Product specification examination method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a product specification examination method and device, electronic equipment and a storage medium. The method includes: acquiring a product specification; determining the templating degree of the product specification according to the content of the product specification; determining a target agent adaptive to the product specification from a plurality of preset agents according to the templating degree of the product specification; wherein different preset intelligent agents are used for examining product specifications with different templating degrees; and the product specification is examined through the target agent, and an examination result is obtained. According to the template degree of the specification content, different agents are dynamically allocated for review, limitation of fixed rules and a single process in a traditional method is avoided, the review efficiency is remarkably improved, the review time is shortened, the requirement for manual intervention is reduced, the targeted agents are adopted for review according to the characteristics of different types of content, and the review efficiency is improved. The flexibility and accuracy of examination are enhanced, and the possibility of misjudgment and missed judgment is reduced.

Description

Technical Field

[0001] This invention relates to the field of intelligent document review technology, and in particular to a method for reviewing product manuals, a device for reviewing product manuals, electronic devices, and computer-readable storage media. Background Technology

[0002] In modern industrial and information-based society, product manuals serve as a crucial bridge connecting manufacturers and users. Their accuracy, completeness, and standardization directly impact product safety, user experience, and after-sales service efficiency. Especially in high-risk or high-tech industries such as medical equipment, aerospace, intelligent manufacturing, and consumer electronics, the quality of manuals not only determines whether users can operate the equipment correctly but may also involve important issues such as legal compliance, intellectual property protection, and product liability determination.

[0003] Traditional manual review primarily relies on manual processes, with a review team consisting of technical document writers, legal counsel, and product engineers meticulously checking each item. While this method ensures accuracy to some extent, it suffers from low efficiency, high subjectivity, susceptibility to omissions, and difficulty in standardization. The limitations of manual review become even more pronounced when the manual is published in multiple languages, versions, and on multiple platforms.

[0004] Currently, the review of documents such as instruction manuals involves manual review or automated systems based on fixed rules. These systems match and check the content against preset rules and templates to ensure compliance. However, this method relies on fixed rules and templates, making it difficult to adapt to the diversity and complexity of instruction manual content. Summary of the Invention

[0005] In view of the above problems, embodiments of the present invention are proposed to provide a method for reviewing a product manual, a device for reviewing a product manual, an electronic device, and a corresponding computer-readable storage medium to overcome or at least partially solve the above problems.

[0006] To address the aforementioned problems, embodiments of the present invention disclose a method for reviewing product manuals, comprising:

[0007] Obtain the product manual;

[0008] Based on the content of the product manual, determine the degree of template-based nature of the product manual;

[0009] Based on the degree of template-ification of the product manual, a target intelligent agent adapted to the product manual is determined from multiple preset intelligent agents; wherein different preset intelligent agents are used to review product manuals with different degrees of template-ification.

[0010] The target intelligent agent reviews the product manual and obtains the review results.

[0011] Optionally, determining the degree of template-based nature of the product manual based on its content includes:

[0012] Identify the keywords in the product manual and determine the frequency of occurrence of the keywords;

[0013] The degree of template-based nature of the product manual is determined based on the keywords and their corresponding frequencies of occurrence.

[0014] Optionally, determining the degree of template-based nature of the product manual based on its content includes:

[0015] Determine the degree of match between the content of the product instruction manual and the preset template;

[0016] The degree of template-ification of the product manual is determined based on the degree of matching.

[0017] Optionally, the preset intelligent agent includes a first type of intelligent agent, and the step of reviewing the product manual through the target intelligent agent to obtain the review result includes:

[0018] If the target intelligent agent is the first type of intelligent agent, then the content of the product manual is compared with the preset template through the first type of intelligent agent to obtain the review result.

[0019] Optionally, the preset intelligent agent includes a second type of intelligent agent, and the step of reviewing the product manual through the target intelligent agent to obtain the review result includes:

[0020] If the target intelligent agent is the second type of intelligent agent, then the content of the product manual is compared with the preset template by the second type of intelligent agent; and semantic analysis is performed on the content of the product manual to obtain the semantic analysis result, and it is determined whether the semantic analysis result meets the preset semantic requirements to obtain the review result.

[0021] Optionally, the preset intelligent agent includes a third type of intelligent agent, and the step of reviewing the product manual through the target intelligent agent to obtain the review result includes:

[0022] If the target intelligent agent is the third type of intelligent agent, then the third type of intelligent agent calls the large language model to determine whether the content of the product manual meets the preset review requirements, so as to obtain the review result; the preset review requirements include at least one of the following: reasonable content, complete coverage, clear description, and logical content.

[0023] Optionally, it also includes:

[0024] If the review results indicate that the product manual contains abnormal or uncertain content, then the target intelligent agent will conduct a collaborative review with other intelligent agents.

[0025] Optionally, determining the target intelligent agent adapted to the product manual from multiple preset intelligent agents based on the degree of template-ification of the product manual includes:

[0026] Based on the product manual's template level being the first level, the target intelligent agent adapted to the product manual is determined as the first type of intelligent agent from multiple preset intelligent agents;

[0027] Based on the second degree of template-based nature of the product manual, a target intelligent agent adapted to the product manual is determined from multiple preset intelligent agents as a second type of intelligent agent; wherein, the second degree is greater than the first degree;

[0028] Based on the product manual's template level being third, a target intelligent agent adapted to the product manual is determined from multiple preset intelligent agents as a third type of intelligent agent; wherein, the third level is greater than the second level.

[0029] A second aspect of the present invention discloses a product instruction manual review apparatus, comprising:

[0030] The acquisition module is used to acquire product manuals;

[0031] The template-level determination module is used to determine the template-level of the product manual based on its content.

[0032] The agent determination module is used to determine the target agent that is compatible with the product manual from multiple preset agents based on the degree of template-ification of the product manual; wherein different preset agents are used to review product manuals with different degrees of template-ification.

[0033] The review module is used to review the product manual through the target intelligent agent and obtain the review result.

[0034] Optionally, the templated degree determination module includes:

[0035] The keyword determination submodule is used to determine the keywords in the content of the product manual and to determine the frequency of occurrence of the keywords;

[0036] The first template-level determination submodule is used to determine the template level of the product manual based on the keywords and their corresponding frequencies of occurrence.

[0037] Optionally, the templated degree determination module includes:

[0038] The matching degree determination submodule is used to determine the degree of matching between the content of the product instruction manual and the preset template;

[0039] The first templated degree determination submodule is used to determine the templated degree of the product manual based on the matching degree.

[0040] Optionally, the preset intelligent agent includes a first type of intelligent agent, and the review module includes:

[0041] The first review submodule is used to compare the content of the product manual with a preset template through the first type of intelligent agent if the target intelligent agent is the first type of intelligent agent, and obtain the review result.

[0042] Optionally, the preset intelligent agent includes a second type of intelligent agent, and the step of reviewing the product manual through the target intelligent agent to obtain the review result includes:

[0043] The second review submodule is used to compare the content of the product manual with a preset template through the second type of intelligent agent if the target intelligent agent is the second type of intelligent agent; and to perform semantic analysis on the content of the product manual to obtain the semantic analysis result, and determine whether the semantic analysis result meets the preset semantic requirements in order to obtain the review result.

[0044] Optionally, the preset intelligent agent includes a third type of intelligent agent, and the step of reviewing the product manual through the target intelligent agent to obtain the review result includes:

[0045] The third review submodule is used to determine whether the content of the product manual meets the preset review requirements by calling the large language model through the third type of intelligent agent if the target intelligent agent is the third type of intelligent agent, so as to obtain the review result; the preset review requirements include at least one of the following: reasonable content, complete coverage, clear description, and logical content.

[0046] Optionally, it also includes:

[0047] The collaborative review module is used to conduct collaborative review with other intelligent agents through the target intelligent agent if the review result indicates that the product manual contains abnormal or uncertain content.

[0048] Optionally, the agent determination module includes:

[0049] The first intelligent agent determination submodule is used to determine the target intelligent agent that is compatible with the product manual as a first type intelligent agent from multiple preset intelligent agents, based on the degree of template-ification of the product manual as the first degree.

[0050] The second agent determination submodule is used to determine, from multiple preset agents, a target agent that is compatible with the product manual as a second type of agent, based on the degree of template-ification of the product manual as a second degree; wherein, the second degree is greater than the first degree;

[0051] The third agent determination submodule is used to determine the target agent that is compatible with the product manual as a third type of agent from multiple preset agents, based on the templated degree of the product manual being a third degree; wherein the third degree is greater than the second degree.

[0052] A third aspect of the present invention discloses an electronic device, comprising: a processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein the computer program, when executed by the processor, implements the steps of the product specification review method as described above.

[0053] A fourth aspect of the present invention discloses a computer-readable storage medium storing a computer program thereon, wherein the computer program, when executed by a processor, implements the steps of the product specification review method as described above.

[0054] The embodiments of the present invention have the following advantages:

[0055] This invention introduces a method for reviewing product manuals, comprising: acquiring the product manual; determining the degree of template-based nature of the product manual based on its content; determining a target intelligent agent adapted to the product manual from multiple preset intelligent agents based on the degree of template-based nature of the product manual; wherein different preset intelligent agents are used to review product manuals with different degrees of template-based nature; and reviewing the product manual through the target intelligent agent to obtain a review result. Dynamically assigning different intelligent agents for review based on the degree of template-based nature of the manual content avoids the limitations of fixed rules and single processes in traditional methods, significantly improving review efficiency, shortening review time, reducing the need for manual intervention, and employing targeted intelligent agents for review based on the characteristics of different types of content, enhancing the flexibility and accuracy of the review, and reducing the possibility of misjudgments and omissions. Attached Figure Description

[0056] Figure 1 This is a flowchart illustrating the steps of a product manual review method provided in an embodiment of the present invention;

[0057] Figure 2This is a flowchart of another product specification review method provided by an embodiment of the present invention;

[0058] Figure 3 This is a system architecture diagram of a product manual review method provided in an embodiment of the present invention;

[0059] Figure 4 This is a structural block diagram of a product manual review device provided in an embodiment of the present invention. Detailed Implementation

[0060] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0061] Existing solutions rely on fixed rules and templates to review instruction manual content. However, the lack of intelligent collaboration and dynamic allocation mechanisms leads to low review efficiency, making it difficult to meet the needs of efficient review. Furthermore, existing solutions use single rules or template matching technology to review instruction manual content. This lack of flexibility in adapting to changes in the content results in insufficient review accuracy, making it difficult to meet the needs of high-precision review.

[0062] One of the core concepts of this invention is that different intelligent agents are dynamically assigned to review the content based on the degree of template-based nature of the specification. Based on the characteristics of different types of content, targeted intelligent agent algorithm technology is adopted, which significantly improves review efficiency, shortens review time, reduces the need for manual intervention, enhances the flexibility and accuracy of review, and reduces the possibility of misjudgment and omission.

[0063] Reference Figure 1 The diagram illustrates a flowchart of a product manual review method according to an embodiment of the present invention. The method may specifically include the following steps:

[0064] Step 101: Obtain the product instruction manual;

[0065] Reviewing instructions such as product manuals and technical documents is a crucial step in ensuring their accuracy, compliance, and usability. Its main functions include:

[0066] Ensure the accuracy and completeness of information, checking the accuracy of technical parameters, operating procedures, and safety warnings to prevent user misoperation or product damage due to incorrect information; identify missing key content (such as compatibility instructions, warranty terms, etc.) to avoid legal or usage risks. Compliance and legal risk control: ensure that the instruction manual complies with industry standards, regional laws and regulations, or mandatory provisions in specific fields (such as medical and children's products); avoid user complaints or legal disputes caused by misleading descriptions or missing safety warnings. Improve user experience: check the clarity of language and the coherence of logic, avoiding the misuse of technical terms and ensuring that ordinary users can understand; for international products, review the accuracy of translations to avoid cultural ambiguity. Maintain brand image: high-quality instruction manuals enhance user trust in the brand; clear instructions reduce the probability of user inquiries or misuse, saving customer service resources. Safety risk control: high-risk operations (such as electrical installation and chemical use) require prominent safety warnings to avoid personal injury; in the event of an accident, the reviewed instruction manual can serve as evidence that the company has fulfilled its obligation to inform. Adaptability and consistency: Ensure that the functions and accessories described in the instruction manual are consistent with the actual product (especially for iteratively updated products); avoid contradictory statements with other documents (such as advertisements, official websites).

[0067] Intelligent document review technology refers to the use of artificial intelligence (AI), machine learning (ML), and other advanced technologies to automatically analyze, understand, and process various types of document content in order to quickly and accurately identify key information, errors, or specific sections requiring attention. This technology is widely used in various fields such as legal document review, financial report auditing, medical record management, and compliance checks.

[0068] Intelligent document review technology typically includes the following functions: text extraction, which extracts textual information from documents of different formats; natural language processing, which understands the content and semantics of the text and performs operations such as information classification and entity recognition; pattern recognition, which identifies recurring patterns or anomalies in documents; and automated decision support, which makes judgments and suggestions based on predefined rules or by learning from historical data.

[0069] In this embodiment of the invention, the product manual is first obtained. The method of obtaining the manual can be by user input, scanning, uploading documents, etc. The product manual document to be reviewed can be in PDF, Word, HTML, or other formats. This invention does not limit the format.

[0070] Step 102: Determine the degree of template-based nature of the product manual based on its content;

[0071] The degree of template standardization in instruction manuals directly impacts the efficiency, accuracy, and compliance of the review process, with different levels of template standardization leading to significant differences in the review outcome. For instruction manuals with a high degree of template standardization, fixed chapter structures and standardized terminology reduce the time spent on manual item-by-item checks; automated review tools (such as AI verification) can quickly match preset rules (such as drug instruction manual templates); mandatory embedding of regulatory requirements (such as the "contraindications" section of a drug) can reduce the risk of missed checks; standardized safety warning icons and expressions avoid legal disputes; and templated content (such as standard terminology and fixed chart numbering) can reduce basic issues such as spelling errors and logical inconsistencies. For instruction manuals with a low degree of template standardization, the review relies on manual sentence-by-sentence analysis, which is time-consuming and prone to omissions; it also relies on the reviewer's personal experience, resulting in higher compliance risks.

[0072] However, high levels of template-based documentation also present several problems. Insufficient flexibility leads to difficulties in adaptation, and highly templated manuals may not be suitable for innovative products or special scenarios. There is also the risk of review fatigue, as repetitive review of fixed content may cause reviewers to overlook subtle changes, such as modifications to key parameters in version updates. Furthermore, over-reliance on automated tools can be problematic, as AI reviewers may fail to recognize semantic issues outside the template, such as ambiguous expressions. Therefore, manuals with varying degrees of template-based documentation each have their advantages and disadvantages, and the review methods also differ.

[0073] In some embodiments provided by this invention, the corresponding degree of templating is determined according to the content of the specification, and corresponding content type tags for the degree of templating are generated, such as high templating, medium templating, and no templating. Further review can improve the accuracy of the review and reduce missed detections and false detections.

[0074] Step 103: Based on the degree of template-ification of the product manual, determine the target intelligent agent that is compatible with the product manual from multiple preset intelligent agents; wherein different preset intelligent agents are used to review product manuals with different degrees of template-ification.

[0075] An intelligent agent is an autonomous software or hardware entity that can perceive its environment and influence it by performing actions to achieve specific goals. Intelligent agents can be simple programs or highly complex systems, and are widely used in various fields such as the Internet of Things, smart homes, e-commerce, game development, and automation control.

[0076] The main characteristics of intelligent agents include: autonomy, meaning that intelligent agents can operate and make decisions independently without direct human intervention; responsiveness, meaning that they can respond promptly to changes in the environment and adjust their behavior accordingly; sociality, meaning that some intelligent agents can communicate and cooperate with other intelligent agents or humans to complete tasks together; learning ability, meaning that through technologies such as machine learning algorithms, intelligent agents can learn from experience and improve their behavioral strategies; and mobility, meaning that for some types, some intelligent agents can migrate within a network to perform tasks in different locations.

[0077] Intelligent agents can be categorized into the following types: simple reflex agents, which react instantly based on currently perceived information without considering historical information; model-based agents, which possess internal state models and can make more complex decisions based on past experience and currently perceived information; goal-based agents, which set explicit goals and take actions to achieve them; utility-based agents, which, in addition to having goals, evaluate the expected utility of each possible action and select the optimal solution; and learning agents, which have the ability to learn from the environment and improve their performance over time.

[0078] An agent-based instruction manual review system employs a distributed artificial intelligence approach, where multiple autonomous software agents collaborate to complete complex instruction manual review tasks. Each agent may focus on processing a specific type of task or document section and can interact with other agents to share information and coordinate actions.

[0079] In this embodiment of the invention, based on the content tag types of different templating levels in the product manual, corresponding target intelligent agents are assigned from multiple different types of preset intelligent agents to adapt to the content of each templating level. Different intelligent agents are used to review content of different templating levels.

[0080] Step 104: The target intelligent agent reviews the product manual and obtains the review result.

[0081] The review of product manuals may include: structural integrity checks, verifying the inclusion of standard sections such as product introduction, safety warnings, installation instructions, usage methods, troubleshooting, maintenance guides, and after-sales service information; terminology consistency checks, verifying the consistency of terminology used (e.g., whether "switch," "power button," and "button" are used interchangeably); checking whether professional terminology conforms to industry standards or the company's internal glossary; compliance checks, verifying compliance with relevant regulations (e.g., medical device instruction manual specifications), and whether necessary legal statements, disclaimers, and safety labels are included; verification of the correctness of logic and operating procedures, ensuring that the steps are clearly described and unambiguous, and that there are no logical errors (e.g., operating B before A but relying on A already being completed); and visual and formatting reviews, such as the accuracy of text and graphics, whether font size, font, and color match the brand style, and the consistency of multilingual versions (e.g., whether the English and Chinese versions are consistent).

[0082] The output of the review results may include, but is not limited to: a review report, which lists the issues and suggested modifications in a structured manner; highlighted documents, which mark the error locations in the original text and provide annotations; a scoring system, which scores the quality of the instruction manual; and automatic repair suggestions, which provide correction suggestions for standardizable issues (such as inconsistent terminology).

[0083] In this embodiment of the invention, different target intelligent agents review the content of their respective product manuals, obtaining highly accurate review results, including detailed explanations of compliant and non-compliant items, and can also provide improvement suggestions to help users improve the manual content. This invention does not impose any limitations on this.

[0084] This invention introduces a method for reviewing product manuals, comprising: acquiring the product manual; determining the degree of template-based nature of the product manual based on its content; determining a target intelligent agent adapted to the product manual from multiple preset intelligent agents based on the degree of template-based nature of the product manual; wherein different preset intelligent agents are used to review product manuals with different degrees of template-based nature; and reviewing the product manual through the target intelligent agent to obtain a review result. Dynamically assigning different intelligent agents for review based on the degree of template-based nature of the manual content avoids the limitations of fixed rules and single processes in traditional methods, significantly improving review efficiency, shortening review time, reducing the need for manual intervention, and employing targeted intelligent agents for review based on the characteristics of different types of content, enhancing the flexibility and accuracy of the review, and reducing the possibility of misjudgments and omissions.

[0085] Reference Figure 2 The diagram illustrates a flowchart of another product specification review method provided by an embodiment of the present invention. The method may specifically include the following steps:

[0086] Step 201: Obtain the product instruction manual;

[0087] Step 202: Determine the degree of template-based nature of the product manual based on its content;

[0088] Highly templated instructions ensure consistency, allowing reviewers to quickly familiarize themselves with the document structure and expected content layout. This helps expedite the review process and reduces confusion caused by formatting or structural differences.

[0089] Standardized formats and content structures make it easier for reviewers to spot missing parts or non-compliance. For example, if every instruction manual is required to include a safety warning section, its absence will be immediately noticeable. For highly templated manuals, it's more likely that automated tools will be developed to assist the review process. For instance, natural language processing techniques could be used to check the content of specific chapters for compliance or to verify the presence of certain key information. When all manuals follow similar templates, new reviewers can get up to speed more quickly because they only need to learn a fixed review process.

[0090] Overly rigid templates may not be suitable for all types of products, especially those with unique functions or use cases. In such cases, templates may force the inclusion of irrelevant information or omit important specific details. If template changes are needed (e.g., to adapt to new regulatory requirements or market changes), it may involve extensive re-editing and review of existing documents, increasing maintenance costs. While templates help maintain consistency, they may also ignore the specific needs of users in different regions. For example, in a multilingual environment, directly translating text from a template without considering cultural and linguistic differences can lead to misunderstandings. Complete reliance on fixed templates may stifle creative expression, thus affecting the effectiveness and user-friendliness of the instruction manual. Sometimes, unconventional designs are more effective at conveying complex information.

[0091] In one embodiment of this invention, Natural Language Processing (NLP) technology, including keyword extraction, semantic analysis, and text classification algorithms, is used to perform a preliminary analysis of the specification content. The degree of template-ification of the content is determined through preset classification rules (such as text structure, keyword frequency, and template matching rules). The content is classified into "highly template-ified," "mediumly template-ified," or "non-template-ified," and content type tags are generated. This provides a basis for subsequent dynamic allocation by intelligent agents, ensuring that different types of content are reviewed by the most suitable intelligent agent.

[0092] Highly template-based content refers to content with a fixed format and structure, following a clear template, usually involving specific fields or industry standards, such as laws and regulations, technical specifications, and safe operation; Mediumly template-based content refers to content with some template references, but allows for a certain degree of adjustment and change, with higher flexibility, such as product instructions; No template-based content refers to content that lacks a fixed template and relies entirely on the characteristics and needs of the content itself, such as user manuals.

[0093] In one embodiment, step 202 may include the following sub-steps:

[0094] Sub-step S11: Determine the keywords in the content of the product manual and determine the frequency of occurrence of the keywords;

[0095] Sub-step S12: Determine the degree of template-based nature of the product manual based on the keywords and their corresponding frequencies of occurrence.

[0096] In one embodiment of the present invention, the degree of template-ification of the product manual can be determined based on keyword frequency. For example, high template-ification is determined by the high frequency of specific keywords (such as "regulations", "norms", "standards", "clauses", etc.) in the content, which are usually related to a fixed template; medium template-ification is determined by the moderate keyword frequency, which may contain some template reference keywords (such as "instructions", "parameters", "precautions", etc.), but allows for a certain degree of variation; and no template-ification is determined by the low keyword frequency, the lack of fixed template-related keywords, and the content focusing more on descriptive or explanatory statements.

[0097] In another embodiment, step 202 may include the following sub-steps:

[0098] Sub-step S21: Determine the degree of matching between the content of the product manual and the preset template;

[0099] Sub-step S22: Determine the degree of template-ization of the product manual based on the degree of matching.

[0100] In one embodiment of the present invention, the degree of template-ification of the product manual can be determined based on the degree of matching between the content of the product manual and the preset template. For example, a high degree of template-ification is determined by the fact that the content of the product manual highly matches the preset template, has a fixed structure, a unified format, and usually involves industry standards or specifications; a medium degree of template-ification is determined by the fact that the content of the product manual has a certain degree of matching with the preset template, but allows for adjustments or additions in some parts, and has a high degree of flexibility; a no-template-ification is determined by the fact that the content of the product manual has a low degree of matching with the template, or even does not rely on the template at all, and focuses more on the logic and completeness of the content itself.

[0101] In addition, in some embodiments provided by this invention, the degree of template-based nature of the product manual can also be determined by other auxiliary factors, such as: text length and complexity. Highly templated content is usually clearly structured and of moderate length, while untemplated content may be longer and more complex, lacking a fixed structure. Semantic consistency is also important; highly templated content is semantically clear and consistent, conforming to fixed template requirements, while untemplated content is more semantically flexible and may contain more explanatory or instructive content. This invention does not impose any limitations on these aspects.

[0102] Step 203: Based on the degree of template-ification of the product manual, determine the target intelligent agent that is compatible with the product manual from multiple preset intelligent agents; wherein different preset intelligent agents are used to review product manuals with different degrees of template-ification.

[0103] The design of an agent-based instruction manual review system allows for the design of different agents based on the specific needs of the review process. These agents could include those responsible for text parsing and those performing grammar and logic checks. Effective communication channels and collaborative frameworks between agents should be established to ensure seamless data exchange and collaborative problem-solving. Agents should possess self-learning capabilities, accumulating experience from each review process to gradually improve review efficiency and accuracy.

[0104] In some embodiments provided by this invention, based on the content type tags of the degree of templateization, a preset intelligent agent capability model is invoked to dynamically allocate tasks, which can support multi-agent collaborative review. For example, multiple intelligent agents can be invoked simultaneously in the review of complex content. The review process is optimized by task scheduling algorithms (such as priority sorting and resource allocation algorithms) to improve efficiency and ensure that the review task is executed by the most suitable intelligent agent, avoiding resource waste and inefficiency.

[0105] In one embodiment, step 203 may include the following sub-steps:

[0106] Sub-step S31: Based on the first degree of templated nature of the product manual, determine the target intelligent agent that adapts to the product manual as the first type of intelligent agent from multiple preset intelligent agents;

[0107] Sub-step S32: Based on the second degree of template-ization of the product manual, determine the target intelligent agent that adapts to the product manual as a second type of intelligent agent from multiple preset intelligent agents; wherein, the second degree is greater than the first degree;

[0108] Sub-step S33: Based on the product manual's template level being the third level, determine the target intelligent agent adapted to the product manual as a third type intelligent agent from multiple preset intelligent agents; wherein, the third level is greater than the second level.

[0109] In some embodiments provided by this invention, based on the degree of template-ification of the product manual content, the first degree is high template-ification, the second degree is medium template-ification, and the third degree is low template-ification. It can be understood that for highly template-ified manual content, review is more formulaic and easier, therefore the matched agent algorithm function may be relatively simple. As template-ification decreases, the agent algorithm functions used for medium and high template-ification become more complex.

[0110] Step 204: The target intelligent agent reviews the product manual and obtains the review result.

[0111] In one embodiment, step 204 may include the following sub-steps:

[0112] If the target intelligent agent is the first type of intelligent agent, then the content of the product manual is compared with the preset template through the first type of intelligent agent to obtain the review result.

[0113] In some embodiments provided by this invention, a first type of intelligent agent is used to process the content of highly templated product manuals. Specifically, the review method involves using data comparison technology to match the input content item by item with a preset standard template, using string matching algorithms (such as Levenshtein distance) to check the consistency between the text and the template, and using regular expressions to match content with fixed formats (such as dates, numerical ranges, etc.). Finally, a review report is automatically generated, marking non-conformities and providing modification suggestions. This first type of intelligent agent can efficiently review highly templated content (such as warranty policies and safety specifications) to ensure that the content conforms to the fixed template requirements.

[0114] In another embodiment, step 204 may include the following sub-steps:

[0115] If the target intelligent agent is the second type of intelligent agent, then the content of the product manual is compared with the preset template by the second type of intelligent agent; and semantic analysis is performed on the content of the product manual to obtain the semantic analysis result, and it is determined whether the semantic analysis result meets the preset semantic requirements to obtain the review result.

[0116] In some embodiments provided by this invention, the second type of intelligent agent is used to process the content of templated product manuals. Specifically, the review method involves using template matching technology combined with semantic analysis to check whether the content conforms to the template requirements; using rule-based matching algorithms (such as conditional judgment rules) to check the format and structure of the content; and using natural language processing technology (such as semantic understanding) to analyze whether the content conforms to semantic requirements (such as whether "prohibited items" are clear and reasonable). Finally, a review report is automatically generated, marking non-conformities and providing modification suggestions. The second type of intelligent agent can review templated content (such as prohibited items and product parameters) to ensure that the content both conforms to the template requirements and is reasonable.

[0117] In another embodiment, step 204 may include the following sub-steps:

[0118] If the target intelligent agent is the third type of intelligent agent, then the third type of intelligent agent calls the large language model to determine whether the content of the product manual meets the preset review requirements, so as to obtain the review result; the preset review requirements include at least one of the following: reasonable content, complete coverage, clear description, and logical content.

[0119] In some embodiments provided by this invention, the second type of intelligent agent is used to process the content of templated product manuals. Specifically, the review method involves using a large language model (such as GPT-3, GPT-4, etc.) combined with contextual analysis to check whether the content is reasonable and comprehensive; using the large language model to generate potential review questions (such as whether the instructions are clear or whether key steps are missing); using semantic understanding technology to check whether the content is logically consistent (such as whether maintenance steps are coherent); using contextual analysis technology to ensure that the content covers all necessary information (such as whether safety precautions are mentioned); and finally, automatically generating a review report, marking non-conformities and providing improvement suggestions. The first type of intelligent agent can review non-templated content (such as instructions for use and maintenance), ensuring that the content is clear, complete, and logically consistent.

[0120] Step 205: If the review result indicates that the product manual contains abnormal or uncertain content, then the target intelligent agent will conduct a collaborative review with other intelligent agents.

[0121] Product manuals are considered complex due to their involvement in multiple fields, complex structure, need for multi-faceted verification, and high dynamism and uncertainty. For example, the manuals for smart home appliances may cover various technical specifications, safety operating guidelines, and troubleshooting procedures. Reviewing such manuals may require verifying whether the technical specifications conform to industry standards, whether the safety operating guidelines are clear and easy to understand, and whether the troubleshooting procedures are comprehensive and effective, demonstrating the complexity of the content and the need for multi-faceted verification. Furthermore, due to continuous technological advancements, manuals need frequent updates to adapt to the introduction of new products or technologies, resulting in highly dynamic and uncertain content. Therefore, the review process needs to be flexible and adaptable to changes, conducting in-depth verification across multiple areas, which necessitates collaborative review by multiple intelligent agents.

[0122] In some embodiments provided by this invention, if uncertain or abnormal content is found during the review process, the specific review methods are as follows: agents request assistance through message passing mechanisms (such as API calls or message queues); agents share review results and analysis data to further confirm the review results; and a collaborative review mechanism (such as multi-agent voting or weighted average algorithms) is used to improve the accuracy of the review results. In practical applications, the agent collaboration module ensures that these uncertain or abnormal contents can be identified and processed in a timely manner during the review process through message passing mechanisms, thereby improving the accuracy and efficiency of the review. This collaborative mechanism not only enhances the flexibility of the system but also improves the reliability and comprehensiveness of the review results.

[0123] This invention introduces a method for reviewing product manuals, comprising: acquiring the product manual; determining the degree of template-based nature of the product manual based on its content; determining a target intelligent agent adapted to the product manual from multiple preset intelligent agents based on the degree of template-based nature of the product manual; wherein different preset intelligent agents are used to review product manuals with different degrees of template-based nature; and reviewing the product manual through the target intelligent agent to obtain a review result. Dynamically assigning different intelligent agents for review based on the degree of template-based nature of the manual content avoids the limitations of fixed rules and single processes in traditional methods, significantly improving review efficiency, shortening review time, and reducing the need for manual intervention. Furthermore, employing targeted intelligent agent algorithm technology based on the characteristics of different types of content enhances the flexibility and accuracy of the review, reducing the possibility of misjudgments and omissions.

[0124] Reference Figure 3 This diagram illustrates a system architecture diagram of a product manual review method provided by an embodiment of the present invention, including a content classification module, an agent dynamic allocation module, an agent review module, an agent collaborative work module, and a review result output module. Arrows in the diagram indicate the process sequence, and rectangles represent functional modules.

[0125] The content classification module is used for user input of instruction manual questions and answers. It analyzes the degree of template-based nature of the instruction manual content and generates content type tags (such as "highly templated," "mediumly templated," and "no template-based"). In other words, the input stage involves obtaining the target instruction manual document and classifying its content according to its degree of template-based nature. The content classification module generates content type tags using natural language processing technology.

[0126] The intelligent agent dynamic allocation module is used to dynamically allocate corresponding intelligent agents for review based on content type tags.

[0127] The intelligent agent review module is divided into: Intelligent Agent A (for highly templated content review module), Intelligent Agent B (for medium templated content review module) and Intelligent Agent C (for non-templated content review module).

[0128] Agent A is used to process highly templated content (such as warranty policies and safety specifications), performs data comparison and review on highly templated content, outputs review results and non-conformity annotations, and ensures that the content meets the fixed template requirements.

[0129] Agent B is used to process templated content (such as prohibited items and product parameters). It performs template matching and semantic analysis on the templated content, and outputs the review results and non-compliance annotations to ensure that the content not only meets the template requirements but also has reasonableness.

[0130] Agent C is used to process template-free content review modules (such as user manuals and maintenance documents), perform semantic analysis review on template-free content, and output review results and improvement suggestions to ensure that the content is clear, complete and logical.

[0131] The agent collaboration module is used to facilitate interaction and confirmation of review results among agents via message passing during the review process. By confirming review results through message passing if uncertain or abnormal content is found during the review, the module improves the accuracy of review results through multi-agent voting or weighted average algorithms.

[0132] The review results output module is used to generate the final review report, including detailed descriptions of compliance and non-compliance items.

[0133] The advantage of this method lies in its flexibility in handling the diversity and complexity of instruction manuals, while simultaneously improving the automation and accuracy of the review process. Furthermore, the use of a multi-agent system enables high-concurrency processing, accelerating the review of large-scale documents.

[0134] It should be noted that, for the sake of simplicity, the method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments of the present invention are not limited to the described order of actions, because according to the embodiments of the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions involved are not necessarily essential to the embodiments of the present invention.

[0135] Reference Figure 4 The diagram shows a structural block diagram of a product instruction manual review device provided in an embodiment of the present invention, which may specifically include the following modules:

[0136] Module 401 is used to obtain product manuals;

[0137] The templated degree determination module 402 is used to determine the templated degree of the product manual based on the content of the product manual;

[0138] The agent determination module 403 is used to determine the target agent that is compatible with the product manual from multiple preset agents based on the degree of template-ification of the product manual; wherein different preset agents are used to review product manuals with different degrees of template-ification.

[0139] The review module 404 is used to review the product manual through the target intelligent agent and obtain the review result.

[0140] In some embodiments provided by the present invention, the templated degree determination module may include the following sub-modules:

[0141] The keyword determination submodule is used to determine the keywords in the content of the product manual and to determine the frequency of occurrence of the keywords;

[0142] The first template-level determination submodule is used to determine the template level of the product manual based on the keywords and their corresponding frequencies of occurrence.

[0143] In some embodiments provided by the present invention, the templated degree determination module may further include the following sub-modules:

[0144] The matching degree determination submodule is used to determine the degree of matching between the content of the product instruction manual and the preset template;

[0145] The first templated degree determination submodule is used to determine the templated degree of the product manual based on the matching degree.

[0146] In some embodiments provided by the present invention, the intelligent agent determination module includes:

[0147] The first intelligent agent determination submodule is used to determine the target intelligent agent that is compatible with the product manual as a first type intelligent agent from multiple preset intelligent agents, based on the degree of template-ification of the product manual as the first degree.

[0148] The second agent determination submodule is used to determine, from multiple preset agents, a target agent that is compatible with the product manual as a second type of agent, based on the degree of template-ification of the product manual as a second degree; wherein, the second degree is greater than the first degree;

[0149] The third agent determination submodule is used to determine the target agent that is compatible with the product manual as a third type of agent from multiple preset agents, based on the templated degree of the product manual being a third degree; wherein the third degree is greater than the second degree.

[0150] In some embodiments provided by the present invention, the review module includes:

[0151] The first review submodule is used to compare the content of the product manual with a preset template through the first type of intelligent agent if the target intelligent agent is the first type of intelligent agent, and obtain the review result.

[0152] In some embodiments provided by the present invention, the review module includes:

[0153] The second review submodule is used to compare the content of the product manual with a preset template through the second type of intelligent agent if the target intelligent agent is the second type of intelligent agent; and to perform semantic analysis on the content of the product manual to obtain the semantic analysis result, and determine whether the semantic analysis result meets the preset semantic requirements in order to obtain the review result.

[0154] In some embodiments provided by the present invention, the review module includes:

[0155] The third review submodule is used to determine whether the content of the product manual meets the preset review requirements by calling the large language model through the third type of intelligent agent if the target intelligent agent is the third type of intelligent agent, so as to obtain the review result; the preset review requirements include at least one of the following: reasonable content, complete coverage, clear description, and logical content.

[0156] In some embodiments provided by the present invention, it further includes:

[0157] The collaborative review module is used to conduct collaborative review with other intelligent agents through the target intelligent agent if the review result indicates that the product manual contains abnormal or uncertain content.

[0158] This invention provides a product manual review device, comprising: acquiring a product manual; determining the degree of template-based nature of the product manual based on its content; determining a target intelligent agent adapted to the product manual from multiple preset intelligent agents based on the degree of template-based nature of the product manual; wherein different preset intelligent agents are used to review product manuals with different degrees of template-based nature; and reviewing the product manual through the target intelligent agent to obtain a review result. By dynamically assigning different intelligent agents for review based on the degree of template-based nature of the manual content, the limitations of fixed rules and single processes in traditional methods are avoided, significantly improving review efficiency, shortening review time, and reducing the need for manual intervention. Furthermore, by employing targeted intelligent agent algorithm technology based on the characteristics of different types of content, the flexibility and accuracy of the review are enhanced, reducing the possibility of misjudgments and omissions.

[0159] As the device embodiment is basically similar to the method embodiment, the description is relatively simple, and relevant parts can be found in the description of the method embodiment.

[0160] This invention also provides an electronic device, including a processor, a memory, and a computer program stored in the memory and capable of running on the processor. When the computer program is executed by the processor, it implements the various processes of the above-described product specification review method embodiment and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0161] This invention also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the various processes of the above-described product specification review method embodiment and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0162] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0163] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus, or computer program products. Therefore, embodiments of the present invention can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of the present invention can take the form of computer program products implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0164] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0165] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0166] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0167] Although preferred embodiments of the present invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present invention.

[0168] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0169] The present invention has provided a detailed description of a method and apparatus for reviewing product manuals. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, those skilled in the art will recognize that there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for reviewing product manuals, characterized in that, include: Obtain the product manual; Based on the content of the product manual, determine the degree of template-based nature of the product manual; Based on the degree of template-ification of the product manual, a target intelligent agent adapted to the product manual is determined from multiple preset intelligent agents; wherein different preset intelligent agents are used to review product manuals with different degrees of template-ification. The target intelligent agent reviews the product manual and obtains the review results.

2. The method for reviewing product manuals according to claim 1, characterized in that, The step of determining the degree of template-based nature of the product manual based on its content includes: Identify the keywords in the product manual and determine the frequency of occurrence of the keywords; The degree of template-based nature of the product manual is determined based on the keywords and their corresponding frequencies of occurrence.

3. The method for reviewing product manuals according to claim 1, characterized in that, The step of determining the degree of template-based nature of the product manual based on its content includes: Determine the degree of match between the content of the product instruction manual and the preset template; The degree of template-ification of the product manual is determined based on the degree of matching.

4. The method for reviewing product manuals according to claim 1, characterized in that, The preset intelligent agent includes a first type of intelligent agent. The step of reviewing the product manual through the target intelligent agent and obtaining the review result includes: If the target intelligent agent is the first type of intelligent agent, then the content of the product manual is compared with the preset template through the first type of intelligent agent to obtain the review result.

5. The method for reviewing product manuals according to claim 1, characterized in that, The preset intelligent agent includes a second type of intelligent agent. The step of reviewing the product manual through the target intelligent agent and obtaining the review result includes: If the target intelligent agent is the second type of intelligent agent, then the content of the product manual is compared with the preset template by the second type of intelligent agent; and semantic analysis is performed on the content of the product manual to obtain the semantic analysis result, and it is determined whether the semantic analysis result meets the preset semantic requirements to obtain the review result.

6. The method for reviewing product manuals according to claim 1, characterized in that, The preset intelligent agent includes a third type of intelligent agent. The process of reviewing the product manual through the target intelligent agent and obtaining the review result includes: If the target intelligent agent is the third type of intelligent agent, then the third type of intelligent agent calls the large language model to determine whether the content of the product manual meets the preset review requirements, so as to obtain the review result; the preset review requirements include at least one of the following: reasonable content, complete coverage, clear description, and logical content.

7. The method for reviewing product manuals according to claim 1, characterized in that, Also includes: If the review results indicate that the product manual contains abnormal or uncertain content, then the target intelligent agent will conduct a collaborative review with other intelligent agents.

8. The method for reviewing product manuals according to claim 1, characterized in that, The step of determining the target intelligent agent adapted to the product manual from multiple preset intelligent agents based on the degree of template-ification of the product manual includes: Based on the product manual's template level being the first level, the target intelligent agent adapted to the product manual is determined as the first type of intelligent agent from multiple preset intelligent agents; Based on the second degree of template-based nature of the product manual, a target intelligent agent adapted to the product manual is determined from multiple preset intelligent agents as a second type of intelligent agent; wherein, the second degree is greater than the first degree; Based on the product manual's template level being third, a target intelligent agent adapted to the product manual is determined from multiple preset intelligent agents as a third type of intelligent agent; wherein, the third level is greater than the second level.

9. A device for reviewing product manuals, characterized in that, include: The acquisition module is used to acquire product manuals; The template-level determination module is used to determine the template-level of the product manual based on its content. The agent determination module is used to determine the target agent that is compatible with the product manual from multiple preset agents based on the degree of template-ification of the product manual; wherein different preset agents are used to review product manuals with different degrees of template-ification. The review module is used to review the product manual through the target intelligent agent and obtain the review result.

10. An electronic device, characterized in that, include: A processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein the computer program, when executed by the processor, implements the steps of the review method of the product specification as described in any one of claims 1-8.

11. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, which, when executed by a processor, implements the steps of the review method for the product specification as described in any one of claims 1-8.

Images