Material standardization method and device
By acquiring, combining, and standardizing material descriptions, the problem of material identification errors was solved, enabling efficient and accurate material identification and standardized processing, thus ensuring product quality and equipment safety for enterprises.
Patent Information
- Application Number
- CN202511685819.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-17
- Publication Date
- 2026-02-10
AI Technical Summary
In the industrial supplies market, the diversity of naming and descriptions of the same industrial material can lead to material identification errors, affecting product quality and equipment safety.
By acquiring material descriptions, determining input types, extracting type text, generating combined description text, extracting attribute key-value pairs, matching standard materials from a standard material library, and supplementing information using a graph reasoning model, the standardization of materials is achieved.
Reduce material identification errors, ensure product quality and equipment safety, and improve the accuracy and efficiency of material identification.
Smart Images
Figure CN121503460A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of computer technology, and more particularly to a method and apparatus for standardizing materials. Background Technology
[0002] The industrial supplies market is a vital component of the global economy, encompassing a wide range of sectors including machinery, electronic components, parts, fasteners, personal protective equipment, and chemical raw materials. In real-world business scenarios, the same industrial material often has multiple names and descriptions. This leads to frequent errors in material identification when interpreting these descriptions, impacting product quality and equipment safety for businesses. Summary of the Invention
[0003] In view of this, embodiments of the present invention provide a method and apparatus for standardizing materials, which can standardize the description of materials, reduce the occurrence of material identification errors, and ensure the product quality and equipment safety of enterprises.
[0004] In a first aspect, embodiments of the present invention provide a method for standardizing materials, comprising:
[0005] Obtain the material description content and determine at least one input type corresponding to the material description content;
[0006] For each input type, determine the material description content for that input type; extract the type text corresponding to the type content;
[0007] Combine the type texts corresponding to each input type to generate the combined description text corresponding to the material description content;
[0008] Extract at least one attribute key-value pair from the combined description text; wherein the attribute key-value pair includes: attribute name and attribute value;
[0009] Based on at least one attribute key-value pair, match the standard material corresponding to the material description from the standard material library.
[0010] Optionally, before determining at least one input type corresponding to the material description content, the method further includes:
[0011] Retrieve the material text from the material description content;
[0012] Several primary standard materials were identified from the standard material library;
[0013] The semantic similarity between the material description and the material text of each first standard material is determined, and the first standard material with the highest semantic similarity is determined as the target standard material.
[0014] If the semantic similarity to the target standard material is greater than the similarity threshold, the target standard material is identified as the standard material corresponding to the material description content.
[0015] In response to the semantic similarity to the target standard material not being greater than a similarity threshold, a step is performed to determine at least one input type corresponding to the material description content.
[0016] Optionally, at least one attribute key-value pair is extracted from the combined description text, including:
[0017] Identify the basic entities in the combined descriptive text;
[0018] In response to the fact that the base entity is an attribute entity, the attribute values corresponding to the base entity are extracted from the combined description text;
[0019] Combine the basic entity and its corresponding attribute values to generate attribute key-value pairs corresponding to the basic entity.
[0020] Optionally, after extracting at least one attribute key-value pair from the combined description text, the method further includes:
[0021] Determine the standard category based on the combined description text;
[0022] Input at least one attribute key-value pair and a standard category into the graph reasoning model to obtain at least one missing key-value pair;
[0023] Merge at least one attribute key-value pair and at least one missing key-value pair to generate a complete key-value pair;
[0024] Based on at least one attribute key-value pair, match the standard materials corresponding to the material description content from the standard material library, including:
[0025] Based on the supplementary key-value pairs, the standard materials corresponding to the material descriptions are matched from the standard material library.
[0026] Optionally, based on at least one attribute key-value pair, standard materials corresponding to the material description content are matched from the standard material library, including:
[0027] Based on the combined description text, determine the standard category corresponding to the material description content;
[0028] Based on at least one attribute key-value pair, match the standard material corresponding to the material description content from multiple standard materials under the standard category in the standard material library.
[0029] Optionally, based on at least one attribute key-value pair, standard materials corresponding to the material description content are matched from the standard material library, including:
[0030] Multiple second standard materials were identified from the standard material library;
[0031] For each second standard material, at least one standard attribute pair corresponding to the second standard material is obtained from the standard material library; the matching degree corresponding to the second standard material is determined based on at least one standard attribute pair and at least one attribute key-value pair.
[0032] The second standard material with the highest matching degree is determined as the standard material corresponding to the material description content.
[0033] Secondly, embodiments of the present invention provide a material standardization apparatus, comprising:
[0034] The content acquisition module is used to acquire material description content and determine at least one input type corresponding to the material description content;
[0035] The text extraction module is used to determine the type content of the material description for each input type and extract the type text corresponding to the type content.
[0036] The text combination module is used to combine the type texts corresponding to each input type to generate the combined description text corresponding to the material description content.
[0037] The key-value pair extraction module is used to extract at least one attribute key-value pair from the combined description text; wherein the attribute key-value pair includes: attribute name and attribute value;
[0038] The standardization module is used to match the standard material corresponding to the material description content from the standard material library based on at least one attribute key-value pair.
[0039] Thirdly, embodiments of the present invention provide an electronic device, including:
[0040] One or more processors;
[0041] A storage device for storing one or more programs that, when executed by one or more processors, cause the one or more processors to implement the methods of any of the above embodiments.
[0042] Fourthly, embodiments of the present invention provide a computer-readable medium having a computer program stored thereon, which, when executed by a processor, implements the method of any of the above embodiments.
[0043] Fifthly, embodiments of the present invention provide a computer program product, including a computer program, wherein the computer program, when executed by a processor, implements the method of any of the above embodiments.
[0044] One embodiment of the above invention has the following advantages or beneficial effects: Type text of the material description content for each input type is extracted separately. The type texts are combined to generate a combined description text. Then, at least one attribute key-value pair is extracted from the combined description text. Finally, using the attribute key-value pairs, the standard material corresponding to the material description content is matched from a standard material library. This enables the standardization of the logistics described in the material description content, reducing material identification errors and ensuring product quality and equipment safety for enterprises.
[0045] Furthermore, the combined description text is generated by combining all extracted type texts. This combined description text contains comprehensive material information. The attribute key-value pairs extracted from the combined description text accurately and specifically represent the material characteristics. Using these attribute key-value pairs for standard logistics matching further improves the effectiveness of material standardization.
[0046] The further effects of the aforementioned unconventional alternative methods will be explained below in conjunction with specific implementation methods. Attached Figure Description
[0047] The accompanying drawings are provided to better understand the invention and are not intended to unduly limit the scope of the invention. Wherein:
[0048] Figure 1 This is a schematic diagram of a material standardization method provided in one embodiment of the present invention;
[0049] Figure 2 This is a schematic diagram of the process of a category determination method provided in an embodiment of the present invention;
[0050] Figure 3 This is a schematic diagram of a material standardization method provided in another embodiment of the present invention;
[0051] Figure 4 This is a schematic diagram of a material standardization method provided in another embodiment of the present invention;
[0052] Figure 5 This is a schematic diagram of the structure of a material standardization device provided in one embodiment of the present invention;
[0053] Figure 6 This is a schematic diagram of the structure of a computer system suitable for implementing terminal devices or servers of the present invention. Detailed Implementation
[0054] The following description, in conjunction with the accompanying drawings, illustrates exemplary embodiments of the present invention, including various details to aid understanding. These details should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the invention. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.
[0055] It should be noted that the acquisition, storage, use, and processing of data in the technical solutions of this invention comply with the relevant provisions of national laws and regulations.
[0056] Figure 1 This is a schematic diagram of a material standardization method provided in one embodiment of the present invention. Figure 1 As shown, the method includes:
[0057] Step 101: Obtain the material description content and determine at least one input type corresponding to the material description content.
[0058] Material descriptions can originate from emails, Requests for Quotation (RFQs), etc. Material descriptions can be categorized into various input types, including text, images, and tables. Specifically, material descriptions can include: text input (material text), images input (material images), and tables input (material tables).
[0059] Step 102: For each input type, determine the material description content for that input type; extract the type text corresponding to the type content.
[0060] If the input type includes text, the material text is extracted from the material description content. Material text is the material content described in words. Material text can be structured or unstructured text. Material text is simply the type of text corresponding to the original text.
[0061] If the input type includes an image, the material image is extracted from the material description. Image recognition technology is then used to extract the image text from the material image. The image text may include information such as the material's appearance, packaging, and labeling. The image text is essentially the type text corresponding to the image.
[0062] If the input type includes a table, the material table is extracted from the material description. Then, the table text is extracted from the material table. The table text may include material specifications, parameters, etc. The table text is the type text corresponding to the table.
[0063] Step 103: Combine the type texts corresponding to each input type to generate the combined description text corresponding to the material description content.
[0064] After combining the type texts corresponding to each input type to generate the combined description text corresponding to the material description content, duplicate content in the combined description text can be deleted to shorten the content of the combined description text and improve the system execution efficiency.
[0065] If the material description content includes: material text entered using text input, material image entered using image input, and material table entered using table input, then the material text, image text, and table text are combined to generate a combined description text corresponding to the material description content.
[0066] Step 104: Extract at least one attribute key-value pair from the combined description text; wherein the attribute key-value pair includes: attribute name and attribute value.
[0067] The attribute name is the name of the material attribute. Attribute names include: brand, material name, product series, version number, product model, storage capacity, screen size, operating system, etc. The attribute value is the corresponding value for the attribute name. For example, the attribute value for storage capacity is 128GB. The attribute value for screen size is 6.1 inches. The attribute value for operating system is iOS 16. The attribute value for color is Space Black.
[0068] In one embodiment of the present invention, extracting at least one attribute key-value pair from the combined description text includes: identifying a basic entity in the combined description text; extracting the attribute value corresponding to the basic entity from the combined description text in response to the basic entity being an attribute entity; and combining the basic entity and its corresponding attribute value to generate an attribute key-value pair corresponding to the basic entity.
[0069] Basic entities can include: product entities, material entities, company entities, geographic location entities, and attribute entities. Entity recognition models can be used to identify these basic entities within the combined descriptive text. Alternatively, context-dependent word representations can be extracted from the combined descriptive text to capture sequence information, and dependencies between labels can be modeled to calculate the optimal label sequence. Then, based on the context-dependent word representations, sequence information, and the optimal label sequence, the basic entities within the combined descriptive text can be identified.
[0070] After identifying the basic entities in the descriptive text, attribute relationship reasoning is performed based on these entities. Attribute relationships include: inter-entity relationships, attribute dependencies, hierarchical relationships, and constraint relationships. Inter-entity relationships are semantic relationships between two entities. Attribute dependencies are logical dependencies between attributes. Hierarchical relationships include: hierarchical relationships, containment relationships, etc. Constraint relationships are the constraints on attribute values. Through entity recognition and attribute relationship reasoning, deep semantic understanding is achieved.
[0071] Attribute entities are entities abstracted from material attributes. Attribute entities can include: material entities, product entities, specification entities, brand entities, model entities, size entities, quantity entities, price entities, etc. If the base entity is an attribute entity, the attribute values corresponding to the base entity can be directly extracted from the combined description text. Alternatively, attribute values corresponding to the base entity can be extracted from the combined description text based on attribute relationships inferred from attribute relationships.
[0072] Step 105: Based on at least one attribute key-value pair, match the standard material corresponding to the material description content from the standard material library.
[0073] The standard material library stores material information for multiple standard materials. Standard materials are materials that have been standardized and normalized. Different names and descriptions of the same material all correspond to a single standard material in the standard material library. For example, "Hex socket head cap screw" and "Allen bolt" both correspond to "Hex socket head cap screw" in the standard material library.
[0074] The material information for standard materials includes: standard material identifier, material description, and attribute values for multiple standard attribute names. Different material categories correspond to different material attributes. Standard attribute names may include: standard category, standard material name, brand, product series, and version number.
[0075] Retrieve the target material name corresponding to the material description. Specifically, determine the attribute pair corresponding to the item name from at least one attribute key-value pair in the item description. Retrieve the target material name corresponding to the material description from the attribute pair corresponding to the item name.
[0076] Retrieve the target category corresponding to the material description. Specifically, determine the attribute pair corresponding to the item category from at least one attribute key-value pair. Retrieve the target category corresponding to the material description from the attribute pair corresponding to the item category.
[0077] From the standard material library, match the target standard material corresponding to the material description. The standard material name of the target standard material matches the target material name, and the standard category of the target standard material matches the target category.
[0078] In one embodiment of the present invention, matching the standard material corresponding to the material description content from the standard material library based on at least one attribute key-value pair includes: determining a plurality of second standard materials from the standard material library; for each second standard material, obtaining at least one standard attribute pair corresponding to the second standard material from the standard material library; determining the matching degree corresponding to the second standard material based on at least one standard attribute pair and at least one attribute key-value pair; and determining the second standard material with the highest matching degree as the standard material corresponding to the material description content.
[0079] The second standard material can be a standard material whose name is the same as or similar to that of the target material. The second standard material can also be a standard material whose standard category is the same as or similar to that of the target material, etc. The standard attribute pairs of the second standard material include: the standard attribute name and the corresponding attribute value.
[0080] Determining the matching degree of the second standard material based on at least one standard attribute pair and at least one attribute key-value pair includes: for each attribute key-value pair corresponding to the item description content, querying the standard attribute pair corresponding to the attribute key-value pair from at least one standard attribute pair. The attribute name in the attribute key-value pair matches the attribute name in the queried standard attribute pair. For example, both the attribute name in the attribute key-value pair and the attribute name in the queried standard attribute pair are material names.
[0081] If the attribute value in the attribute key-value pair matches the attribute value in the retrieved standard attribute pair, then the retrieved standard attribute pair is determined to be a matching attribute pair; the number of matching attribute pairs is taken as the matching degree corresponding to the second standard material. The second standard material with the highest matching degree is taken as the standard material corresponding to the material description content.
[0082] After matching the standard materials corresponding to the material description from the standard material library, the category, brand, specifications, and required attributes of the material description are accurately determined based on the standard materials corresponding to the material description, laying the data foundation for subsequent operations such as supplier matching, price comparison, and product matching.
[0083] In this embodiment of the invention, type text of the material description content for each input type is extracted separately. The extracted type texts are combined to generate a combined description text. Then, at least one attribute key-value pair is extracted from the combined description text. Finally, using the attribute key-value pairs, the standard material corresponding to the material description content is matched from a standard material library. This embodiment of the invention can standardize material description content, reduce material identification errors, and ensure product quality and equipment safety for enterprises.
[0084] Furthermore, the combined description text is generated by combining the extracted type texts. This combined description text contains comprehensive material information. The attribute key-value pairs extracted from the combined description text can accurately and specifically describe the material's characteristics, further improving the effectiveness of material standardization.
[0085] In one embodiment of the present invention, matching the standard material corresponding to the material description content from the standard material library based on at least one attribute key-value pair includes: determining the standard category corresponding to the material description content based on the combined description text; and matching the standard material corresponding to the material description content from multiple standard materials under the standard category in the standard material library based on at least one attribute key-value pair.
[0086] To manage standard materials, multi-level categories can be set up to solve problems such as the complexity of industrial product category systems and vague customer descriptions. Based on industry standards and historical data, a multi-level category tree containing tens of thousands of nodes is constructed. After the category tree is constructed, an adaptive mechanism is adopted to automatically adjust and expand the category tree structure according to new products and market changes.
[0087] When it is necessary to determine the standard category corresponding to the material description content, the combined description text and category tree are input into the improved tree LSTM network to realize dynamic path recognition from the root node to the leaf node.
[0088] Figure 2 This is a schematic diagram illustrating the flow of a category determination method according to an embodiment of the present invention. Figure 2 As shown, the material description is: 304 stainless steel M12 anti-loosening nut. If a three-level category recognition architecture is used, the first-level category corresponding to this material description is: Industrial Parts. The second-level category is: Fasteners. The third-level category is: Anti-loosening nut. Based on the categories corresponding to this material description, multiple material attribute values are identified.
[0089] The standard category corresponding to the material description can be any category within the multi-level categories corresponding to the material description. Matching only multiple standard materials under the standard category can improve the efficiency and accuracy of standard product matching.
[0090] Figure 3 This is a schematic diagram of a material standardization method provided in another embodiment of the present invention. Figure 3 As shown, the method includes:
[0091] Step 301: Obtain the material description content and the material text within it.
[0092] Material text refers to the text content within the material description. Material text is the type of text corresponding to the text input type.
[0093] Step 302: Identify multiple first standard materials from the standard material library.
[0094] Multiple primary standard materials can be randomly selected from the standard material library. Alternatively, the standard category corresponding to the material description can be determined, and the standard material under the standard category in the standard material library can be selected as the primary standard material.
[0095] Step 303: Determine the semantic similarity between the material description and the material text of each first standard material, and determine the first standard material with the highest semantic similarity as the target standard material.
[0096] Generate a first semantic feature for the material description of the first standard material. Generate a second semantic feature for the material text. Determine the semantic similarity between the first semantic feature and the second semantic feature as the material description of the first standard material and the material text.
[0097] Step 304: In response to the semantic similarity of the target standard material being greater than the similarity threshold, the target standard material is determined as the standard material corresponding to the material description content.
[0098] If the semantic similarity of the target standard material is greater than the similarity threshold, then the material described in the material text is highly similar to the target standard material, and the target standard material is directly identified as the standard material corresponding to the material description content.
[0099] Step 305: In response to the semantic similarity of the target standard material not being greater than the similarity threshold, determine the multiple input types corresponding to the material description content.
[0100] If the semantic similarity to the target standard material is not greater than the similarity threshold, then the similarity between the material described in the material text and the target standard material is low. It is necessary to combine the type content of each input type to determine the standard material corresponding to the material description content.
[0101] Step 306: For each input type, determine the material description content for that input type; extract the type text corresponding to the type content.
[0102] Step 307: Combine the type texts corresponding to each input type to generate the combined description text corresponding to the material description content.
[0103] Step 308: Extract at least one attribute key-value pair from the combined description text; wherein the attribute key-value pair includes: attribute name and attribute value.
[0104] Step 309: Based on at least one attribute key-value pair, match the standard material corresponding to the material description content from the standard material library.
[0105] In the embodiment of this invention, the semantic similarity between the material description and the material text of each first standard material is first determined, and the target standard material that is semantically closest to the material text is identified. Then, based on the semantic similarity between the target standard material and the material text, it is determined how to annotate the material description content. This method can accelerate the material annotation process while accurately matching the standard material corresponding to the material description content.
[0106] Figure 4 This is a schematic diagram of a material standardization method provided in another embodiment of the present invention. Figure 4 As shown, the method includes:
[0107] Step 401: Obtain the material description content and determine at least one input type corresponding to the material description content.
[0108] Step 402: For each input type, determine the material description content for that input type; extract the type text corresponding to the type content.
[0109] Step 403: Combine the type texts corresponding to each input type to generate the combined description text corresponding to the material description content.
[0110] Step 404: Extract at least one attribute key-value pair from the combined description text; wherein the attribute key-value pair includes: attribute name and attribute value.
[0111] Step 405: Determine the standard category based on the combined description text.
[0112] Step 406: Input at least one attribute key-value pair and standard category into the graph reasoning model to obtain at least one missing key-value pair.
[0113] Construct a knowledge graph for the industrial products domain. The knowledge graph contains information such as material attributes, category relationships, application scenarios, and industry standards. Based on graph attention networks and the knowledge graph, construct a graph reasoning model. This model transforms structured knowledge in the industrial products domain into computable vector representations, providing powerful knowledge reasoning capabilities for material standardization.
[0114] Input at least one attribute key-value pair and a standard category into the graph inference model. The graph inference model outputs at least one inference key-value pair. The inference key-value pair includes: attribute name and attribute value. Compare the attribute key-value pairs in the material description with the inference key-value pairs output by the graph inference model to identify missing key-value pairs. Missing key-value pairs do not exist in the attribute key-value pairs of the material description. Missing key-value pairs exist in the inference key-value pairs output by the graph inference model.
[0115] Step 407: Merge at least one attribute key-value pair and at least one missing key-value pair to generate a supplementary key-value pair.
[0116] Step 408: Based on the supplementary key-value pairs, match the standard materials corresponding to the material description content from the standard material library.
[0117] In this embodiment of the invention, a graph inference model is used to supplement the attribute key-value pairs of the material description. The supplemented attribute key-value pairs include richer material feature information. By utilizing the supplemented attribute key-value pairs, a more accurate material annotation effect can be obtained.
[0118] This invention also provides a user feedback interface for collecting evaluation and error correction information on standardized results. An incremental learning mechanism is implemented, periodically updating model parameters with new data. An active learning strategy is adopted, prioritizing the manual annotation of high-value samples. Contrastive learning technology is introduced to improve the model's ability to distinguish similar materials. The knowledge graph is dynamically updated based on market changes and emerging material information, enabling adaptive adjustment of the graph inference model. The identification strategy is automatically optimized based on the characteristics of different industries and regions.
[0119] The solutions proposed in this invention can effectively address key issues in the intelligent identification of industrial materials, achieving efficient, accurate, and safe material identification and standardization, and providing strong technical support for industrial supply chain management.
[0120] Before acquiring the material description content, the process also includes: constructing a training dataset; and using the training dataset to train a standardized model. The standardized model includes: a content extraction model for extracting type-based content, an entity recognition model for identifying entities, a key-value pair generation model for generating key-value pairs, and a graph inference model, etc.
[0121] Before constructing the training dataset, multi-dimensional data was collected. This multi-dimensional data came from historical business opportunity material data, industry standard material library data, etc. The multi-dimensional data included information such as material name, brand, specifications, parameters, and images. Data cleaning, deduplication, text standardization, and image preprocessing were performed on the multi-dimensional data to construct the training dataset.
[0122] For the text content in the material description, a text feature model is used to encode the text content. First, the text feature model outputs the sequence vectors of each text sequence. Then, average pooling is used on the sequence vectors of each text sequence to output the text features. If the text content is structured data, such as specifications, feature engineering is performed on the text content.
[0123] For the images in the material description, a pre-trained image feature model is used to extract image features. Image features may include: material appearance, packaging, labels, etc. Text features and image features are used for subsequent multi-stage semantic understanding processing.
[0124] Multi-stage semantic understanding processing includes: using entity recognition models to identify basic entities in combined descriptive text; and performing attribute relationship reasoning based on basic entities.
[0125] The customer inputs a material description, initiating the business processing flow. The system extracts at least one attribute key-value pair from the combined description text corresponding to the material description. A knowledge graph is used to verify and complete these attribute key-value pairs, ensuring the consistency and completeness of the recognition results. Based on the identified attribute key-value pairs, the material description is mapped to a standard material library to obtain the standard material with the highest matching degree. Based on the obtained standard material, information completion is performed on the material description, resulting in more detailed and specific material data.
[0126] Figure 5 This is a schematic diagram of the structure of a material standardization device provided in one embodiment of the present invention. Figure 5 As shown, the device includes:
[0127] The content acquisition module 501 is used to acquire material description content and determine at least one input type corresponding to the material description content;
[0128] The text extraction module 502 is used to determine the type content of the material description for each input type and extract the type text corresponding to the type content.
[0129] The text combination module 503 is used to combine the type texts corresponding to each input type to generate the combined description text corresponding to the material description content.
[0130] The key-value pair extraction module 504 is used to extract at least one attribute key-value pair from the combined description text; wherein the attribute key-value pair includes: attribute name and attribute value;
[0131] The standardization module 505 is used to match the standard material corresponding to the material description content from the standard material library based on at least one attribute key-value pair.
[0132] Optionally, it also includes:
[0133] The similarity determination module is used to obtain the material text in the material description content;
[0134] Several primary standard materials were identified from the standard material library;
[0135] The semantic similarity between the material description and the material text of each first standard material is determined, and the first standard material with the highest semantic similarity is determined as the target standard material.
[0136] If the semantic similarity to the target standard material is greater than the similarity threshold, the target standard material is identified as the standard material corresponding to the material description content.
[0137] In response to the semantic similarity to the target standard material not being greater than a similarity threshold, a step is performed to determine at least one input type corresponding to the material description content.
[0138] Optionally, the key-value pair extraction module 504 is specifically used for:
[0139] Identify the basic entities in the combined descriptive text;
[0140] In response to the fact that the base entity is an attribute entity, the attribute values corresponding to the base entity are extracted from the combined description text;
[0141] Combine the basic entity and its corresponding attribute values to generate attribute key-value pairs corresponding to the basic entity.
[0142] Optionally, it also includes:
[0143] The category determination module is used to determine the standard category based on the combined description text;
[0144] Input at least one attribute key-value pair and a standard category into the graph reasoning model to obtain at least one missing key-value pair;
[0145] Merge at least one attribute key-value pair and at least one missing key-value pair to generate a complete key-value pair;
[0146] Standardized module 505 is specifically used for:
[0147] Based on the supplementary key-value pairs, the standard materials corresponding to the material descriptions are matched from the standard material library.
[0148] Optionally, the standardized module 505 is specifically used for:
[0149] Based on the combined description text, determine the standard category corresponding to the material description content;
[0150] Based on at least one attribute key-value pair, match the standard material corresponding to the material description content from multiple standard materials under the standard category in the standard material library.
[0151] Optionally, the standardized module 505 is specifically used for:
[0152] Multiple second standard materials were identified from the standard material library;
[0153] For each second standard material, at least one standard attribute pair corresponding to the second standard material is obtained from the standard material library; the matching degree corresponding to the second standard material is determined based on at least one standard attribute pair and at least one attribute key-value pair.
[0154] The second standard material with the highest matching degree is determined as the standard material corresponding to the material description content.
[0155] This invention provides an electronic device, comprising:
[0156] One or more processors;
[0157] A storage device for storing one or more programs that, when executed by one or more processors, cause the one or more processors to implement the methods of any of the above embodiments.
[0158] This invention provides a computer program product, including a computer program that, when executed by a processor, implements the method of any of the above embodiments.
[0159] The following is for reference. Figure 6 It shows a schematic diagram of the structure of a computer system 600 suitable for implementing a terminal device of the present invention. Figure 6 The terminal device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.
[0160] like Figure 6 As shown, the computer system 600 includes a central processing unit (CPU) 601, which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) 602 or programs loaded from storage section 608 into random access memory (RAM) 603. The RAM 603 also stores various programs and data required for the operation of the system 600. The CPU 601, ROM 602, and RAM 603 are interconnected via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.
[0161] The following components are connected to I / O interface 605: an input section 606 including a keyboard, mouse, etc.; an output section 607 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN card, modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to I / O interface 605 as needed. A removable medium 611, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 610 as needed so that computer programs read from it can be installed into storage section 608 as needed.
[0162] In particular, according to the embodiments disclosed in this invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 609, and / or installed from removable medium 611. When the computer program is executed by central processing unit (CPU) 601, it performs the functions defined above in the system of this invention.
[0163] It should be noted that the computer-readable medium shown in this invention can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-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 thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this invention, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.
[0164] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.
[0165] The modules described in the embodiments of the present invention can be implemented in software or hardware. The described modules can also be located in a processor, and for example, can be described as: a content acquisition module, a text extraction module, a text combination module, a key-value pair extraction module, and a standardization module. The names of these modules do not necessarily limit the module itself; for example, the content acquisition module can also be described as "a module that acquires material description content and determines at least one input type corresponding to the material description content."
[0166] In another aspect, the present invention also provides a computer-readable medium, which may be included in the device described in the above embodiments; or it may exist independently and not assembled into the device. The computer-readable medium carries one or more programs, which, when executed by the device, cause the device to include:
[0167] Obtain the material description content and determine at least one input type corresponding to the material description content;
[0168] For each input type, determine the material description content for that input type; extract the type text corresponding to the type content;
[0169] Combine the type texts corresponding to each input type to generate the combined description text corresponding to the material description content;
[0170] Extract at least one attribute key-value pair from the combined description text; wherein the attribute key-value pair includes: attribute name and attribute value;
[0171] Based on at least one attribute key-value pair, match the standard material corresponding to the material description from the standard material library.
[0172] According to the technical solution of this invention, type text of material description content for each input type is extracted separately. The extracted type texts are combined to generate a combined description text. Then, at least one attribute key-value pair is extracted from the combined description text. Finally, using the attribute key-value pairs, the standard material corresponding to the material description content is matched from a standard material library. The solution of this invention can standardize material description content, reduce material identification errors, and ensure product quality and equipment safety for enterprises.
[0173] Furthermore, the combined description text is generated by combining the extracted type texts. This combined description text contains comprehensive material information. The attribute key-value pairs extracted from the combined description text can accurately and specifically describe the material's characteristics, further improving the effectiveness of material standardization.
[0174] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.
Claims
1. A method for standardizing materials, characterized in that, include: Obtain the material description content and determine at least one input type corresponding to the material description content; For each input type, determine the type content of the material description for that input type; extract the type text corresponding to the type content; Combine the type texts corresponding to each of the input types to generate the combined description text corresponding to the material description content; From the combined description text, at least one attribute key-value pair is extracted; wherein the attribute key-value pair includes: attribute name and attribute value; Based on the at least one attribute key-value pair, the standard material corresponding to the material description content is matched from the standard material library.
2. The method according to claim 1, characterized in that, Before determining at least one input type corresponding to the material description content, the method further includes: Retrieve the material text from the material description content; Several primary standard materials were identified from the standard material library; The semantic similarity between the material description and the material text of each first standard material is determined, and the first standard material with the highest semantic similarity is determined as the target standard material. In response to the semantic similarity of the target standard material being greater than a similarity threshold, the target standard material is determined as the standard material corresponding to the material description content; In response to the fact that the semantic similarity corresponding to the target standard material is not greater than a similarity threshold, a step of determining at least one input type corresponding to the material description content is performed.
3. The method according to claim 1, characterized in that, Extracting at least one attribute key-value pair from the combined description text includes: Identify the basic entities in the combined descriptive text; In response to the fact that the basic entity is an attribute entity, the attribute value corresponding to the basic entity is extracted from the combined description text; Combine the basic entity and its corresponding attribute values to generate attribute key-value pairs corresponding to the basic entity.
4. The method according to claim 1, characterized in that, After extracting at least one attribute key-value pair from the combined description text, the method further includes: Based on the combined description text, determine the standard category; Input the at least one attribute key-value pair and the standard category into the graph reasoning model to obtain at least one missing key-value pair; Merge the at least one attribute key-value pair and the at least one missing key-value pair to generate a supplementary key-value pair; The step of matching the standard material corresponding to the material description content from the standard material library based on the at least one attribute key-value pair includes: Based on the supplementary key-value pairs, the standard materials corresponding to the material description content are matched from the standard material library.
5. The method according to claim 1, characterized in that, The step of matching the standard material corresponding to the material description content from the standard material library based on the at least one attribute key-value pair includes: Based on the combined description text, determine the standard category corresponding to the material description content; Based on the at least one attribute key-value pair, the standard material corresponding to the material description content is matched from multiple standard materials under the standard category in the standard material library.
6. The method according to claim 1, characterized in that, The step of matching the standard material corresponding to the material description content from the standard material library based on the at least one attribute key-value pair includes: Multiple second standard materials were identified from the standard material library; For each of the second standard materials, at least one standard attribute pair corresponding to the second standard material is obtained from the standard material library; the matching degree corresponding to the second standard material is determined based on the at least one standard attribute pair and the at least one attribute key-value pair. The second standard material with the highest matching degree is determined as the standard material corresponding to the material description content.
7. A standardization device for materials, characterized in that, include: The content acquisition module is used to acquire material description content and determine at least one input type corresponding to the material description content; The text extraction module is used to determine the type content of the material description content for each input type; and to extract the type text corresponding to the type content. The text combination module is used to combine the type texts corresponding to each of the input types to generate the combined description text corresponding to the material description content. The key-value pair extraction module is used to extract at least one attribute key-value pair from the combined description text; wherein, the attribute key-value pair includes: attribute name and attribute value; The standardization module is used to match the standard material corresponding to the material description content from the standard material library based on the at least one attribute key-value pair.
8. An electronic device, characterized in that, include: One or more processors; Storage device for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-6.
9. A computer-readable medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1-6.
10. A computer program product comprising a computer program that, when executed by a processor, implements the method according to any one of claims 1-6.