Electromechanical product design cognitive scheme generation method and system based on pre-training model

Abstract

The invention provides an electromechanical product design cognitive scheme generation method and system based on a pre-training model. The method and system are used for solving the problem that the efficiency of a design scheme of a complex engineering problem product is low. The method comprises the following steps: identifying improvement parameters and deterioration parameters from an engineering problem text of a target mechanical and electrical product; querying an invention principle number from a pre-constructed contradiction matrix knowledge base according to the improvement parameter and the deterioration parameter, wherein the contradiction matrix knowledge base stores a mapping relationship between the improvement parameter and the deterioration parameter and the invention principle number; querying the invention principle information corresponding to the invention principle number from an invention principle knowledge base, wherein the corresponding relationship between the invention principle number and the invention principle information is stored in the invention principle knowledge base; and reasoning the engineering problem text and the invention principle information through the pre-training language model to obtain a design cognition scheme of the target mechanical and electrical product.

Description

Technical Field

[0001] This application relates to the technical field of artificial intelligence, machine learning and engineering technology intersection, and specifically to a method and system for generating cognitive schemes for electromechanical product design based on pre-trained models. Background Technology

[0002] Currently, most product design solutions in the electromechanical field are completed through the work of product designers. However, this manual design model relies on the subjective experience and creativity of designers, and it suffers from significant efficiency bottlenecks. In particular, product designers are less efficient at designing solutions for complex engineering problems in the electromechanical field. Summary of the Invention

[0003] The purpose of this application is to provide a method and system for generating cognitive design schemes for electromechanical products based on pre-trained models, in order to improve the low efficiency of design schemes for complex engineering problems.

[0004] This application provides a method for generating a design cognitive scheme for electromechanical products based on a pre-trained model, comprising: identifying improvement parameters and deterioration parameters from the engineering problem text of the target electromechanical product; querying the invention principle number from a pre-constructed contradiction matrix knowledge base based on the improvement parameters and deterioration parameters, wherein the contradiction matrix knowledge base stores the mapping relationship between the improvement parameters and deterioration parameters and the invention principle number; querying the invention principle information corresponding to the invention principle number from the invention principle knowledge base, wherein the invention principle knowledge base stores the correspondence between the invention principle number and the invention principle information; and inferring the design cognitive scheme of the target electromechanical product by reasoning on the engineering problem text and the invention principle information through a pre-trained language model. In the implementation of the above scheme, by combining the mapping relationship of the contradiction matrix knowledge base with the generalization reasoning ability of the pre-trained language model, the pre-trained language model, when analyzing engineering problems, not only extracts the invention principle based on the mapping relationship of the contradiction matrix knowledge base, but also dynamically associates potential contradiction parameters or non-explicit technical requirements that are not explicitly labeled based on its implicit learning of massive technical texts. This integration makes the design process no longer limited to the static matching of known parameters, but actively explores the deep technical conflicts behind the problem and proposes design solutions for electromechanical products through the semantic understanding and logical deduction of the model, thereby effectively improving the efficiency of design solutions for complex engineering problem products.

[0005] Optionally, in this embodiment, identifying improvement parameters and deterioration parameters from the engineering problem text of the target electromechanical product includes: acquiring a standard engineering parameter knowledge base and a parameter prompt template; filling the parameter prompt template with the standard engineering parameter knowledge base and the engineering problem text to obtain parameter prompt engineering text; and inputting the parameter prompt engineering text into a trained parameter recognition model so that the pre-trained language model outputs improvement parameters and deterioration parameters. In the implementation of the above scheme, by filling the parameter prompt template with the standard engineering parameter knowledge base and the engineering problem text, the engineering problem text can be systematically associated with the standard engineering parameter knowledge base. This allows the parameter recognition model to more accurately understand the context and semantics of the engineering problem text. This systematic processing avoids the complexity and subjectivity of manually analyzing and interpreting the engineering problem text, making the parameter recognition process more efficient and reliable, thereby improving the accuracy of identifying improvement parameters and deterioration parameters.

[0006] Optionally, in this embodiment, before inputting the parameter prompt engineering text into the trained parameter recognition model, the method further includes: obtaining a training dataset, which includes: sample problem text, sample improvement parameters, and sample deterioration parameters. The sample improvement parameters and sample deterioration parameters are parameters from a standard engineering parameter knowledge base and are labeled for the sample problem text. The parameter recognition model is then fine-tuned using the sample problem text as training data and the sample improvement parameters and sample deterioration parameters as training labels. In the implementation of the above scheme, by using sample improvement parameters and sample deterioration parameters from a standard engineering parameter knowledge base as training labels, the parameter recognition model can learn the parameter improvement and deterioration patterns related to specific problem texts. This training method enables the model to more accurately determine which parameters are beneficial to problem solving and which are harmful when recognizing parameters, thereby improving the model's decision-making quality in practical applications. Furthermore, by using labeled sample problem text as training data, the parameter recognition model can learn a more accurate mapping relationship between text features and parameters. This helps the model to more quickly and accurately identify relevant improvement and deterioration parameters when processing new problem texts, thereby improving the model's response speed and processing efficiency.

[0007] Optionally, in this embodiment, reasoning on engineering problem text and invention principle information using a pre-trained language model includes: obtaining a solution prompt template; filling the engineering problem text and invention principle information into the solution prompt template to obtain solution prompt engineering text; and inputting the solution prompt engineering text into the pre-trained language model so that the pre-trained language model outputs a design cognition solution for the target product. In the implementation of the above solution, by using a pre-trained language model for reasoning, complex engineering problem text and invention principle information can be automatically processed and analyzed, thereby significantly improving the efficiency and accuracy of generating design cognition solutions. Furthermore, by introducing a solution prompt template, the input of engineering problem text and invention principle information is structured, which helps the model better understand and process this information, thereby improving the accuracy and consistency of reasoning.

[0008] This application also provides a system for generating design cognitive schemes for electromechanical products based on a pre-trained model, including: a terminal device and a server; the terminal device is used to send engineering problem text of a target electromechanical product to the server; the server is used to identify improvement parameters and deterioration parameters from the engineering problem text of the target electromechanical product; based on the improvement parameters and deterioration parameters, it queries the invention principle number from a pre-built contradiction matrix knowledge base, which stores the mapping relationship between improvement parameters and deterioration parameters and invention principle numbers; it queries the invention principle information corresponding to the invention principle number from the invention principle knowledge base, which stores the correspondence between invention principle numbers and invention principle information; and it uses a pre-trained language model to reason about the engineering problem text and invention principle information to obtain a design cognitive scheme for the target electromechanical product.

[0009] This application also provides a device for generating design cognitive schemes for electromechanical products based on a pre-trained model, comprising: an engineering parameter identification module for identifying improvement parameters and deterioration parameters from the engineering problem text of the target electromechanical product; a principle number query module for querying the invention principle number from a pre-built contradiction matrix knowledge base based on the improvement parameters and deterioration parameters, wherein the contradiction matrix knowledge base stores the mapping relationship between improvement parameters and deterioration parameters and invention principle numbers; a principle information query module for querying the invention principle information corresponding to the invention principle number from the invention principle knowledge base, wherein the invention principle knowledge base stores the correspondence between invention principle numbers and invention principle information; and a scheme reasoning generation module for reasoning on the engineering problem text and invention principle information through a pre-trained language model to obtain a design cognitive scheme for the target electromechanical product.

[0010] Optionally, in this embodiment, the engineering parameter recognition module includes: a prompt template acquisition submodule, used to acquire a standard engineering parameter knowledge base and a parameter prompt template; a prompt template filling submodule, used to fill the standard engineering parameter knowledge base and engineering problem text into the parameter prompt template to obtain parameter prompt engineering text; and a prompt text input submodule, used to input the parameter prompt engineering text into the trained parameter recognition model so that the pre-trained language model outputs improved parameters and worsened parameters.

[0011] Optionally, in this embodiment, the engineering parameter identification module further includes: a training data acquisition submodule, used to acquire a training dataset, the training dataset including: sample problem text, sample improvement parameters and sample deterioration parameters, the sample improvement parameters and sample deterioration parameters are parameters from a standard engineering parameter knowledge base and labeled on the sample problem text; and a model fine-tuning training submodule, used to fine-tune the parameter identification model using the sample problem text as training data and the sample improvement parameters and sample deterioration parameters as training labels.

[0012] Optionally, in this embodiment, the solution reasoning generation module includes: a solution template acquisition submodule, used to acquire a solution prompt template; an engineering text acquisition submodule, used to fill the engineering problem text and invention principle information into the solution prompt template to obtain solution prompt engineering text; and an engineering text input submodule, used to input the solution prompt engineering text into a pre-trained language model so that the pre-trained language model outputs the design cognition solution of the target product.

[0013] This application also provides an electronic device, including a processor and a memory, wherein the memory stores machine-readable instructions executable by the processor, and the machine-readable instructions are executed by the processor to perform the methods described above.

[0014] This application also provides a computer-readable storage medium storing a computer program, which is executed by a processor to perform the methods described above.

[0015] This application also provides a computer program product, including: a computer program or computer instructions, which are executed by a processor to perform the method described above. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 The diagram shown is a flowchart illustrating the electromechanical product design cognitive scheme generation method based on a pre-trained model provided in an embodiment of this application. Figure 2 The diagram shown is a schematic of a pre-trained model-based electromechanical product design cognitive scheme generation system provided in an embodiment of this application. Figure 3 The diagram shown is a structural schematic of the electromechanical product design cognitive scheme generation device based on a pre-trained model provided in an embodiment of this application. Figure 4 The diagram shows a structural schematic of an electronic device provided in an embodiment of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the accompanying drawings in the embodiments of this application are for illustrative and descriptive purposes only and are not intended to limit the protection scope of the embodiments of this application. Furthermore, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in the embodiments of this application illustrate operations implemented according to some embodiments of this application. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without logical contextual relationships may be reversed or implemented simultaneously. In addition, those skilled in the art, guided by the content of the embodiments of this application, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts.

[0019] Furthermore, the described embodiments are merely a part of the embodiments of this application, and not all of them. The components of the embodiments of this application described and illustrated herein can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed embodiments of this application, but merely to illustrate selected embodiments of this application.

[0020] It is understood that the terms "first" and "second" in the embodiments of this application are used to distinguish similar objects. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and that "first" and "second" do not necessarily imply that they are different. In the description of the embodiments of this application, the term "and / or" is merely a description of the association relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship. The term "multiple" refers to two or more (including two), and similarly, "multiple groups" refers to two or more groups (including two groups).

[0021] It should be noted that the electromechanical product design cognitive scheme generation method based on pre-trained models provided in this application can be executed by electronic devices. Here, electronic devices refer to device terminals or servers with the function of executing computer programs. Device terminals include, for example, smartphones, personal computers, tablets, personal digital assistants, or mobile internet devices. Servers refer to devices that provide computing services through a network. Servers include, for example, x86 servers and non-x86 servers. Non-x86 servers include, for example, mainframes, minicomputers, and UNIX servers.

[0022] In related technologies, product design solutions in the electromechanical field are mostly completed through the design work of product designers. However, electromechanical systems typically involve the intersection of multiple disciplines such as mechanics, electronics, control, and materials, requiring designers to comprehensively consider the synergistic effects of various subsystems. The design of complex electromechanical systems often requires multiple modifications, with each adjustment necessitating re-verification of structure, function, and performance. For example, the design of aero-engines requires repeated testing of material performance and energy consumption balance, with a single iteration potentially taking weeks or even months. The rapid iteration of materials, new processes, and new equipment demands continuous learning from designers, but manual design relies on accumulated experience, making it difficult to keep up with emerging technologies. Furthermore, designers need to reduce material costs (such as alternatives to high-performance alloys) and manufacturing costs (such as simplifying processing techniques) while ensuring performance, but manual experience is insufficient for systematically optimizing the cost-performance balance. This manual design model relies on the subjective experience and creativity of designers, requiring manual processing of massive amounts of parameters, repeated verification of solutions, and coordination of multiple needs. It is also limited by personal experience and industry knowledge boundaries, thus exhibiting significant efficiency bottlenecks.

[0023] Please see Figure 1The illustrated flowchart illustrates a method for generating electromechanical product design cognitive solutions based on a pre-trained model, as provided in this application embodiment. The main idea of ​​this method is to combine the mapping relationship of a contradiction matrix knowledge base with the generalization reasoning ability of a pre-trained language model to proactively uncover the deep-seated technical conflicts behind problems and propose design solutions for electromechanical products, thereby effectively improving the efficiency of designing complex engineering problems. The implementation methods of the aforementioned method for generating electromechanical product design cognitive solutions based on a pre-trained model may include: Step S110: Identify the improvement parameters and deterioration parameters from the engineering problem text of the target electromechanical product.

[0024] Target electromechanical products refer to electromechanical equipment or systems that need to solve specific engineering problems, typically involving multiple disciplines such as mechanics, electronics, and control. Examples of target electromechanical products include smartwatches, industrial robots, and electric valves.

[0025] Engineering problem text describes technical contradictions or design flaws in the practical application of a target electromechanical product. A technical contradiction or design flaw typically implies two opposing parameters, also known as a contradictory parameter pair. However, these are not explicitly expressed in the user-inputted engineering problem text. Therefore, a pre-trained language model is needed to extract these two opposing parameters from the user-input engineering problem text. These parameters are mutually conflicting, such as improvement parameters (e.g., indicators that the user wants to improve or promote in the engineering problem) and deterioration parameters (e.g., indicators that are reduced or suppressed in the engineering problem). The aforementioned engineering problem text can include the problem phenomenon, the required objectives, and potential constraints.

[0026] Improvement parameters refer to the performance indicators that need to be optimized in engineering problems, such as the indicators that users want to improve or promote in engineering problems, which usually correspond to the core objectives of user needs.

[0027] Deteriorating parameters refer to other performance indicators that may be worsened during the optimization and improvement process in engineering problems. This deterioration means that the parameters change in a direction that the user does not want, such as indicators that are reduced or suppressed in engineering problems.

[0028] Step S120: Based on the improvement parameters and deterioration parameters, retrieve the invention principle number from the pre-built contradiction matrix knowledge base. The contradiction matrix knowledge base stores the mapping relationship between the improvement parameters and deterioration parameters and the invention principle number.

[0029] The contradiction matrix is ​​a tool matrix built based on the Theory of Inventive Problem Solving (TRIZ). For example, in early TRIZ theory, a mapping relationship matrix was established between 39 general engineering parameters (such as strength, weight, cost, etc.) and 40 inventive principles (such as division, reverse action, etc.). The contradiction matrix is ​​in the form of a 39×39 matrix, where each row or column represents 39 general engineering parameters.

[0030] The contradiction matrix knowledge base is a knowledge base constructed from a contradiction matrix composed of standard engineering parameters. This contradiction matrix knowledge base can be built using a Redis cache database or memcached. In practice, two data entries in the contradiction matrix knowledge base can be represented in JSON format as {"14-1":["1","8","40","15"],"10-22":["14","15"]}. Taking the first data entry as an example, the "14" in the first "14-1" represents an improvement parameter like "strength," while the "1" represents a deterioration parameter like "weight of the moving object." The "1," "8," "40," and "15" in the square brackets represent the inventive principle numbers, such as "1" for the segmentation principle, "8" for the weight compensation principle, "15" for the dynamic characteristic principle, and "40" for the principle of using composite materials.

[0031] Step S130: Query the invention principle information corresponding to the invention principle number from the invention principle knowledge base. The invention principle knowledge base stores the correspondence between invention principle numbers and invention principle information.

[0032] An invention principle knowledge base refers to a knowledge base that stores the correspondence between invention principle numbers and invention principle information. The aforementioned invention principle information is related to the invention principle of a general solution proposed in TRIZ theory. This invention principle information may include: invention principle number (ID), invention principle name, invention principle definition (such as detailed description), and / or engineering application examples corresponding to the invention principle, etc.

[0033] Step S140: Use a pre-trained language model to reason about the engineering problem text and invention principle information to obtain a design cognitive solution for the target electromechanical product.

[0034] Pre-trained language models are large language models (LLMs) trained on large-scale data. They possess the ability to understand natural language and generate design cognitive solutions, and are used to reason about engineering problem texts and invention principle information and output analysis results. The aforementioned pre-trained language models can employ Deepseek series models (such as Deepseek R1-7B), Qwen series models, Llama series models (such as Llama 3-8B), etc.

[0035] Design cognition schemes are innovative design schemes generated by combining pre-trained language models with engineering problem analysis and invention principle information. The aforementioned design cognition schemes may include: conceptual scheme name, detailed description of engineering problem solution, functional description, and expected effects.

[0036] In the implementation of the above scheme, by combining the mapping relationship of the contradiction matrix knowledge base with the generalization reasoning ability of the pre-trained language model, the pre-trained language model, when analyzing engineering problems, not only extracts the invention principle based on the mapping relationship of the contradiction matrix knowledge base, but also dynamically associates potential contradiction parameters or non-explicit technical requirements that are not explicitly labeled based on its implicit learning of massive technical texts. This integration makes the design process no longer limited to the static matching of known parameters, but actively explores the deep technical conflicts behind the problem and proposes design solutions for electromechanical products through the semantic understanding and logical deduction of the model, thereby effectively improving the efficiency of design solutions for complex engineering problem products.

[0037] Optionally, before identifying the improvement parameters and deterioration parameters from the engineering problem text of the target electromechanical product, a pre-set input template can guide the user to fill in an engineering problem text containing "improvement objectives" (corresponding to improvement parameters) and "potential negative impacts" (corresponding to deterioration parameters). The input template content could include, for example: What kind of electromechanical product do you want to design? What specific improvements are needed for this electromechanical product? What negative impacts might these improvements cause? As an optional implementation of step S110 above, the implementation of identifying improvement parameters and deterioration parameters from the engineering problem text of the target electromechanical product may include: Step S111: Obtain the standard engineering parameter knowledge base and parameter prompt template.

[0038] The Standard Engineering Parameter Knowledge Base is a knowledge base composed of standard engineering parameters. This knowledge base can be represented in JSON format as: {"1":"Weight of the moving object","2":"Weight of the stationary object","3":"Length of the moving object","4":"Length of the stationary object","5":"Area of ​​the moving object","6":"Area of ​​the stationary object","7":"Volume of the moving object","8":"Volume of the stationary object","9":"Velocity","10":"Force","11":"Stress or pressure","12":"Shape","13":"Structural stability","14":"Strength","15":"Time of action for the moving object","16":"Time of action for the stationary object","17":"Temperature","18":"Illuminance","19":"Illuminance","19":"Illuminance","10":"Weight of the moving object","11":"Time of action for the stationary object","17":"Temperature","18":"Illuminance","19 ... ":"Energy of a moving object","20":"Energy of a stationary object","21":"Power","22":"Energy loss","23":"Material loss","24":"Information loss","25":"Time loss","26":"Quantity of matter or things","27":"Reliability","28":"Testing accuracy","29":"Manufacturing accuracy","30":"Sensitivity to external harmful factors","31":"Harmful factors generated by the object","32":"Manufacturability","33":"Operability","34":"Maintainability","35":"Adaptability and versatility","36":"Complicated device","37":"Difficulty of monitoring and testing","38":"Degree of automation","39":"Productivity"}.

[0039] Parameter prompt templates are prompt word templates used by the model to identify improvement parameters and deterioration parameters in the engineering problem text. For example: "You are a TRIZ expert. Please identify the technical contradictions from the following engineering problem description and output them in the format [Improvement Parameter No..Name, Deterioration Parameter No..Name]. Parameters should be selected from the 39 standard engineering parameters of TRIZ. Engineering problem description: {user_input}".

[0040] For example, the implementation of step S111 above can be achieved by using an executable program to obtain a standard engineering parameter knowledge base and parameter prompt template. The standard engineering parameters mentioned above refer to the general engineering parameters in TRIZ theory, such as the 39 general engineering parameters in early TRIZ theory. The standard engineering parameter knowledge base mentioned above refers to a knowledge base composed of multiple standard engineering parameters.

[0041] Step S112: Fill the parameter prompt template with the standard engineering parameter knowledge base and engineering problem text to obtain the parameter prompt engineering text.

[0042] Understandably, assuming the engineering problem text is "We want to improve the strength of the product without increasing its weight," and the pre-trained language model is fine-tuned after being trained with a standard engineering parameter knowledge base, then it's also possible to simply fill the engineering problem text into the parameter hint template. The resulting parameter hint engineering text could be: "You are a TRIZ expert. Please identify the technical contradictions from the following engineering problem text description and output them in the format [Improvement Parameter Number.Name, Deterioration Parameter Number.Name]. Parameters should be selected from the 39 standard engineering parameters of TRIZ. Engineering problem text description: We want to improve the strength of the product without increasing its weight."

[0043] For example, in implementing step S112 above: Assume the engineering problem text is "We want to improve the strength of the product without increasing its weight," and the parameter prompt template is "You are a TRIZ expert. Please identify the technical contradictions from the following engineering problem text description and output them in the format [Improvement Parameter Number.Name, Deterioration Parameter Number.Name]. The parameters should be selected from the 39 standard engineering parameters of TRIZ. Engineering problem text description: {user_input}, Standard engineering parameter knowledge base: {parameters_knowledge_base}." Then the obtained parameter prompt engineering text could be "You are a TRIZ expert. Please identify the technical contradictions from the following engineering problem text description and output them in the format [Improvement Parameter Number.Name, Deterioration Parameter Number.Name]." Output in the specified format. Parameters should be selected from the 39 standard engineering parameters of TRIZ. Engineering problem text description: We want to improve the strength of the product without increasing its weight. Standard engineering parameter knowledge base: {"1":"Weight of moving object","2":"Weight of stationary object","3":"Length of moving object","4":"Length of stationary object","5":"Area of ​​moving object","6":"Area of ​​stationary object","7":"Volume of moving object","8":"Volume of stationary object","9":"Velocity","10":"Force","11":"Stress or pressure","12":"Shape","13":"Structural stability","14":"Strength","15":"Time of action of moving object","16":"Time of action of stationary object","17":"Temperature","18":} "Illuminance","19":"Energy of moving objects","20":"Energy of stationary objects","21":"Power","22":"Energy loss","23":"Material loss","24":"Information loss","25":"Time loss","26":"Quantity of matter or things","27":"Reliability","28":"Test accuracy","29":"Manufacturing accuracy","30":"Sensitivity to external harmful factors","31":"Harmful factors generated by the object","32":"Manufacturability","33":"Operability","34":"Maintainability","35":"Adaptability and versatility","36":"Complicatedness of the device","37":"Difficulty of monitoring and testing","38":"Degree of automation","39":"Productivity"}.

[0044] Step S113: Input the parameter prompt engineering text into the trained parameter recognition model so that the pre-trained language model outputs improved parameters and deteriorated parameters.

[0045] For example, step S113 can be implemented by inputting the parameter prompt engineering text into the trained parameter recognition model through the calling program of the pre-trained language model, so that the pre-trained language model outputs the improvement parameters and the deterioration parameters. Assuming the engineering problem text in the parameter prompt engineering text is "We want to improve the strength of the product without increasing its weight," the output of the pre-trained language model can be represented in JSON format as "[14. Strength, 2. Weight of stationary object]". Here, "[14. Strength, 2. Weight of stationary object]" represents the improvement parameter as "strength" (number 14) and the deterioration parameter as the weight of stationary object (number 2).

[0046] In implementing the above scheme, by using the trained parameter recognition model, the powerful natural language processing capabilities of the pre-trained language model can be fully utilized. This allows the model to extract key improvement and deterioration parameters from complex engineering problem texts. Because the pre-trained language model has high generalization ability and robustness when processing large amounts of text data, it can maintain high recognition accuracy and consistency when handling engineering problem texts of different types and complexities. Furthermore, by inputting the parameter-hinted engineering text into the trained parameter recognition model, the process of identifying improvement and deterioration parameters from engineering problem texts can be automated, reducing the time and cost of manual intervention and operation. This automated processing flow not only improves work efficiency but also avoids human error, ensuring the accuracy and reliability of the recognition results.

[0047] As an optional implementation of step S113 above, before inputting the parameter prompt engineering text into the trained parameter recognition model, the pre-trained language model can also be trained. This implementation may include: Step S113a: Obtain the training dataset, which includes: sample problem text, sample improvement parameters and sample deterioration parameters. The sample improvement parameters and sample deterioration parameters are parameters from the standard engineering parameter knowledge base and are labeled on the sample problem text.

[0048] For example, the implementation of step S113a above can be as follows: an executable program can be used to obtain a training dataset. A single data point in this training dataset can be represented in JSON format as “{"input":"We want to improve the strength of the product without increasing its weight","improve":{"id":14,"name":"strength"},"worsen":{"id":2,"name":"weight of a stationary object"}}”. Here, “We want to improve the strength of the product without increasing its weight” is the sample problem text, {"id":14,"name":"strength"} represents the sample improvement parameter, and {"id":2,"name":"weight of a stationary object"} represents the sample deterioration parameter.

[0049] Step S113b: Use the sample problem text as training data and the sample improvement parameters and sample deterioration parameters as training labels to fine-tune the parameter recognition model.

[0050] An example implementation of step S113b above is as follows: A parameter recognition model is used to perform inference and prediction on the sample problem text to obtain predicted improvement parameters and predicted deterioration parameters. Then, a first loss value is calculated between the predicted improvement parameters and the sample improvement parameters, and a second loss value is calculated between the predicted deterioration parameters and the sample deterioration parameters. A total loss value is determined based on the first and second loss values. Based on this total loss value, the weight parameters of the parameter recognition model are updated during the fine-tuning training process. It is understood that since the sample problem texts in the training dataset are similar to the problem texts in actual application scenarios, fine-tuning training enables the model to better adapt to problems in actual applications, reducing model errors in practical applications and improving the model's generalization ability.

[0051] As an optional implementation of step S120 above, the implementation of retrieving the invention principle number from a pre-built contradiction matrix knowledge base based on improvement parameters and deterioration parameters may include: retrieving the invention principle number from a pre-built contradiction matrix knowledge base based on improvement parameters and deterioration parameters. For example: assuming that two data entries in the contradiction matrix knowledge base are represented in JSON format as {"14-1":["1","8","40","15"],"10-22":["14","15"]}, and the number of the improvement parameter is 14, and the number of the deterioration parameter is 1, then the invention principle numbers 1, 8, 40, and 15 can be directly retrieved from the contradiction matrix knowledge base. Here, 1 represents the segmentation principle, 8 represents the weight compensation principle, 15 represents the dynamic characteristic principle, and 40 represents the principle of using composite materials.

[0052] As an optional implementation of step S130 above, the invention principle information corresponding to the invention principle number is queried from the invention principle knowledge base, which stores the correspondence between the invention principle number and the invention principle information.

[0053] The Invention Principle Knowledge Base refers to a knowledge base that stores the correspondence between invention principle numbers and invention principle information.

[0054] For example, the implementation of step S130 above is as follows: Assuming the invention principle number is 1, and the invention principle knowledge base stores the correspondence between the invention principle number and the invention principle information, then the invention principle information corresponding to the invention principle number can be queried from the invention principle knowledge base. The invention principle information can be "{"Invention Principle Number":"1","Invention Principle Name":"Segmentation","Invention Principle Definition":"Divide an object into independent parts","Engineering Application Example":"For example, setting up different recycling bins for the recycling of different materials (such as glass, paper, iron cans, etc.)"}.

[0055] As an optional implementation of step S140 above, the above-described implementation of reasoning about engineering problem text and invention principle information using a pre-trained language model may include: Step S141: Obtain the solution prompt template.

[0056] For example, the implementation of step S141 above can be as follows: Use an executable program to obtain a solution prompt template, which can be: "You are a product design expert. Based on the following engineering problem description and inventive principle, please generate a specific design concept solution. Engineering problem: {user_input}, recommended inventive principle: {principle_text}; please output the design concept solution name, detailed explanation of the engineering problem solution, functional description and expected effect."

[0057] Step S142: Fill the engineering problem text and invention principle information into the solution prompt template to obtain the solution prompt engineering text.

[0058] For example, in implementing step S142 above: Assume the engineering problem text is "We want to increase the strength of the product without increasing its weight," and the inventive principle information is "{"Inventive Principle Number":"1","Inventive Principle Name":"Segmentation","Inventive Principle Definition":"Divide an object into independent parts","Engineering Application Example":"For example, setting up different recycling bins for different materials (such as glass, paper, tin cans, etc.)"}". Then, the engineering problem text and inventive principle information can be filled into the solution prompt template, resulting in the following solution prompt engineering text: "You are a product design expert. Based on the following engineering problem description and inventive principle, please generate a specific design concept solution. Engineering Problem: We want to increase the strength of the product without increasing its weight. Recommended Inventive Principle: {"Inventive Principle Number":"1","Inventive Principle Name":"Segmentation","Inventive Principle Definition":"Divide an object into independent parts" "parts), "Engineering Application Examples": "For example, setting up different recycling bins for the recycling of different materials (such as glass, paper, tin cans, etc.)"; Please output the design concept scheme name, detailed description of the engineering problem solution, functional description and expected effect.

[0059] Step S143: Input the solution prompt engineering text into the pre-trained language model so that the pre-trained language model outputs the design cognition solution of the target product.

[0060] For example, step S143 can be implemented by inputting the engineering text of the solution prompt into the pre-trained language model through a calling program, so that the pre-trained language model outputs a design cognition solution for the target product. This method for generating electromechanical product design cognition solutions based on a pre-trained model can be used to develop and create a product innovation design technology contradiction analysis system. By applying the technology contradiction analysis method from TRIZ theory to an automated system for analyzing product innovation design technology contradictions through a pre-trained language model, it can improve problems affecting the main functions of electromechanical products and enhance their performance and innovation design efficiency. In the implementation of this solution, by filling the engineering problem text and invention principle information into the solution prompt template and inputting it into the pre-trained language model, the powerful reasoning ability of the model can be utilized to generate customized design cognition solutions for specific engineering problems. This not only improves the flexibility of the design process but also reduces the possibility of human error.

[0061] This application also provides a method for generating a design cognitive scheme for electromechanical products based on a pre-trained model. The method includes: sending an engineering problem text of a target electromechanical product to a server via a terminal device; identifying improvement parameters and deterioration parameters from the engineering problem text of the target electromechanical product via the server; querying the invention principle number from a pre-built contradiction matrix knowledge base based on the improvement parameters and deterioration parameters, the contradiction matrix knowledge base storing the mapping relationship between improvement parameters, deterioration parameters, and invention principle numbers; querying the invention principle information corresponding to the invention principle number from an invention principle knowledge base, the invention principle knowledge base storing the correspondence between invention principle numbers and invention principle information; and inferring the design cognitive scheme of the target electromechanical product using a pre-trained language model based on the engineering problem text and invention principle information. It is understood that the implementation of this method is the same as or similar to the implementation of steps S110 to S140 above, and therefore will not be described again.

[0062] Please see Figure 2 The diagram shown is a schematic of a pre-trained model-based electromechanical product design cognitive scheme generation system provided in an embodiment of this application; the above-mentioned pre-trained model-based electromechanical product design cognitive scheme generation system 200 may include: a terminal device 210 and a server 220.

[0063] Terminal device 210 is used to send engineering problem text of the target electromechanical product to the server.

[0064] Server 220 is used to identify improvement parameters and deterioration parameters from the engineering problem text of the target electromechanical product; to query the invention principle number from a pre-built contradiction matrix knowledge base based on the improvement parameters and deterioration parameters, which stores the mapping relationship between improvement parameters, deterioration parameters and invention principle numbers; to query the invention principle information corresponding to the invention principle number from the invention principle knowledge base, which stores the correspondence between invention principle number and invention principle information; and to obtain the design cognitive scheme of the target electromechanical product by reasoning on the engineering problem text and invention principle information through a pre-trained language model.

[0065] In implementing the above solution, a pre-trained language model is used to reason about engineering problem texts and invention principle information. By combining contextual information and complex semantic relationships, more practical and innovative design cognition solutions can be generated. This language model-based reasoning method overcomes the limitations of traditional rule or template generation, providing more flexible and diverse solutions. Furthermore, the system integrates the collaborative work of terminal devices and servers, enabling the generation of design cognition solutions not only to be performed locally but also through more complex calculations and knowledge base queries on the server. This achieves cross-device and cross-platform generation of design cognition solutions, enhancing the system's flexibility and scalability.

[0066] Please see Figure 3 The diagram shown is a structural schematic of the electromechanical product design cognitive scheme generation device based on a pre-trained model provided in this application embodiment; this application embodiment provides an electromechanical product design cognitive scheme generation device 300 based on a pre-trained model, including: The engineering parameter identification module 310 is used to identify improvement parameters and deterioration parameters from the engineering problem text of the target electromechanical product.

[0067] The principle number query module 320 is used to query the invention principle number from the pre-built contradiction matrix knowledge base based on the improvement parameters and deterioration parameters. The contradiction matrix knowledge base stores the mapping relationship between the improvement parameters and deterioration parameters and the invention principle number.

[0068] The principle information query module 330 is used to query the invention principle information corresponding to the invention principle number from the invention principle knowledge base. The invention principle knowledge base stores the correspondence between the invention principle number and the invention principle information.

[0069] The scheme reasoning generation module 340 is used to reason about the engineering problem text and invention principle information through a pre-trained language model to obtain the design cognitive scheme of the target electromechanical product.

[0070] As an optional implementation of the above-mentioned device, the engineering parameter identification module includes: The prompt template retrieval submodule is used to retrieve the standard engineering parameter knowledge base and parameter prompt templates.

[0071] The prompt template filling submodule is used to fill the parameter prompt template with the standard engineering parameter knowledge base and engineering problem text to obtain the parameter prompt engineering text.

[0072] The prompt text input submodule is used to input parameter prompt engineering text into the trained parameter recognition model, so that the pre-trained language model outputs improved and worsened parameters.

[0073] As an optional implementation of the above-mentioned device, the engineering parameter identification module further includes: The training data acquisition submodule is used to acquire the training dataset, which includes: sample problem text, sample improvement parameters, and sample deterioration parameters. The sample improvement parameters and sample deterioration parameters are parameters from the standard engineering parameter knowledge base and are labeled on the sample problem text.

[0074] The model fine-tuning training submodule is used to fine-tune the parameter recognition model using sample problem text as training data and sample improvement parameters and sample deterioration parameters as training labels.

[0075] As an optional implementation of the above-mentioned device, the scheme reasoning generation module includes: The solution template retrieval submodule is used to retrieve solution prompt templates.

[0076] The Engineering Text Acquisition Submodule is used to fill the engineering problem text and invention principle information into the solution prompt template to obtain the solution prompt engineering text.

[0077] The engineering text input submodule is used to input the engineering text prompts into the pre-trained language model, so that the pre-trained language model can output the design and cognitive scheme of the target product.

[0078] It should be understood that this device corresponds to the above-described embodiment of the electromechanical product design cognitive scheme generation method based on pre-trained models, and is capable of performing the various steps involved in the above-described method embodiments. The specific functions of this device can be found in the description above, and detailed descriptions are appropriately omitted here. The device includes at least one software functional module that can be stored in memory or embedded in the device's operating system (OS) in the form of software or firmware.

[0079] Please see Figure 4 The diagram shows a structural schematic of an electronic device provided in an embodiment of this application. An electronic device 400 provided in this application includes a processor 410 and a memory 420. The memory 420 stores machine-readable instructions executable by the processor 410. When the machine-readable instructions are executed by the processor 410, the method described above is performed.

[0080] This application embodiment also provides a computer-readable storage medium 430, on which a computer program is stored. This computer program is executed by a processor 410 to perform the methods described above. The computer-readable storage medium 430 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0081] This application also provides a computer program product, including: a computer program or computer instructions, which are executed by a processor to perform the method described above.

[0082] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For apparatus embodiments, since they are basically similar to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0083] It should be understood that the disclosed apparatus and methods can also be implemented in other ways, as provided in the embodiments of this application. The apparatus embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of this application. In this regard, each block in the flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the accompanying drawings. For example, two consecutive blocks may actually be executed substantially in parallel, or they may sometimes be executed in reverse order, depending primarily on the functions involved.

[0084] Furthermore, the functional modules of each embodiment in this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part. In addition, in the description of this specification, the reference to terms such as "one embodiment," "some embodiments," "example," "specific example," "some examples," etc., means that the specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. Furthermore, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0085] The above description is only an optional implementation of the embodiments of this application, but the protection scope of the embodiments of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the embodiments of this application should be covered within the protection scope of the embodiments of this application.

Claims

1. A method for generating cognitive design schemes for electromechanical products based on a pre-trained model, characterized in that, include: Identify improvement parameters and deterioration parameters from the engineering problem text of the target electromechanical product; The invention principle number is retrieved from a pre-built contradiction matrix knowledge base based on the improvement parameters and the deterioration parameters. The contradiction matrix knowledge base stores the mapping relationship between the improvement parameters, the deterioration parameters and the invention principle number. The invention principle information corresponding to the invention principle number is retrieved from the invention principle knowledge base, which stores the correspondence between invention principle numbers and invention principle information. By reasoning about the engineering problem text and the invention principle information through a pre-trained language model, a design cognitive scheme for the target electromechanical product can be obtained.

2. The method according to claim 1, characterized in that, The identification of improvement parameters and deterioration parameters from the engineering problem text of the target electromechanical product includes: Access the standard engineering parameter knowledge base and parameter prompt templates; The standard engineering parameter knowledge base and the engineering problem text are filled into the parameter prompt template to obtain the parameter prompt engineering text; The parameter prompt engineering text is input into the trained parameter recognition model so that the pre-trained language model outputs the improved parameters and the deteriorated parameters.

3. The method according to claim 2, characterized in that, Before inputting the parameter prompt engineering text into the trained parameter recognition model, the method further includes: Obtain a training dataset, which includes: sample problem text, sample improvement parameters, and sample deterioration parameters. The sample improvement parameters and sample deterioration parameters are parameters from the standard engineering parameter knowledge base and are annotated for the sample problem text. Using the sample problem text as training data, and the sample improvement parameters and sample deterioration parameters as training labels, the parameter recognition model is fine-tuned and trained.

4. The method according to claim 1, characterized in that, The step of reasoning about the engineering problem text and the invention principle information using a pre-trained language model includes: Get solution prompt template; Fill the engineering problem text and the invention principle information into the solution prompt template to obtain the solution prompt engineering text; The proposed solution is input into the pre-trained language model, so that the pre-trained language model outputs the design and cognitive solution for the target product.

5. A method for generating cognitive design schemes for electromechanical products based on a pre-trained model, characterized in that, include: Send engineering problem text of the target electromechanical product to the server via terminal device; The server identifies improvement parameters and deterioration parameters from the engineering problem text of the target electromechanical product; based on the improvement parameters and deterioration parameters, it queries the invention principle number from a pre-built contradiction matrix knowledge base, which stores the mapping relationship between the improvement parameters, the deterioration parameters and the invention principle number; it queries the invention principle information corresponding to the invention principle number from the invention principle knowledge base, which stores the correspondence between the invention principle number and the invention principle information; and it uses a pre-trained language model to reason about the engineering problem text and the invention principle information to obtain the design cognitive scheme of the target electromechanical product.

6. A system for generating cognitive design schemes for electromechanical products based on a pre-trained model, characterized in that, include: Terminal devices and servers; The terminal device is used to send engineering problem text of the target electromechanical product to the server; The server is used to identify improvement parameters and deterioration parameters from the engineering problem text of the target electromechanical product; to query the invention principle number from a pre-built contradiction matrix knowledge base based on the improvement parameters and the deterioration parameters, the contradiction matrix knowledge base storing the mapping relationship between the improvement parameters and the deterioration parameters and the invention principle number; to query the invention principle information corresponding to the invention principle number from the invention principle knowledge base, the invention principle knowledge base storing the correspondence between the invention principle number and the invention principle information; and to obtain the design cognitive scheme of the target electromechanical product by reasoning on the engineering problem text and the invention principle information through a pre-trained language model.

7. A device for generating cognitive design schemes for electromechanical products based on a pre-trained model, characterized in that, include: The engineering parameter identification module is used to identify improvement parameters and deterioration parameters from the engineering problem text of the target electromechanical product. The principle number query module is used to query the invention principle number from a pre-built contradiction matrix knowledge base based on the improvement parameters and the deterioration parameters. The contradiction matrix knowledge base stores the mapping relationship between the improvement parameters, the deterioration parameters and the invention principle number. The principle information query module is used to query the invention principle information corresponding to the invention principle number from the invention principle knowledge base, which stores the correspondence between invention principle numbers and invention principle information. The scheme reasoning generation module is used to reason about the engineering problem text and the invention principle information through a pre-trained language model to obtain the design cognitive scheme of the target electromechanical product.

8. An electronic device, characterized in that, include: A processor and a memory, the memory storing machine-readable instructions executable by the processor, the machine-readable instructions being executed by the processor to perform the method of any one of claims 1 to 5.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the method according to any one of claims 1 to 5.

10. A computer program product, characterized in that, include: A computer program or computer instructions that, when executed by a processor, perform the method described in any one of claims 1 to 5.

Images