., mainly including report content generation, reference citation, text polishing, text export and other functions. 1) Report content generation: The enterprise digital diagnosis report content generation function aims to automatically generate a detailed, systematic and targeted diagnostic report based on the results of the enterprise digital diagnosis, mainly including outline generation, content generation, paragraph division and format adjustment, data visualization and chart citation and other functions. 2) Reference citation and understanding: By referring to relevant literature, reports, data and other documents, the diagnostic team can have a more comprehensive and in-depth understanding of the digital status of the enterprise, and provide the enterprise with more accurate diagnostic results and improvement suggestions. This function can automatically identify and cite relevant documents, reports, data and other files to ensure the authority and traceability of the diagnostic report; the intelligent agent can regularly retrieve and update relevant files to ensure that the reference information in the diagnostic report is the latest, and can use natural language processing technology (NLP) to parse reference files and extract key information, opinions and suggestions; through semantic analysis, understand the relevance and context between files to ensure the coherence and depth of the diagnostic report; build a domain knowledge graph to store knowledge points, cases and other structured information in reference files to facilitate rapid retrieval and application of the intelligent agent during the diagnosis process. 3) The text polishing function is for the generated report content, supporting users to continue writing, expand content, streamline content, etc. online. 4) The text export function supports the extraction of data, analysis reports or other documents generated during the diagnosis process from specific software or systems, and saves and transmits them in text format.

[0055] Through the above technical solutions, focusing on various aspects of enterprise digital transformation, we build enterprise digital intelligent diagnosis and analysis functions based on big models, mainly including enterprise digital diagnosis intelligent body, enterprise data collection and storage module, enterprise information extraction and diagnosis module. Through the application of big model technology, a comprehensive and accurate assessment of the enterprise's digital level is achieved, and the device supports real-time diagnosis and optimization, can quickly identify problems and provide targeted suggestions, and help enterprises efficiently promote digital transformation.

[0056] like Figure 4 As shown, the following is an embodiment of the enterprise digital diagnosis method provided by the embodiment of the present disclosure, which belongs to the same inventive concept as the enterprise digital diagnosis device of the above-mentioned embodiments. For details not described in detail in the embodiment of the enterprise digital diagnosis method, please refer to the embodiment of the above-mentioned enterprise digital diagnosis device.

[0057] S400: Collect enterprise data from the enterprise's business system and external data sources, construct a data set, and store the data set in a memory.

[0058] S401: It is constructed by receiving and jointly embedding structured, semi-structured, and unstructured data sets and performing domain fine-tuning on a general large model based on the industry digital transformation evaluation standard document, and is used for the preliminary diagnosis of enterprise digitization.

[0059] S402: Extract feature information from the unstructured data in the data set based on the NLP natural language processing large model, and input the feature information into a prediction model, so that the prediction model predicts potential risks according to the correlation between the feature information and the structured data in the data set. In a specific embodiment, first, a large number of sample data and evaluation criteria are injected through the enterprise digital diagnosis agent for model fine-tuning and adaptation. Then, the data collection and storage module automatically collects enterprise operation data from the enterprise internal system (such as ERP, CRM) and external data sources (such as industry reports, market data), securely stores the data through distributed storage technologies (such as Hadoop, MongoDB), and cleans and preprocesses the data. Again, the information extraction and diagnosis module uses large model technologies (such as GPT, BERT) to deeply analyze the data, identify bottleneck problems in the digital transformation, extract key information from unstructured data through natural language processing (NLP), and perform correlation analysis with structured data. Predict potential risks (such as supply chain disruptions, market demand fluctuations) and generate optimization suggestions (for example, production process improvement suggestions, equipment replacement suggestions, digital software suggestions); finally, through the diagnosis report generation module, according to the diagnosis results, automatically generate a multi-dimensional analysis report covering the current situation analysis, problem summary, and improvement path, realize data visualization through forms such as charts and dashboards, and support the report to be output in forms such as PDF, HTML, and PPT.

[0060] Figure 5 It is a schematic diagram of the hardware structure of an electronic device for implementing various embodiments of the present invention.

[0061] The enterprise digital diagnosis method provided by the embodiments of this application can be applied to an electronic device. Those skilled in the art can understand that the structure of the electronic device involved in the embodiments of the present invention does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements. In the embodiments of the present invention, the electronic device includes, but is not limited to, a laptop computer, a desktop computer, a workbench, a personal digital assistant, a server, a blade server, a mainframe computer, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the embodiments of this application described herein and / or claimed.

[0062] The electronic device may include a processor, an external memory interface, an internal memory, a universal serial bus (USB) interface, a charging management module, a power management module, a battery, a wireless communication module, an audio module, a speaker, a microphone, a sensor module, keys, a camera, a display screen, and a SIM card interface, etc.

[0063] It can be understood that the structure schematically shown in the embodiments of this application does not constitute a specific limitation on the electronic device. In other embodiments of this application, the electronic device may include more or fewer components than shown in the figure, or combine certain components, or split certain components, or have different component arrangements. The components shown in the figure can be implemented in hardware, software, or a combination of software and hardware.

[0064] The processor may include one or more processing units. For example, the processor may include a central processing unit (CPU), etc., an application processor (AP), a modem processor, a graphics processing unit (GPU), an image signal processor (ISP), a controller, a memory, a video codec, a digital signal processor (DSP), a baseband processor, and / or a neural-network processing unit (NPU), etc. Among them, different processing units may be independent devices or integrated in one or more processors.

[0065] Among them, the processor can be the nerve center and command center of the electronic device. The controller can generate operation control signals according to the instruction operation code and timing signals to complete the control of instruction fetching and execution.

[0066] A memory can also be set in the processor for storing instructions and data. In some embodiments, the memory in the processor is a cache memory. This memory can save the instructions or data that the processor has just used or recycled. If the processor needs to use the instruction or data again, it can directly call it from this memory. This avoids repeated accesses, reduces the waiting time of the processor, and thus improves the system efficiency.

[0067] The external memory interface can be used to connect an external memory card, such as a MicroSD card, to implement the storage capacity expansion of the electronic device. The external memory card communicates with the processor through the external memory interface to achieve the data storage function. For example, files such as music and videos are saved in the external memory card.

[0068] The internal memory can be used to store computer-executable program codes, and the computer-executable program codes include instructions. The processor executes various functional applications and data processing of the electronic device by running the instructions stored in the internal memory. The internal memory can include a program storage area and a data storage area. The internal memory can include a high-speed random access memory and can also include a non-volatile memory, such as at least one disk storage device, a flash memory device, a universal flash storage (UFS), etc.

[0069] The wireless communication function of the electronic device can be implemented through an antenna, a wireless communication module, a modulation and demodulation processor, a baseband processor, etc.

[0070] The wireless communication module can provide wireless communication solutions applied to the electronic device, including wireless local area networks (WLANs) (such as wireless fidelity (Wi-Fi) networks), Bluetooth (BT), global navigation satellite systems (GNSSs), frequency modulation (FM), near field communication (NFC), infrared technology (IR), etc.

[0071] The electronic device can implement audio functions, etc. through an audio module, a speaker, a receiver, a microphone, a headphone interface, an application processor, etc.

[0072] An electronic device can implement a shooting function through an ISP, a camera, a video codec, a GPU, a display screen, an application processor, etc.

[0073] An electronic device can implement a display function through a GPU, a display screen, an application processor, etc.

[0074] The GPU is a microprocessor for image processing, connecting the display screen and the application processor. The GPU is used to perform mathematical and geometric calculations for graphics rendering. The processor may include one or more GPUs, which execute program instructions to generate or change display information.

[0075] The display screen is used to display images, videos, etc. The display screen includes a display panel.

[0076] In the storage medium provided by this application, there is a program product that can implement the enterprise digital diagnosis method.

[0077] The enterprise digital diagnosis method includes: collecting enterprise data from the enterprise's business system and external data sources, constructing a data set, and storing the data set in a memory; receiving and jointly embedding structured, semi-structured, and unstructured data sets, and constructing based on the industry digital transformation evaluation standard document for domain fine-tuning of a general large model for the preliminary diagnosis of enterprise digitization; extracting feature information from the unstructured data in the data set based on the NLP natural language processing large model, and inputting the US feature information into a prediction model, so that the prediction model predicts potential risks according to the association relationship between the feature information and the structured data in the data set. In some possible implementation manners, the subject matter name of the present disclosure, the enterprise digital diagnosis method and system, can be implemented in the form of a program product, which includes program code. When the program product runs on a terminal device, the program code is used to cause the terminal device to execute the steps according to various exemplary embodiments of the present disclosure described in the above "Exemplary Method" section of this specification.

[0078] The storage medium of the present disclosure can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0079] The foregoing description of the disclosed embodiments enables those skilled in the art to practice or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An enterprise digital diagnosis device, characterized in that, Including: An enterprise data collection and storage module, which is used to collect enterprise data from the enterprise's business systems and external data sources, construct a data set, and store the data set in a memory; An enterprise digital diagnosis agent, which is constructed by receiving and jointly embedding structured, semi-structured and unstructured data sets and fine-tuning a general large model based on an industry digital transformation evaluation standard document, and is used for the preliminary diagnosis of enterprise digitization; An information extraction and diagnosis module, which is used to extract feature information from unstructured data in the data set based on an NLP natural language processing large model, and input the feature information into a prediction model, so that the prediction model predicts potential risks according to the correlation between the feature information and the structured data in the data set.

2. The enterprise digital diagnosis device according to claim 1, wherein The enterprise digital diagnosis agent includes: A knowledge-driven decision-making core, which includes an industry knowledge graph and a transformation practice library, and realizes digital evaluation in four dimensions of technical architecture, business process, data management, and organizational culture through dynamic knowledge retrieval, multi-round dialogue reasoning, and semantic association analysis; A dynamic evolution module, which continuously optimizes the large model parameters through historical diagnosis feedback data and constructs an industry adaptive learning mechanism to improve the cross-domain diagnosis accuracy; A collaborative interaction interface, which is configured to perform multi-round semantic interaction with enterprise managers, obtain requirements and visualize the diagnosis process, and form a closed-loop diagnosis link.

3. The enterprise digital diagnosis device according to claim 1, characterized in that, The process of the enterprise digital diagnosis agent receiving and jointly embedding structured, semi-structured and unstructured data sets includes: For structured data, through a pre-trained entity alignment model, map the table fields in the structured data to the diagnosis standard ontology to generate a relational vector with semantic labels; For semi-structured data, construct a tree structure, extract node features, and generate a multi-dimensional graph vector in combination with a multi-head attention mechanism; For unstructured data, adopt a two-stream encoder architecture to extract feature vectors, and perform alignment processing on the extracted feature vectors to eliminate modality differences. Among them, for text streams, perform layout analysis, extract hierarchical features of title-paragraph-table, and generate multi-dimensional semantic vectors. For speech streams, perform ASR transcription and extract prosodic features and semantic features; Input the relational vector, multi-dimensional graph vector, multi-dimensional semantic vector, and prosodic features and semantic features into a gated fusion network to allocate learning weights and output a multi-dimensional joint embedding vector; According to the industry to which the input data belongs, splice the multi-dimensional joint embedding vector to form a domain-sensitive feature.

4. The enterprise digital diagnosis device according to claim 3, characterized in that The process of constructing an enterprise digital diagnosis agent by receiving and jointly embedding structured, semi-structured and unstructured data sets and fine-tuning a general large model based on an industry digital transformation evaluation standard document includes: Perform corpus annotation on the industry digital transformation evaluation standard document using extraction rules to construct a domain corpus; Simulate enterprise digital scenarios based on the knowledge graph and the domain corpus to generate dialogue samples with diagnosis labels; Build a multi-task fine-tuning framework. For the metric classification task, use cross-entropy loss; for the entity annotation task, use the BIOES tag system plus a CRF layer; for the inference chain generation task, use a Seq2Seq architecture for training. Freeze the first 24 layers of the general large model, fine-tune the rank decomposition matrix of the last 6 layers, and use dialogue samples to train the general large model. Among them, a domain-specific prompt is concatenated before the input data to enhance the model's sensitivity to the evaluation criteria.

5. The enterprise digital diagnosis device according to claim 3, characterized in that, The information extraction and diagnosis module extracts feature information from the unstructured data in the dataset based on the NLP natural language processing large model, and inputs the feature information into the prediction model, so that the prediction model predicts potential risks according to the correlation between the feature information and the structured data in the dataset, including: Connect the unstructured features and structured metrics through an ontology relationship to form an inferable association rule, and adjust the association strength coefficient according to the industry to which the enterprise belongs. Build a dual-time-axis alignment model including a structured axis and an unstructured axis, and predict potential risks by matching the patterns of the structured axis and the unstructured axis. Calculate the attention weights of the unstructured features for each structured metric through the Transformer mechanism and generate a risk heat map. When the correlation degree between the unstructured features and the structured metrics is detected to exceed a preset threshold, generate a risk chain, perform an intervention simulation on the generated risk chain, verify the effectiveness of the suggestions through causal inference, and output a confidence score. Combine the confidence score with the preliminary diagnosis result to optimize the prediction result.

6. The enterprise digital diagnosis device according to claim 1, wherein, The enterprise digital diagnosis device further includes: A diagnosis report generation module for generating a diagnosis report, where the diagnosis report includes a digital maturity score, a visual analysis of key bottlenecks, and prioritized transformation suggestions.

7. The enterprise digital diagnosis device according to claim 6, characterized in that The generated diagnosis report includes outline generation, content generation, paragraph division and format adjustment, data visualization, and chart citation.

8. An enterprise digital diagnosis method applied to the enterprise digital diagnosis device according to any one of claims 1-7, characterized in that, Including: Collect enterprise data from the enterprise's business system and external data sources, build a dataset, and store the dataset in a memory. It is built by receiving and jointly embedding structured, semi-structured, and unstructured datasets and performing domain fine-tuning on the general large model based on the industry digital transformation evaluation standard document for the preliminary diagnosis of enterprise digitization. Extract feature information from the unstructured data in the dataset based on the NLP natural language processing large model, and input the feature information into the prediction model, so that the prediction model predicts potential risks according to the correlation between the feature information and the structured data in the dataset.

9. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the enterprise digital diagnosis method as described in claim 8.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the enterprise digital diagnosis method as described in claim 8.

Images