Method, computing device, non-transitory computer-readable recording medium, and computer program product for tracing and verifying for construction inspection form

Abstract

The present application provides a method for tracing and verifying for a construction inspection form, which is performed by a computing device after loading and executing a computer program product. The method includes the following steps: receiving a construction inspection form, which is associated with a construction regulatory text and has a plurality of construction inspection items and a plurality of construction inspection criteria corresponding to the plurality of construction inspection items, respectively; inputting the construction inspection form into a embedding model; by the embedding model, performing word embedding vectorization on the contents of the construction inspection form, and then generating and outputting a construction inspection vectorization result corresponding to the construction inspection form; comparing the construction inspection vectorization result with a construction regulatory vectorization result stored in a construction regulatory vector database for similarity, and then searching a plurality of key content segments corresponding to a traceability verification item in the plurality of construction inspection items from the construction regulatory vector database; scoring and ranking the plurality of key content segments, and then extracting a best key paragraph from the plurality of key content segments; and comparing the contents of the best key paragraph with the contents of a traceability verification criterion in the plurality of construction inspection criteria for semantic consistency, and then generating the verified comparison result; wherein the traceability verification criterion corresponds to the traceability verification item, and the construction regulatory vectorization result is a result generated after performing word embedding vectorization on the content of the construction regulatory text by the embedding model. Thereby, the method can be applied to traceability verification of construction inspection forms, and solve the problem that inspectors are unable to effectively perform traceability verification. In addition, a computing device, a non-transitory computer-readable recording medium, and a computer program product for tracing and verifying for a construction inspection form are also provided.

Description

Technical Field

[0001] This application relates to a method for verification, a computer device, a non-transitory computer-readable recording medium, and a computer program product, and more particularly to a method for tracing and verifying construction inspection forms, a computer device, a non-transitory computer-readable recording medium, and a computer program product. Prior Technology

[0002] Firstly, for construction projects, in order to ensure the construction quality of each project, it is often necessary to conduct engineering inspections. More specifically, inspectors will conduct engineering inspections based on the contents of the inspection form to confirm that the construction project has indeed been carried out in accordance with the provisions of the construction specifications during the construction process, thereby ensuring the construction quality of the project.

[0003] Furthermore, before using inspection forms for engineering inspection, in practice it is often necessary to first check the contents of the inspection forms against the content specified in the construction specifications. This is to confirm whether the inspection items and criteria in the inspection forms are consistent with the content specified in the construction specifications, thereby ensuring that the inspectors can correctly conduct engineering inspections based on inspection forms whose contents are consistent with the construction specifications, and avoiding erroneous inspection results caused by inspection forms whose contents are inconsistent with the construction specifications. Summary of the Invention

[0004] Generally speaking, the verification of inspection forms usually requires manpower to carefully examine hundreds or thousands of pages of construction specifications. However, this method is not only time-consuming and laborious, but also difficult to implement in practice, which often leads to the failure to implement the verification of inspection forms in practice.

[0005] According to the statistical results of common deficiencies in the quality management system for public works construction inspection mentioned in "Exploring the Implementation Effectiveness of the Three-Tier Quality Management System in Public Works Construction from an Engineering Practice Perspective" (the statistics are from January 1, 2016 to June 30, 2020), the most common deficiency is "the self-inspection checklist is not implemented, or the inspection standards are not quantified, allowable error values ​​are not defined, or the inspection values ​​are not accurately recorded," with a deficiency rate as high as 84.01%. Meanwhile, the fifth and eighth ranked items in this statistical result are "the lack of established quality management standards for various materials / equipment and construction, or their non-compliance" (with a deficiency rate as high as 41.91%) and "the lack of established quality management standards for various sub-projects" (with a deficiency rate as high as 31.71%), respectively. These two are also related to the lack of established or non-compliance of inspection management standards. Therefore, the aforementioned statistical results reflect the current practical difficulty in implementing the inspection of checklists.

[0006] Furthermore, with the evolution of artificial intelligence, various industries and technological fields have introduced technologies such as generative artificial intelligence (GAI) in the hope of leveraging the high-speed computing capabilities of computer devices to solve practical problems caused by the need for human resources. However, taking large-scale language models as an example, although current large-scale language models possess powerful natural language processing capabilities, enabling them to parse text content in various technological fields and understand the contextual semantics and intent of the text to a certain extent, thereby generating corresponding results, current large-scale language models still have significant limitations. For instance, current large-scale language models can only acquire response capabilities based on the training data. This leads to a situation where, when faced with questions about new knowledge or specialized fields, large-scale language models often can only infer their perceived possible answers based on limited training data. These possible answers are often inaccurate or inappropriate, i.e., they generate inaccurate responses.

[0007] Therefore, how to solve the above-mentioned problems encountered by traditional inspection methods and effectively and accurately determine whether the content of the construction inspection form conforms to the content of the construction specification text has become an urgent problem that this technical field hopes to solve.

[0008] To address the aforementioned problems, this application provides a method for tracing and verifying construction inspection forms. This method is executed after a computer program is loaded and run on a computer device. The method includes the following steps: receiving a construction inspection form, wherein the construction inspection form is associated with a construction specification text, and the construction inspection form has a plurality of construction inspection items and a plurality of construction inspection standards corresponding to each of the plurality of construction inspection items; inputting the construction inspection form into a vectorization model; performing word embedding vectorization processing on the content of the construction inspection form through the vectorization model, and generating and outputting a construction inspection vectorization result corresponding to the construction inspection form; and comparing the construction inspection vectorization result with the construction specification vector database stored in a construction specification vector database. The similarity of the vectorized construction specification results is compared, and for a traceability inspection item among the plurality of construction inspection items, a plurality of key content fragments corresponding to the traceability inspection item are searched from the vector database of the construction specification. The plurality of key content fragments are scored and ranked, and an optimal key paragraph is extracted from the plurality of key content fragments. The semantic consistency of the content of the optimal key paragraph with the content of a traceability inspection standard among the plurality of construction inspection standards is compared, and an inspection comparison result is generated. The traceability inspection standard corresponds to the traceability inspection item. The vectorized result of the construction specification is the result generated after word embedding vectorization processing of the content of the construction specification text through the vectorization model.

[0009] In some embodiments, the vectorization result of the construction specification includes a plurality of vectorization results of specification fragments, and the plurality of vectorization results of specification fragments are generated through the following steps: receiving the construction specification text; segmenting the content of the construction specification text into a plurality of construction specification fragments according to a chapter arrangement structure; inputting the plurality of construction specification fragments into the vectorization model; and performing word embedding vectorization processing on the content of the plurality of construction specification fragments through the vectorization model, and generating and outputting a plurality of specification fragment vectorization results corresponding to each of the plurality of construction specification fragments.

[0010] In some embodiments, the plurality of key content segments are scored and sorted according to the traceability verification item and the traceability verification standard.

[0011] In some embodiments, the steps of scoring and ranking the plurality of key content segments include: extracting at least one independent factual information based on the content of the traceability verification item and the content of the traceability verification standard; determining whether the content of each of the plurality of key content segments satisfies the at least one independent factual information, and generating a plurality of scoring results corresponding to each of the plurality of key content segments; and ranking the plurality of key content segments based on the plurality of scoring results.

[0012] In some embodiments, the plurality of key content fragments are searched based on the traceability verification item and the traceability verification standard.

[0013] In some embodiments, the method further includes the following steps: when the verification result is a content discrepancy, correcting the content of the traceability verification standard based on the content of the best key paragraph, and generating a correction verification form.

[0014] In some embodiments, the method further includes the following steps: storing the traceability inspection item, the traceability inspection standard, and the best key paragraph as a historical inspection item, a historical inspection standard, and a historical best paragraph, respectively, in the construction specification vector database.

[0015] Furthermore, this application also provides a computer device for tracing and verifying construction inspection forms, comprising: a storage module configured to store a computer program product; and a processing module configured to be coupled to the storage module; wherein, after loading and executing the computer program product, the processing module is capable of executing any of the methods described in this application for tracing and verifying construction inspection forms.

[0016] Furthermore, this application also provides a non-transitory computer-readable recording medium for tracing and verifying construction inspection forms. After a computer device loads a computer program product stored in the non-transitory computer-readable recording medium and executes the computer program product, the computer device can execute any of the methods described in this application for tracing and verifying construction inspection forms.

[0017] Furthermore, this application also provides a computer program product for tracing and verifying construction inspection forms. After a computer device loads and executes the computer program product, the computer device can execute any of the methods described in this application for tracing and verifying construction inspection forms.

[0018] Therefore, the technical means provided in this application can achieve advantages that were previously unattainable. Specifically, one advantage achieved by this application is the ability to utilize artificial intelligence to trace and verify construction inspection forms, thereby effectively and accurately determining whether the content of the construction inspection forms conforms to construction specifications. This enables relevant personnel to implement the verification work of the inspection forms, and allows inspectors to correctly conduct project inspections based on construction inspection forms that conform to the content of the construction specifications, thereby ensuring the construction quality of each construction project. In other words, this application not only provides more effective assistance in project management and execution, but also improves the efficiency and accuracy of project supervision. Simple Explanation of the Diagram

[0019] Figure 1 is a block diagram illustrating the computer device used in this application for tracing and verifying construction inspection forms. Figure 2 is a flowchart illustrating the establishment of a construction specification vector database for traceability verification in this application. Figure 3 is a flowchart illustrating the method for tracing and verifying construction inspection forms in this application. Implementation

[0020] This application will be described in detail through the embodiments described below and the accompanying drawings, so as to help those skilled in the art to which this application pertains to understand the purpose, features and effects of this application.

[0021] It should be noted that the various steps described in this application may be performed sequentially, in reverse order, or by appropriately changing or skipping the order during control processing. It should be noted that "the first step may be performed after the second step" as described in this application may mean "the first step is performed directly after the second step is performed" and / or "other steps (e.g., the third step) are performed after the second step is performed, and then the first step is performed."

[0022] Furthermore, it should be noted that in the description of this application, terms such as "first," "second," and "third" are used to distinguish between elements, rather than to limit the elements themselves or indicate a specific order of elements. It should also be noted that in the description below, the same elements or steps may be represented by the same numbering.

[0023] Furthermore, the term "coupled" as described in this application can be interpreted as "directly connected" and / or "indirectly connected". Specifically, "the first element is configured to be coupled to the second element" can be interpreted as "the first element is configured to be directly connected to the second element" and / or "the first element is configured to be indirectly connected to the second element".

[0024] For the sake of brevity, although the steps in the method for tracing and verifying construction inspection forms described in this application are executed through a single computer device, in some embodiments, the steps may also be executed through multiple computer devices. That is, the steps in the method for tracing and verifying construction inspection forms described in this application may also be implemented through the collaborative operation of multiple computer devices (e.g., a computer device and a remote server).

[0025] Please refer to Figure 1, which is a block diagram illustrating the computer device 200 for tracing and verifying construction inspection forms in this application.

[0026] In some embodiments, the computer device 200 may be a finished product known to those skilled in the art to which this application pertains, specifically such as various models or specifications of desktop computers, notebook computers, laptop computers, tablet computers, or other electronic devices with equivalent configurations, but not limited thereto. The computer device 200 includes a processing module 220 and a storage module 230; additionally, in some embodiments, the computer device 200 further includes a receiving module 210, an output module 240, and / or a display module 250. Taking FIG1 as an example, the computer device 200 may include a receiving module 210, a processing module 220, a storage module 230, an output module 240, and a display module 250. The following will describe each module in more detail.

[0027] The receiving module 210 is configured to receive various data, images, and / or instructions from a remote server (not shown). For example, in this application, the receiving module 210 can receive construction specification texts and / or construction inspection forms from an electronic device (not shown) such as a scanner via a physical signal cable, so as to perform subsequent processing on the received construction specification texts and / or construction inspection forms. In some embodiments, the receiving module 210 can be a finished product known to those skilled in the art, such as various types or specifications of input / output interfaces, but is not limited thereto. In some embodiments, the file format of the construction specification texts and / or construction inspection forms received by the computer device 200 can be, for example, a portable document format (PDF), but is not limited thereto. Furthermore, the computer device 200 can also receive construction specification texts from a cloud database (not shown) or a cloud server (not shown) via virtual transmission.

[0028] Processing module 220 is configured to be coupled to storage module 230 and configured to perform the steps of any of the methods described in this application for tracing and verifying construction inspection forms. More specifically, after loading and executing a computer program product, processing module 220 is able to perform the steps of any of the methods described in this application for tracing and verifying construction inspection forms, thereby implementing any of the methods described in this application for tracing and verifying construction inspection forms. In some embodiments, processing module 220 may be a finished product known to those skilled in the art, such as various models or specifications of central processing units or graphics processors, but is not limited thereto.

[0029] Furthermore, the processing module 220 can be configured to be coupled to the receiving module 210, the output module 240 and the display module 250, so as to receive specific data, images and / or instructions via the receiving module 210, output specific data, images and / or instructions via the output module 240 and display specific data and / or images via the display module 250.

[0030] Storage module 230 is configured to store a computer program product, such that after processing module 220 loads and executes the stored computer program product, processing module 220 can execute the various steps of any of the methods described in this application for tracing and verifying construction inspection forms. The computer program product described in this application may include a series of code and / or instruction sets, particularly including specific code and / or instruction sets corresponding to the respective steps of any of the methods described in this application for tracing and verifying construction inspection forms.

[0031] In some embodiments, the storage module 230 may include one or more non-volatile memory and one or more volatile memory. In some embodiments, the volatile memory may be a finished product known to those skilled in the art to which this application pertains, such as various types of dynamic random access memory or static random access memory, but is not limited thereto. In some embodiments, the non-volatile memory may be a finished product known to those skilled in the art to which this application pertains, such as various types of read-only memory or flash memory, but is not limited thereto.

[0032] Output module 240 is configured to output various data, images, and / or instructions to devices such as remote servers (not shown), terminal devices (not shown), or cloud databases (not shown). Taking this application as an example, output module 240 can output the content of construction inspection forms, the inspection comparison results corresponding to the traceability inspection items and standards in the construction inspection forms, and / or the content of the best key paragraphs, but is not limited thereto. In some embodiments, output module 240 can be a finished product known to those skilled in the art to which this application pertains, such as various models or specifications of input / output interfaces, but is not limited thereto.

[0033] Display module 250 is configured to display specific data and / or images. Taking this application as an example, display module 250 can display to the user the contents of a construction inspection form, the inspection comparison results corresponding to the traceability inspection items and standards in the construction inspection form, and / or the content of the best key paragraphs, etc., but is not limited thereto. In some embodiments, display module 250 can be a finished product known to those skilled in the art to which this application pertains, such as various models or specifications of displays or display panels, but is not limited thereto.

[0034] Additionally, taking Figure 1 as an example, the computer device 200 also includes a construction specification vector database 410 to store the vectorized results of construction specifications, but is not limited thereto. The aforementioned vectorized results of construction specifications refer to the vectorized data generated after word embedding vectorization processing of the content of the construction specification text through the vectorization model described later. Furthermore, the content of the construction specification text corresponding to each vectorized result of construction specifications can also be stored in the construction specification vector database 410. In addition, metadata such as source files associated with the construction specification text, page numbers of each content, and the number of words from beginning to end can also be stored in the construction specification vector database 410. In some embodiments, the construction specification vector database 410 can be a finished product known to those skilled in the art to which this application pertains, such as Chroma DB, but is not limited thereto. Furthermore, the computer device 200 can also read various data stored in the construction specification vector database 410.

[0035] Furthermore, taking Figure 1 as an example, the computer device 200 also includes a vectorization model 310 and a language model 320. The following will provide a more detailed explanation of each model.

[0036] The vectorization model 310 is configured to convert string content in a file into corresponding vectorized results (i.e., high-dimensional vector representations) so that the computer device 200 can understand the string content in the file through the vectorized results. More specifically, taking this application as an example, after the construction specification text, construction specification fragments, and construction inspection forms are respectively input into the vectorization model 310, the vectorization model 310 can perform word embedding vectorization processing on the content of the construction specification text, the content of the construction specification fragments, and the content of the construction inspection forms, respectively, and then generate and output the corresponding vectorized results, namely, the construction specification vectorized result, the specification fragment vectorized result, and the construction inspection vectorized result. Thanks to the operational characteristics of the vectorization model 310, the computer device 200 can capture the semantic similarity and correlation between words in the file, thereby improving the accuracy and efficiency of the computer device 200 in understanding the string content in the file. In some embodiments, the vectorization model 310 may be a finished product known to those skilled in the art to which this application pertains, such as OpenAI's text-embedding-ada-002 or open-source Google-bert / bert-base-chinese or shibing624 / text2vec-base-chinese on Hugging Face, but is not limited thereto.

[0037] The language model 320 is configured to score and rank a plurality of key content segments (described later), and extract the best key paragraph from the plurality of key content segments. More specifically, in some embodiments, after the source tracing verification criteria (described later) and a plurality of key content segments corresponding to the source tracing verification criteria are input into the language model 320, the language model 320 can analyze the content of the source tracing verification criteria and the content of the plurality of key content segments, and then score and rank the plurality of key content segments according to the source tracing verification criteria, and extract the key content segment with the highest score as the best key paragraph from the plurality of key content segments; furthermore, in some embodiments, in the source tracing verification criteria (described later), the language model 320 can score and rank the plurality of key content segments according to the source tracing verification criteria, and extract the key content segment with the highest score as the best key paragraph; additionally, in some embodiments, in the source tracing verification criteria (described later), the language model 320 can score and rank the plurality of key content segments according to the source tracing verification criteria, and extract the key content segment with the highest score as the best key paragraph. After the verification items, the corresponding traceability verification standards, and a plurality of key content fragments corresponding to the traceability verification items are input into the language model 320, the language model 320 can analyze the content of the traceability verification items, the content of the traceability verification standards, and the content of the plurality of key content fragments. Then, based on the traceability verification items and the traceability verification standards, it scores and ranks the plurality of key content fragments, and extracts the highest-scoring key content fragment as the best key paragraph. In some embodiments, the language model 320 can be a finished product known to those skilled in the art, such as OpenAI's GPT series (GPT-4o, GPT-4o mini, o1, o3-min), Google's Gemini series (gemini-1.5-flash, gemin-2.0-flash), or Meta's Llama series (Llama 3.1, Llama 3.2, Llama 3.3), but is not limited thereto.

[0038] With the above configuration, the computer device 200 can execute the steps of any of the methods described in this application for tracing and verifying construction inspection forms, thereby implementing any of the methods described in this application for tracing and verifying construction inspection forms, and effectively and accurately determining whether the content of the construction inspection form conforms to the content of the construction specification text.

[0039] Please refer to Figure 2, which is a flowchart illustrating the construction of a construction specification vector database for traceability verification in this application. As shown in Figure 2, this flowchart may include steps S210, S220, S230, S240, and S250, and each step can be executed through the processing module 220 of the computer device 200 shown in Figure 1.

[0040] In some embodiments, step S220 may be executed after step S210, step S230 may be executed after step S220, step S240 may be executed after step S230, and step S250 may be executed after step S240.

[0041] In step S210, the construction specification text is received. More specifically, the computer device 200 can receive the construction specification text from an electronic device such as a scanner or a cloud database. The construction specification text may include various documents relating to the regulations of a construction project, such as construction contracts, but is not limited to these. Furthermore, the content of the construction specification text may specifically include the regulations for each construction step corresponding to various construction methods during the construction process, such as the timing of construction, construction methods, material quantity, material quality, material proportions, inspection methods and standards, and many other matters, but is not limited to these. In some embodiments, the file format of the construction specification text received by the computer device 200 may be, for example, a portable file format, but is not limited to this.

[0042] In step S220, the content of the construction specification text is divided into a plurality of construction specification segments. More specifically, the computer device 200 can divide the content of the construction specification text into a plurality of construction specification segments according to segmentation conditions such as the word length of each segment and the number of words to be repeated in the preceding and following segments. The aforementioned segmentation conditions may be, for example, chunk_size=150, chunk_overlap=15 or separators=["\n"], but are not limited to these. This will help to effectively and well preserve the content in the construction specification text.

[0043] Furthermore, in some embodiments, the computer device 200 can also segment the content of the construction specification text into multiple construction specification segments based on the segmentation conditions of the chapter arrangement structure. The aforementioned chapter arrangement structure can be, for example, chapters, sections, clauses, items, and clauses, but is not limited to these. In other words, the computer device 200 can first segment the content of each chapter based on the chapter arrangement structure, and then segment the content of each section within each chapter based on the section content, and so on. This segmentation method helps to preserve the complete content of each paragraph of the construction specification text, thereby improving the accuracy of the computer device 200 in understanding the content of the construction specification text.

[0044] In step S230, a plurality of construction specification fragments are input into a vectorization model. More specifically, the computer device 200 can input a plurality of construction specification fragments cut from the construction specification text into a vectorization model (e.g., OpenAI's text-embedding-ada-002 or open-source Google-bert / bert-base-chinese or shibing624 / text2vec-base-chinese on Hugging Face, but not limited to these), so that the vectorization model can perform word embedding vectorization processing on the input plurality of construction specification fragments respectively, i.e., step S240 described later.

[0045] In step S240, word embedding vectorization processing is performed on the content of a plurality of construction specification fragments, and a plurality of vectorized result fragments are generated and output. More specifically, after a plurality of construction specification fragments are input into the vectorization model, the vectorization model can perform word embedding vectorization processing on the content of each construction specification fragment separately, thereby generating and outputting corresponding results, that is, a plurality of vectorized result fragments, where each of the plurality of vectorized result fragments corresponds to the content of each construction specification fragment.

[0046] By leveraging the operational characteristics of vectorization models (converting each word into a high-dimensional vector representation), the computer device 200 is able to capture the semantic similarity and correlation between words in a plurality of input construction specification fragments, thereby improving the accuracy and efficiency of the computer device 200 in understanding the string content in a plurality of construction specification fragments; that is, the computer device 200 is better able to understand the content in the construction specification text.

[0047] In step S250, the vectorization results of the plurality of construction specifications are stored in a construction specification vector database. More specifically, the computer device 200 can store the vectorization results of each construction specification in a construction specification vector database (such as Chroma DB, but not limited thereto).

[0048] Through the aforementioned steps, the computer device 200 can construct a construction specification vector database. This database stores vectorized construction inspection results corresponding to the content of the construction specification text, or vectorized results of various specification segments corresponding to the segmented construction specification fragments. These vectorized results are then used for subsequent traceability verification of construction inspection forms associated with the specified construction specification. Simultaneously, these steps enable the construction of independent construction specification vector databases. Each database can store vectorized construction inspection results corresponding to a specific construction specification text. This allows for the use of independent databases to address various traceability verification needs for different construction specification texts, ensuring that all relevant specifications are stored without being limited by the language model's word count, thus effectively preventing the language model from responding to content beyond the database's scope. In other words, this helps to treat each construction specification vector database as an expert thoroughly familiar with the construction specification text, enabling the computer device 200 to utilize each database separately for traceability verification of associated construction inspection forms based on its specific verification needs.

[0049] Please refer to Figure 3, which is a flowchart illustrating the method for tracing and verifying construction inspection forms according to this application. The method for tracing and verifying construction inspection forms can be executed through the processing module 220 of the computer device 200 shown in Figure 1, and the method may include steps S310, S320, S330, S340, S350 and S360.

[0050] In some embodiments, step S320 may be executed after step S310, step S330 may be executed after step S320, step S340 may be executed after step S330, step S350 may be executed after step S340, and step S360 may be executed after step S350.

[0051] It should be noted that the method for tracing and verifying construction inspection forms, as shown in Figure 3, may include the steps shown in Figure 2. First, a construction specification vector database for tracing and verification is established by executing the steps shown in Figure 2. Then, the construction inspection forms are traced and verified by executing the steps shown in Figure 3. In other words, in some embodiments, step S310 may be executed immediately after step S250 as shown in Figure 2.

[0052] Furthermore, after establishing the construction specification vector database for traceability verification by executing the steps shown in Figure 2, the computer device 200 can repeatedly execute the steps shown in Figure 3 to perform traceability verification on multiple construction inspection forms. That is, it first executes steps S310 to S360 in the first round to perform traceability verification on the first construction inspection form, then executes steps S310 to S360 in the second round to perform traceability verification on the second construction inspection form, and so on. In other words, when the construction specification vectorization result generated after the content of the construction specification text has been processed by word embedding vectorization is stored in the construction specification vector database, the computer device 200 can perform traceability verification on the construction inspection forms associated with the aforementioned construction specification text by executing the steps shown in Figure 3.

[0053] In step S310, a construction inspection form is received. More specifically, the computer device 200 can receive the construction inspection form from an electronic device such as a scanner or a cloud database. The construction inspection form has a plurality of construction inspection items and a plurality of construction inspection standards, and each construction inspection standard corresponds to a specific construction inspection item; that is, a first construction inspection standard corresponds to a first construction inspection item, and a second construction inspection standard corresponds to a second construction inspection item, and so on. In some embodiments, the file format of the construction inspection form received by the computer device 200 may be, for example, a portable file format, but is not limited thereto.

[0054] Furthermore, since the content of the construction inspection form originates from the construction specification text, the construction inspection form that the computer device 200 needs to trace back to is associated with the construction specification text. That is, the computer device 200 can receive the first construction inspection form associated with the first construction specification text, and then trace the content of the first construction inspection form back to the first construction specification text for inspection.

[0055] Furthermore, during the traceability verification process, the computer device 200 can perform traceability verification on each construction inspection item and standard in the construction inspection form. For example, the computer device 200 can first use the first construction inspection item and the first construction inspection standard as the traceability verification inspection item and traceability verification standard described later, and then use the second construction inspection item and the second construction inspection standard as the traceability verification inspection item and traceability verification standard described later, and so on.

[0056] In step S320, the construction inspection form is input into the vectorization model. More specifically, the computer device 200 can input the received construction inspection form into the vectorization model (e.g., OpenAI's text-embedding-ada-002 or open-source Google-bert / bert-base-chinese or shibing624 / text2vec-base-chinese on Hugging Face, but not limited to these), so that the vectorization model can perform word embedding vectorization processing on the input construction inspection form, i.e., step S330 described later.

[0057] In step S330, word embedding vectorization processing is performed on the content of the construction inspection form, and the construction inspection vectorization result is generated and output. More specifically, after the construction inspection form is input into the vectorization model, the vectorization model can perform word embedding vectorization processing on the content of the construction inspection form, thereby generating and outputting the corresponding result, that is, the construction inspection vectorization result.

[0058] In some embodiments, the vectorization model can perform word embedding vectorization processing on the content of multiple construction inspection items and multiple construction inspection standards in the construction inspection form, and then generate and output corresponding results, namely, multiple inspection item vectorization results and multiple inspection standard vectorization results, wherein each of the multiple inspection item vectorization results corresponds to the content of each construction inspection item, and each of the multiple inspection standard vectorization results corresponds to the content of each construction inspection standard.

[0059] By leveraging the operational characteristics of the vectorization model (converting each word into a high-dimensional vector representation), the computer device 200 is able to capture the semantic similarity and correlation between words in the input construction inspection form, thereby improving the accuracy and efficiency of the computer device 200 in understanding the string content in the construction inspection form. In other words, the computer device 200 is better able to understand the content of each construction inspection item and each construction inspection standard in the construction inspection form.

[0060] Furthermore, since the construction specification text and the construction inspection form are both processed through word embedding vectorization using the same vectorization model, the computer device 200 can have similar or identical vectorization results for similarities or similarities between the content of the construction specification text and the content of the construction inspection form. That is, the computer device 200 can have similar or identical understanding of similar or identical content, which will help to ensure the accuracy of tracing the construction inspection form back to the construction specification text for verification.

[0061] In step S340, the similarity of the construction inspection vectorization result with the construction specification vectorization result stored in the construction specification vector database is compared, and a plurality of key content fragments are searched from the construction specification vector database. More specifically, the computer device 200 can compare the similarity of the construction inspection vectorization result with the construction specification vectorization result stored in the construction specification vector database, that is, perform a correlation search to calculate the comparison, and then calculate the key content fragments (Top K) with the shortest distance in the vector space, for example, the top 10 key content fragments with the shortest distance (Top 10). In some embodiments, the aforementioned similarity comparison can be performed in a manner known to those skilled in the art, such as Euclidean distance (also known as L2 distance), but is not limited to this. This application uses L2 distance as the criterion for judging the degree of similarity. In addition to Chroma DB itself supporting this correlation judgment standard, L2 distance better reflects the difference in position between the two in the vector space, rather than the difference in direction. Since the positional differences between the construction specification text and the construction inspection form can roughly reflect the correlation between the statements after they are converted into vector space representation, this application chooses L2 distance as the main criterion for calculating the degree of similarity.

[0062] In some embodiments, the computer device 200 can process traceability inspection items and / or traceability inspection standards in a construction inspection form, wherein the aforementioned traceability inspection items are selected from a plurality of construction inspection items in the construction inspection form, and the aforementioned traceability inspection standards are selected from a plurality of construction inspection standards in the construction inspection form, and the traceability inspection standards correspond to the traceability inspection items. Meanwhile, since the computer device 200 has performed word embedding vectorization processing on the content of the construction inspection form through step S330, this means that the vectorization model has performed word embedding vectorization processing on the traceability inspection items and / or traceability inspection standards respectively, and generated and output the corresponding results respectively, namely, the vectorization results of the traceability inspection items and / or the vectorization results of the traceability inspection standards. Therefore, the computer device 200 can also perform similarity comparison between the vectorization results of the traceability inspection items and / or the vectorization results of the traceability inspection standards and the vectorization results of the construction specifications stored in the construction specification vector database through step S340, and then calculate the key content segments with shorter distances in the vector space for each piece of data in the construction inspection form.

[0063] Taking traceability inspection items as an example, the computer device 200 can compare the similarity between the vectorized results of traceability inspection items and / or the vectorized results of traceability inspection standards with the vectorized results of various construction specifications stored in the construction specification vector database. This allows the computer device to calculate a plurality of construction specification vectorized results that are close in distance in the vector space for the traceability inspection items in the construction inspection form. Then, it can search for the corresponding plurality of construction specification contents in the construction specification vector database and use these construction specification contents as a plurality of key content fragments corresponding to the traceability inspection items.

[0064] Alternatively, in some embodiments, the computer device 200 can also treat the traceability inspection items and standards in the construction inspection form as a traceability inspection object. It can then generate and output the vectorized result of the traceability inspection object corresponding to the traceability inspection object through a vectorization model. The similarity of the vectorized result of the traceability inspection object is compared with the vectorized result of the construction specifications stored in the construction specification vector database. Furthermore, for each traceability inspection object in the construction inspection form, the key content fragments with the shortest distance in the vector space are calculated. In this approach, the computer device 200 can more accurately grasp the content described by each traceability inspection object and ensure the correlation between the content of each traceability inspection object and the content of the construction specification text.

[0065] Subsequently, after calculating the similarity comparison results, the computer device 200 can search the construction specification vector database for a plurality of key content fragments (Top K) corresponding to the traceability inspection item. That is, the construction specification vector database of this application can also be used to store the content of the construction specification text. This allows the computer device 200 to directly read the corresponding text content from the construction specification vector database using the similarity comparison results. These key content fragments are considered to be the content in the construction specification text most relevant to the traceability inspection item, and these key content fragments are highly likely to be the design source of this traceability inspection item. In this way, the source of each traceability inspection item can be successfully traced, thereby completing the traceability inspection of the construction inspection form. For example, the computer device 200 can search for the top 10 key content fragments (Top 10) corresponding to a certain traceability inspection item.

[0066] In step S350, the plurality of key content segments are scored and ranked, and the best key paragraph is extracted from the plurality of key content segments. More specifically, the computer device 200 can use traceability verification standards to score and rank the plurality of key content segments corresponding to each traceability verification standard, and then extract the best key paragraph from each of the plurality of key content segments. For example, the computer device 200 can use a certain traceability verification standard to score and rank the plurality of key content segments corresponding to this traceability verification standard, and then extract the best key paragraph corresponding to this traceability verification standard from the plurality of key content segments, thereby finding the most relevant content in the construction specification text. In some embodiments, the aforementioned scoring and ranking can be implemented in a manner known to those skilled in the art to which this application pertains, such as through prompt engineering by incorporating Chain of Thought (CoT) or Few-Shot Learning paradigms, but is not limited thereto. Furthermore, in some embodiments, those skilled in the art to which this application pertains may score and rank each key content segment based on whether it is supported by a single independent factual information or whether it is supported by multiple independent factual information.

[0067] Furthermore, in some embodiments, the computer device 200 can use corresponding traceability inspection items and traceability inspection standards to score and rank a plurality of key content segments corresponding to the traceability inspection items and traceability inspection standards, and then extract the best key paragraphs corresponding to the traceability inspection items and traceability inspection standards from the plurality of key content segments, thereby finding the most relevant content in the construction specification text. More specifically, the computer device 200 can teach a large language model (such as OpenAI's GPT series, Google's Gemini series, or Meta's Llama series, but not limited to these) to break down the traceability inspection items and criteria into multiple independent sub-conditions using prompting words. It then scores whether each sub-condition is satisfied by a plurality of key content segments (Top K), and averages the scores of each sub-condition to represent the score of each key content segment. Based on this score, the plurality of key content segments are reordered to find the highest-scoring key content segment, which is then selected as the optimal key segment. This approach helps improve the accuracy of the computer device 200's traceability inspection of the construction inspection form's content.

[0068] For example, corresponding traceability check items and traceability check criteria are input into a large language model. The large language model first performs factual breakdown on the traceability check items and traceability check criteria that need to be met, and determines the number of facts that need to be met (for example, if both size and casting time are mentioned in a traceability check item, this traceability check item will be broken down into two independent factual information). This allows at least one independent factual information to be extracted based on the content of the traceability check items and the content of the traceability check criteria. Subsequently, after multiple key content fragments are input into the large language model, the large language model will infer whether the content of each key content fragment meets the requirements of the aforementioned independent factual information, and score it according to the degree of satisfaction. Then, these key content fragments are sorted according to the aforementioned scoring results.

[0069] Furthermore, taking the traceability and verification standard of "the temperature of concrete pouring should be greater than 13 degrees Celsius and less than 32 degrees Celsius" as an example, the large-scale language model can break down the information into three independent facts (i.e., "temperature control is required when pouring concrete," "the temperature of concrete pouring should be kept above 13 degrees Celsius," and "the temperature of concrete pouring should not exceed 32 degrees Celsius") by learning from thought chains or a small number of sample examples. Also, taking the first key content segment as an example, the large-scale language model will analyze the content of the first key content segment and determine whether its semantics satisfies these three independent facts. Assuming that the semantics of the first key content segment only satisfies two of them, the score for the first key content segment would be, for example, 67 points (i.e.: ), and so on.

[0070] It should be noted that although this application uses a large language model as an example for illustration, this application is not limited to being implemented with a large language model. That is, this application can also be implemented using other language models such as small language models.

[0071] In step S360, the content of the best key paragraph is semantically compared with the content of the traceability verification standards among the plurality of construction inspection standards, and a verification comparison result is generated. More specifically, the computer device 200 can perform semantic consistency comparison between the content of the corresponding best key paragraphs and the content of the traceability verification standards, thereby generating a verification comparison result. The verification comparison result can be content matching information or content non-matching information, and the content matching information or content non-matching information can be represented by flag bits, for example, a high logic level ("1") indicates that the content of the traceability verification standard conforms to the content of the construction specification text, while a low logic level ("0") indicates that the content of the traceability verification standard does not conform to the content of the construction specification text, but is not limited thereto. In some embodiments, when the semantics of the traceability verification standard fully conforms to the requirements of the construction specification text, the computer device 200 can generate a verification comparison result of content matching information, but is not limited thereto. By comparing semantic consistency, the computer device 200 can identify whether there are semantic and / or numerical discrepancies between the content of the best key paragraph and the content of the source verification standard. In some embodiments, the aforementioned semantic consistency comparison can be performed in a manner known to those skilled in the art, such as teaching a large language model to perform semantic consistency comparison using thought chains, small sample examples, or prompting engineering, or training a small language model, such as Bidirectional Encoder Representations from Transformers (BERT), to judge semantic consistency, but is not limited to these methods.

[0072] Taking large-scale language modeling technology as an example, it teaches large-scale language models to make inferences through thought chains, cue engineering, and a small number of sample examples. The inference results ultimately determine whether the traceability inspection standards in the construction inspection form meet the requirements of the construction specification text. If the semantics are completely consistent, a content consistency judgment is given, and a content consistency information check and comparison result is generated; if any content is inconsistent, a content inconsistency judgment is given, and a content inconsistency information check and comparison result is generated.

[0073] Furthermore, taking a small language model (such as BERT) as an example, during the training process, special tokens can be added to represent the correlation between two statements using Boolean values ​​(0 / 1), thereby helping the model learn and judge their degree of consistency. The model can also be trained using labeled data. In the labeled data, the traceability inspection standards in the construction inspection form must fully comply with the requirements of the construction specification text. Only under these circumstances will it be considered a positive sample.

[0074] Through the operations described above, the computer device 200 can perform traceability checks on the contents of the construction inspection form, thereby effectively and accurately determining whether the contents of the construction inspection form conform to the contents of the associated construction specification text. This enables relevant personnel to utilize artificial intelligence to implement the inspection work of the construction inspection form, and allows inspectors to correctly conduct project inspections based on the construction inspection form that conforms to the contents of the construction specification text, thereby ensuring the construction quality of each construction project.

[0075] It should be noted that the method described in this application for tracing and verifying construction inspection forms can be used for batch processing of multiple construction inspection forms, that is, tracing and verifying each construction inspection form separately.

[0076] Furthermore, in some embodiments, the method shown in FIG3 further includes the following step: storing the traceability inspection items, traceability inspection standards, and best key paragraphs as historical inspection items, historical inspection standards, and historical best paragraphs respectively in a construction specification vector database. This step can also be performed after, for example, step S360 as shown in FIG3. This step helps to examine the accuracy and usability of the method for traceability inspection of construction inspection forms described in this application using the characteristics of the database, and allows the computer device 200 to use the data in the database to determine whether there are any anomalies in the results generated after each traceability inspection.

[0077] Furthermore, in some embodiments, the method shown in FIG3 further includes the following steps: when the verification result is content discrepancy information, the content of the traceability verification standard is corrected according to the content of the best key paragraph, and a correction verification form is generated. This step can be performed after, for example, step S360 as shown in FIG3. More specifically, when the verification result is content discrepancy information, this indicates that the content of the traceability verification standard in the construction verification form does not conform to the content in the construction specification text. The computer device 200 can correct the non-conforming content in the traceability verification standard, that is, the content of the best key paragraph overwrites the non-conforming content in the traceability verification standard, thereby generating a new construction verification form, and using the new construction verification form as the correction verification form. This step utilizes accurate information from the construction specifications to correct any discrepancies in the construction inspection forms. This facilitates the provision of corrected construction inspection forms to inspectors after traceability verification, thereby improving user convenience and enabling inspectors to conduct accurate project inspections based on construction inspection forms that conform to the construction specifications, thus ensuring the construction quality of each project.

[0078] In some embodiments, those skilled in the art to which this application pertains may use integrated development environments (IDEs) such as Jupyter Notebook, Visual Studio Code, PyCharm, and Sublime Text; programming languages ​​such as Python and PyTorch; and libraries and toolkits such as chroma-hnswlib, chromadb, chromadbx, langchain, langchain_community, neo4j, openai, sentence-transformers, tokenizers, and transformers to implement any of the methods described in this application for tracing and verifying construction inspection forms.

[0079] Please refer to Table (I), which illustrates the results of verifying the construction inspection forms by tracing them back to the construction specification text using the method of this application for tracing back to the source of the construction inspection forms. Table (1)

[0080] Please refer to Table (II), which illustrates the results of scoring and ranking multiple key content segments in this application. Table (II)

[0081] Please refer to Table (III), which illustrates the results of verifying the construction inspection forms by tracing them back to the construction specification text using the method of this application for tracing back to the construction inspection forms. Table (III)

[0082] In some embodiments, the steps in the method for tracing and verifying construction inspection forms described in this application may be further combined, replaced, repeatedly performed, and / or modified to produce new embodiments without departing from the scope disclosed in this application.

[0083] In some embodiments, the steps of the method for tracing and verifying construction inspection forms described in this application can be stored in a non-transitory computer-readable recording medium, such as a hard disk, optical disk, magnetic disk, USB flash drive, or a database accessible via a network, but not limited thereto. After the non-transitory computer-readable recording medium loads a computer program product into memory through a computer device and executes the computer program product, the computer device is able to implement any of the methods described in this application for tracing and verifying construction inspection forms.

[0084] In some embodiments, the computer program product for tracing and verifying construction inspection forms described in this application may include a series of code and / or instruction sets, particularly specific code and / or instruction sets corresponding to each step of any of the methods for tracing and verifying construction inspection forms described in this application, so that after a computer device loads and executes the computer program product, the computer device can implement any of the methods for tracing and verifying construction inspection forms described in this application.

[0085] This application has been further described through the above embodiments and accompanying drawings. However, those skilled in the art to which this application pertains can still make many modifications and variations without departing from the scope and spirit set forth in the claims of this application. Therefore, the scope of protection of this application should still be determined by the claims of the patent applications and should not be limited by the content disclosed in the specification.

[0086] 200: Computer device 210: Receiver Module 220: Processing Module 230: Storage Module 240: Output Module 250: Display Module 310: Vectorized Model 320: Language Model 410: Construction Specification Vector Database S210, S220, S230: Steps S240, S250: Steps S310, S320, S330: Steps S340, S350, S360: Steps

Claims

1. A method for tracing and verifying a construction inspection form, the method being executed after loading and running a computer program product via a computer device, the method comprising the following steps: receiving a construction inspection form, wherein the construction inspection form is associated with a construction specification text, and the construction inspection form has a plurality of construction inspection items and a plurality of construction inspection standards corresponding to the plurality of construction inspection items; inputting the construction inspection form into a vectorization model; performing word embedding vectorization processing on the content of the construction inspection form through the vectorization model, and generating and outputting a construction inspection vectorization result corresponding to the construction inspection form; comparing the similarity of the construction inspection vectorization result with the construction specification vectorization result stored in a construction specification vector database, and searching the construction specification vector database for a traceability verification inspection item among the plurality of construction inspection items to retrieve a plurality of key content fragments corresponding to the traceability verification inspection item; The process involves scoring and ranking the multiple key content segments, extracting the best key paragraph from these segments, and comparing the semantic consistency of the best key paragraph with the content of one traceability inspection standard among the multiple construction inspection standards, generating an inspection comparison result. The traceability inspection standard corresponds to the traceability inspection item; and the vectorization result of the construction specification is the result generated after word embedding vectorization processing of the content of the construction specification text through the vectorization model.

2. The method as described in request item 1, wherein, The vectorization result of the construction specification includes a plurality of vectorized specification fragments, which are generated through the following steps: receiving the construction specification text; dividing the content of the construction specification text into a plurality of construction specification fragments according to a chapter arrangement structure; inputting the plurality of construction specification fragments into the vectorization model; and through the vectorization model, performing word embedding vectorization processing on the content of the plurality of construction specification fragments, and generating and outputting a plurality of vectorized specification fragments corresponding to each of the plurality of construction specification fragments.

3. The method as described in request item 1, wherein, These multiple key content segments are scored and ranked based on the traceability inspection and verification items and the traceability inspection and verification standards.

4. The method as described in request item 3, wherein, The steps for scoring and ranking the plurality of key content segments include: extracting at least one independent factual information based on the content of the traceability inspection item and the content of the traceability inspection standard; determining whether the content of each of the plurality of key content segments satisfies the at least one independent factual information, and generating a plurality of scoring results corresponding to each of the plurality of key content segments; and ranking the plurality of key content segments based on the plurality of scoring results.

5. The method as described in request item 1, wherein, The multiple key content segments were searched based on the traceability verification item and the traceability verification standard.

6. The method as described in request item 1, wherein, The method also includes the following steps: When the check comparison result is a content discrepancy, the content of the traceability check standard is corrected according to the content of the best key paragraph, and a correction check form is generated.

7. The method as described in request item 1, wherein, The method also includes the following steps: storing the traceability inspection item, the traceability inspection standard, and the best key paragraph as a historical inspection item, a historical inspection standard, and a historical best paragraph in the construction specification vector database, respectively.

8. A computer device for tracing and verifying construction inspection forms, comprising: a storage module configured to store a computer program product; and a processing module configured to be coupled to the storage module; wherein, After loading and executing the computer program product, the processing module is able to perform the method for tracing and verifying construction inspection forms as described in any of requests 1 to 7.

9. A non-transitory computer-readable recording medium for tracing and verifying construction inspection forms, wherein after a computer device loads and executes a computer program product stored in the non-transitory computer-readable recording medium, the computer device is capable of performing the method for tracing and verifying construction inspection forms as described in any one of claims 1 to 7.

10. A computer program product for tracing and verifying construction inspection forms, wherein after the computer program product is loaded and executed on a computer device, the computer device is able to perform the method for tracing and verifying construction inspection forms as described in any one of claims 1 to 7.

Images