CN-121504307-B - Cross-border logistics flow automatic generation method and device based on natural language analysis

CN121504307BCN 121504307 BCN121504307 BCN 121504307BCN-121504307-B

Abstract

The application relates to the technical field of artificial intelligence and discloses a method and a device for automatically generating a cross-border workflow based on natural language analysis, wherein the method comprises the steps of carrying out cross-language semantic analysis and task structuring on natural language instructions to generate structured task description comprising logistics operation logic, target area information and equipment interaction points; the method comprises the steps of carrying out logic reasoning and process construction on a large language model based on structured task description, generating an executable logistics process comprising equipment control nodes and compliance document generation nodes, carrying out protocol dynamic mapping on the equipment control nodes based on an equipment function vector space alignment mechanism, generating equipment driving strategies, carrying out compliance content dynamic generation processing on the compliance document generation nodes based on a multi-mode rule base and real-time policy information, obtaining a format compliance document, and finally integrating to form a logistics process execution package and deploying. The method can effectively improve the customization efficiency, suitability and compliance of the cross-border logistics flow.

Inventors

WANG XUETENG
ZENG WEIJIA
CHEN DAWEI
XU LINGZI
HE ZHONGQING
XU KUNYANG
ZHAO SHAN
Xie Qiongbing

Assignees

深圳市明心数智科技有限公司

Dates

Publication Date: 20260505
Application Date: 20260113

Claims (8)

1. A natural language analysis-based cross-border logistics process automatic generation method is characterized by comprising the following steps: Performing cross-language semantic analysis and task structuring on the received natural language instruction to generate structured task description comprising logistics operation logic, target area information and equipment interaction points; Based on the structured task description, driving a large language model to carry out logic reasoning and flow construction, and generating an executable logistics flow comprising equipment control nodes and compliance document generation nodes; Analyzing the equipment control node, extracting corresponding functional requirements, converting the functional requirements into standardized requirement feature vectors, respectively inputting the requirement feature vectors and functional description vectors of each candidate protocol into a pre-trained feature encoder and mapping the pre-trained feature encoder and the functional description vectors to a shared vector representation space, respectively calculating the matching degree score between the requirement feature vectors and each candidate protocol vector in the shared vector representation space by adopting a similarity calculation function based on distance measurement, sequencing all candidate protocols according to the matching degree score, determining the candidate protocol with the highest score as an optimal matching target protocol, compiling the logic operation of the equipment control node into a plug-and-play equipment driving package according to the communication specification and the data format of the optimal matching target protocol, and forming a corresponding equipment driving strategy based on the equipment driving package; carrying out dynamic generation processing on the compliance content based on a multi-mode rule base and real-time policy information on a compliance document generation node in the executable logistics flow to obtain a format compliance document meeting the requirement of a target area; Integrating the executable logistics flow, the equipment driving strategy and the format compound specification file to form a logistics flow execution package and deploying.
2. The method for automatically generating a cross-border logistics flow based on natural language parsing according to claim 1, wherein the step of performing cross-language semantic parsing and task structuring on the received natural language instruction to generate a structured task description including logistics operation logic, target area information and equipment interaction points comprises the steps of: Inputting the received multilingual natural language instruction into a pre-trained multilingual language semantic model to obtain a preliminary semantic coding sequence; based on an attention mechanism, the preliminary semantic coding sequence and preset regional culture feature knowledge are interacted, and a feature enhancement vector fusing regional culture context is obtained through calculation; and generating a structured task description comprising logistics operation logic, target area information and equipment interaction points after the entity relation joint decoding processing according to the characteristic enhancement vector.
3. The automatic generation method of cross-border logistics flow based on natural language parsing according to claim 2, wherein the interaction between the preliminary semantic coding sequence and the preset regional culture feature knowledge based on the attention mechanism is calculated to obtain a feature enhancement vector of the fusion regional culture context, and the method comprises the following steps: according to the region keywords obtained by the primary semantic coding sequence identification, loading corresponding multidimensional cultural feature vectors from a region cultural feature library; calculating the correlation weight of each position in the preliminary semantic coding sequence and the multidimensional cultural feature vector through a cross attention layer; And dynamically weighting and aggregating the multidimensional culture feature vectors based on the relevance weights to generate the feature enhancement vectors related to the current instruction semantics.
4. The automatic generation method of cross-border logistics flow based on natural language parsing according to claim 1, wherein the driving a large language model to perform logical reasoning and flow construction based on the structured task description generates an executable logistics flow including a device control node and a compliance document generation node, and the method comprises the following steps: Converting the structured task description into large language model prompt words containing context and constraints; Driving the large language model to gradually infer based on the prompt words, and generating an initial flow logic diagram represented by nodes and edges; and performing executable compiling on the initial flow logic diagram, binding a specific API call, a data source or a document template for each node, and marking the equipment interaction point and the compliance output point as nodes of a specific type to form the executable logistics flow.
5. The automatic generation method of cross-border logistics flow based on natural language analysis according to claim 1, wherein the generating node of the compliance document in the executable logistics flow dynamically generates and processes the compliance content based on a multi-mode rule base and real-time policy information to obtain a format compound rule file meeting the requirements of a target area, and the method comprises the following steps: According to the target area and the document type associated with the compliance document generation node, text rule terms, image form templates and associated real-time policy update abstracts are retrieved from a multi-mode compliance rule base; Carrying out structural constraint extraction on the retrieved text rule clauses, carrying out key field positioning and format analysis on the image form template, and carrying out timeliness verification and conflict resolution on the extracted structural constraint and key field by combining the real-time policy updating abstract; And automatically filling service data required by the running of the executable logistics flow according to the verified structural constraint and field format, and generating the format compound specification file.
6. The method for automatically generating a cross-border logistics flow based on natural language parsing according to claim 5, wherein the steps of extracting structural constraints from the retrieved text rule terms, performing key field positioning and format parsing on the image form template, and performing timeliness verification and conflict resolution on the extracted structural constraints and key fields in combination with the real-time policy update abstract include: Inputting the text rule terms and the real-time policy update abstract into a text analysis model, and extracting a text constraint condition list and an effective time range thereof; inputting the image form template into a document understanding model, and identifying corresponding fixed fields, variable fields and typesetting specifications; Mapping the constraint condition list and the variable field and checking compliance to ensure that the filled data meets all effective constraint conditions and accords with a form layout specification.
7. The method for automatically generating a cross-border logistics flow based on natural language parsing according to claim 6, wherein the updating mode of the multi-mode compliance rule base comprises: Continuously monitoring the release state and revision dynamics of policies, regulations and standard forms of a target area by connecting a plurality of official sources; When the release state or revision dynamic change is monitored, automatically capturing updated content, and analyzing the updated content into a structural rule object by utilizing a multi-mode understanding model; and storing the structured rule object, the version number, the effective date and the source region in association with the multi-mode compliance rule base, and automatically triggering the compliance reevaluation of the related history flow.
8. The device for automatically generating the cross-border logistics flow based on natural language analysis is characterized by comprising the following components: The instruction analysis module is used for performing cross-language semantic analysis and task structuring on the received natural language instruction to generate structured task description comprising logistics operation logic, target area information and equipment interaction points; the flow construction module is used for driving a large language model to carry out logic reasoning and flow construction based on the structured task description, and generating an executable logistics flow comprising equipment control nodes and compliance document generation nodes; The system comprises a policy generation module, a logic operation module and a logic operation module, wherein the policy generation module is used for analyzing the equipment control node, extracting corresponding functional requirements, converting the functional requirements into standardized requirement feature vectors, inputting the requirement feature vectors and functional description vectors of each candidate protocol respectively, inputting the pre-trained feature encoders and mapping the functional description vectors to a shared vector representation space, adopting a similarity calculation function based on distance measurement in the shared vector representation space to calculate the matching degree score between the requirement feature vectors and each candidate protocol vector respectively, sorting all candidate protocols according to the matching degree score, determining the candidate protocol with the highest score as an optimal matching target protocol, compiling the logic operation of the equipment control node into a plug-and-play equipment driving package according to the communication specification and the data format of the optimal matching target protocol, and forming a corresponding equipment driving policy based on the equipment driving package; The compliance file module is used for dynamically generating and processing compliance content based on a multi-mode rule base and real-time policy information for a compliance document generation node in the executable logistics flow to obtain a format compliance file meeting the requirements of a target area; and the integration and deployment module is used for integrating the executable logistics flow, the equipment driving strategy and the format combination rule file to form a logistics flow execution package and deploying the logistics flow execution package.

Description

Cross-border logistics flow automatic generation method and device based on natural language analysis Technical Field The application relates to the technical field of artificial intelligence, in particular to a method and a device for automatically generating a cross-border workflow based on natural language analysis. Background Under the background of globalization trade rapid development, cross-border logistics scenes cover core requirements such as multi-region cooperation, multi-device linkage, multi-compliance requirement adaptation and the like, and the requirements of enterprises on rapid construction, flexible adaptation and compliance landing of logistics flows are increasingly urgent. At present, all links of the existing logistics flow are mainly finished manually. The natural language requirements put forward by business personnel are manually translated into technical schemes and code development by technical experts, and the process is time-consuming and easy to generate understanding deviation. Meanwhile, various equipment communication protocols related to the logistics site are different, and an adaptive code needs to be developed in a targeted mode, so that the expansibility of a system is poor, and the online period of a new flow is long. In addition, the requirements of regulations, cultures and customs clearance in different countries or regions are complex and dynamically changed, and the difficulty and risk of manually ensuring the compliance of the whole process are extremely high. Therefore, the unstructured business instructions are difficult to quickly and accurately convert into end-to-end executable logistics flows which can be deployed immediately and meet regional requirements, and the problems of low customization efficiency, and insufficient flexibility and adaptability of cross-border logistics flows are caused. The foregoing description is provided for general background information and does not necessarily constitute prior art. Disclosure of Invention The embodiment of the application provides a method and a device for automatically generating a cross-border logistics flow based on natural language analysis, which can automatically convert a natural language instruction related to the cross-border logistics into an end-to-end executable logistics flow and complete deployment, and solve the problems of low flow customization efficiency, equipment suitability and insufficient compliance guarantee in the prior art. In a first aspect, an embodiment of the present application provides a method for automatically generating a cross-border logistics flow based on natural language parsing, including: Performing cross-language semantic analysis and task structuring on the received natural language instruction to generate structured task description comprising logistics operation logic, target area information and equipment interaction points; Based on the structured task description, driving a large language model to carry out logic reasoning and flow construction, and generating an executable logistics flow comprising equipment control nodes and compliance document generation nodes; performing protocol dynamic mapping on the equipment control node in the executable logistics flow based on an equipment function vector space alignment mechanism to generate a corresponding equipment driving strategy; carrying out dynamic generation processing on the compliance content based on a multi-mode rule base and real-time policy information on a compliance document generation node in the executable logistics flow to obtain a format compliance document meeting the requirement of a target area; Integrating the executable logistics flow, the equipment driving strategy and the format compound specification file to form a logistics flow execution package and deploying. Further, in some embodiments of the present application, the performing cross-language semantic parsing and task structuring on the received natural language instruction generates a structured task description including logistics operation logic, target area information and device interaction points, including: Inputting the received multilingual natural language instruction into a pre-trained multilingual language semantic model to obtain a preliminary semantic coding sequence; based on an attention mechanism, the preliminary semantic coding sequence and preset regional culture feature knowledge are interacted, and a feature enhancement vector fusing regional culture context is obtained through calculation; and generating a structured task description comprising logistics operation logic, target area information and equipment interaction points after the entity relation joint decoding processing according to the characteristic enhancement vector. Further, in some embodiments of the present application, the interaction between the preliminary semantic coding sequence and the preset regional culture feature knowledge based on the at