Building design method, system and equipment based on large language model

Abstract

The invention is suitable for the technical field of artificial intelligence, and particularly relates to a building design method based on a large language model, and the method comprises the steps: inputting description information into a target large language model, and obtaining an initial building scheme; optimizing the initial building scheme according to sustainable data corresponding to the application scene to obtain a target building scheme; and outputting the target building scheme. The invention further provides a building design system based on the large language model, terminal equipment and a computer readable storage medium. By means of the scheme, the problems that a building scheme generated by a large language model is not professional enough and does not meet the requirements of the specific field can be solved, and better building design experience is provided for users.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of artificial intelligence, and particularly relates to a building design method and system based on a large language model, a terminal device, and a computer readable storage medium. BACKGROUND

[0002] In the prior art, an artificial intelligence model related to building design cannot provide corresponding assistance to architects when assisting architects in work because a general large language model (LLM) cannot understand semantics in text content and image content, and cannot provide professional and specific field-compliant responses to architects, resulting in low quality of building schemes obtained by architects from the general large language model and affecting the overall experience of architects in using the general large language model in the design process.

[0003] How to obtain professional and specific field-compliant building schemes through a large language model is a problem that needs to be solved by the application. SUMMARY

[0004] The embodiments of the application provide a building design method and system based on a large language model, a terminal device, and a computer readable storage medium, which can solve the problem that building schemes generated by a large language model are not professional and do not comply with specific fields.

[0005] In a first aspect, the application provides a building design method based on a large language model, comprising:

[0006] inputting description information into a target large language model to obtain an initial building scheme;

[0007] optimizing the initial building scheme according to sustainable data corresponding to an application scenario to obtain a target building scheme, the description information being used to describe the application scenario;

[0008] outputting the target building scheme.

[0009] The application integrates a large language model in the building field into a building design process, uses the large language model to help architects sort out design ideas, and provides creative inspiration for architects. By inputting description information and other information related to building design by architects, the large language model in the building field generates a preliminary building design scheme that meets the requirements, for reference and selection by architects, and assists architects in completing building design work.

[0010] Optionally, before the description information is input into the target large language model to obtain the initial building scheme, the method further comprises:

[0011] analyzing an application scenario of the input information according to the input information of the user.

[0012] Get the user's description of the application scenario.

[0013] Optionally, the application scenario of the input information is analyzed based on the user's input information, including

[0014] Obtaining text information and / or image information input by the user;

[0015] Analyze textual and / or graphical information for construction information relevant to the construction industry;

[0016] Analyze application scenarios based on building information.

[0017] Optionally, before inputting the description information into the target large language model to obtain the initial building plan, the method further includes:

[0018] Get design labels based on design semantics and design language in the construction industry;

[0019] The historical dataset stored in the database is annotated using design labels to obtain the architectural dataset;

[0020] The initial large language model is trained using the building dataset to obtain the target large language model.

[0021] Optionally, the architectural design method further comprises:

[0022] Obtaining feedback data input by users regarding the target building plan;

[0023] Optimize the target large language model based on the feedback data.

[0024] Optionally, before outputting the target building plan, the following is also included:

[0025] When there are more than two target building schemes, the professional score of each target building scheme is calculated based on the sustainable data standards in the application scenario;

[0026] The final target architectural plan is selected based on professional scores.

[0027] Optionally, the initial building plan is optimized according to the sustainable data corresponding to the application scenario to obtain a target building plan, including:

[0028] Determine the parameter types of sustainable data in the application scenario;

[0029] The target building plan is obtained by optimizing the initial building plan according to the preset parameters in the parameter type.

[0030] In a second aspect, the present application provides an architectural design system based on a large language model, the architectural design system comprising:

[0031] An input module is configured to input description information into the target large language model to obtain an initial building scheme, and the description information is used to describe an application scenario.

[0032] An optimization module is configured to optimize the initial building scheme according to sustainable data corresponding to the application scenario to obtain a target building scheme.

[0033] An output module is configured to output the target building scheme.

[0034] In a third aspect, the present application provides a terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the method of any one of the above aspects when executing the computer program.

[0035] In a fourth aspect, the present application provides a computer-readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the method of any one of the above aspects.

[0036] It can be understood that the beneficial effects of the above-mentioned second aspect to fourth aspect can be referred to the related description in the above-mentioned first aspect, which will not be repeated here.

[0037] The beneficial effects of the embodiments of the present application compared with the prior art are as follows:

[0038] The present application interacts with the large language model by inputting the description information by the architect. The description information includes information such as functional requirements and site conditions of the building. The large language model will generate a preliminary building design scheme according to these information, and combine the learned knowledge in the field of architecture to meet the requirements. Then, according to the corresponding sustainable data in the building application scenario, the preliminary building design scheme is optimized to obtain a building design scheme that better meets the building industry. The sustainable data includes parameters such as window-wall ratio, carbon emission of building materials, and light reflectivity. The present application takes the building design scheme as a reference and selection in the work of the architect, helps the architect to quickly sort out the design ideas, stimulates the creative inspiration of the architect, and thus better completes the building design work. BRIEF DESCRIPTION OF DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0040] Figure 1 is a flowchart of the building design method based on the large language model provided by the embodiments of the present application;

[0041] Figure 2 is a sub-flowchart of the building design method based on a large language model provided by an embodiment of the present application;

[0042] Figure 3 is a flowchart of step S201 in an embodiment of the present application;

[0043] Figure 4 is another sub-flowchart of the building design method based on a large language model provided by an embodiment of the present application;

[0044] Figure 5 is still another sub-flowchart of the building design method based on a large language model provided by an embodiment of the present application;

[0045] Figure 6 is yet another sub-flowchart of the building design method based on a large language model provided by an embodiment of the present application;

[0046] Figure 7 is a flowchart of step S102 in an embodiment of the present application;

[0047] Figure 8 is a structural diagram of the building design system based on a large language model provided by an embodiment of the present application;

[0048] Figure 9 is a structural diagram of a terminal device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0049] In the following description, specific details are set forth in order to provide a thorough understanding of embodiments of the application. However, persons having ordinary skill in the art will readily recognize that embodiments of the application can be practiced without these specific details. In other instances, well-known structures have not been shown or described in order not to obscure the description of embodiments of the application.

[0050] It should be understood that the term "comprises" when used in this specification and the appended claims indicates the presence of the described features, integers, steps, operations, elements, and / or components but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0051] It should also be understood that the term "and / or" when used in this specification and the appended claims indicates that the associated listed items can be present one or more of the associated listed items and all possible combinations of the associated listed items.

[0052] As used in the specification and the appended claims herein, the term “if’ can be construed to mean “when” or “once,” or “in response to a determination” or “in response to detecting,” that a stated condition or event has occurred. Similarly, the phrase “if it is determined” or “if [a stated condition or event] is detected” can be construed to mean “once it is determined” or “in response to a determination,” or “once [a stated condition or event] is detected” or “in response to detecting [a stated condition or event],” depending on the context.

[0053] In addition, in the description of the present application and the appended claims, the terms “first,” “second,” “third,” and the like are used only to distinguish descriptions, and cannot be understood as indicating or implying relative importance.

[0054] In the present application, the reference “one embodiment” or “some embodiments” and the like means that the specific features, structures or characteristics described in connection with the embodiment are included in one or more embodiments of the present application. Therefore, the statements “in one embodiment,” “in some embodiments,” “in other some embodiments,” “in further some embodiments” and the like appearing in different places in the specification are not necessarily all referring to the same embodiment, but mean “one or more but not all embodiments,” unless otherwise specifically emphasized. The terms “include,” “contain,” “have” and their variants mean “including but not limited to,” unless otherwise specifically emphasized.

[0055] In the embodiments of the present application, the large language model has the ability to process massive text data, and can deeply understand the meaning of the text and generate natural language responses.

[0056] Please refer to Figure 1 , Figure 1 is a flowchart of a building design method based on a large language model provided by the embodiments of the present application. The building design method based on a large language model specifically includes the following steps:

[0057] Step S101, inputting description information into a target large language model to obtain an initial building scheme.

[0058] The description information is used to describe the application scenario, including but not limited to user's demand, geographical location, environment, climate, regulations and planning, etc. The input form of the description information includes but is not limited to text, picture and text plus picture, etc. The user in the embodiments of the present application is generally an architect, and other users interested in building design can also apply the embodiments of the present application.

[0059] The large language model (LLM) mentioned in the present application is a generative artificial intelligence (GenAI) platform named EarlyArchi, which includes an artificial intelligence technology driven natural language processing tool (Chat Generative Pre-trained Transformer, ChatGPT). The initial architectural scheme includes all the information of the current architectural design scheme, ensuring that it can fully adapt to and meet the architectural design needs of users.

[0060] In the present embodiment, the architect inputs the description information by conversing with ChatGPT in natural language. The description information includes data on geographical location, environment, climate, regulations and planning, etc. ChatGPT, as a large language model, can understand and analyze the content of the user's input and feed back the corresponding content to the architect. Specifically, in the architectural design industry, the terminal device inputs data with the architect's design requirements into the target large language model to obtain an initial architectural scheme.

[0061] In the present embodiment, the terminal device communicates with the architect through ChatGPT, interacts according to the context of the chat, and converts the natural language input by the architect into labels that are more easily recognized by the target large language model, effectively improving the efficiency of the architect using the target large language model.

[0062] The architectural design method based on a large language model provided in the present embodiment utilizes large language model technology in a specific field to provide support for early architectural design concept development for architects. By deeply integrating the advanced functions of general large language models, EarlyArchi achieves the goal of improving design efficiency in the architectural field, thereby promoting innovation and development in the architectural design industry.

[0063] Step S102, optimizing the initial architectural scheme according to the sustainable data corresponding to the application scenario to obtain a target architectural scheme.

[0064] In some embodiments, the application scenarios include residential building scenarios, commercial building scenarios, office building scenarios, cultural building scenarios, educational building scenarios and industrial building scenarios, etc. The sustainable data corresponding to the application scenarios includes data on energy efficiency, environmental impact and indoor environmental quality, etc.

[0065] In other embodiments, the sustainable data includes parameters such as window-to-wall ratio, carbon emissions of building materials and light reflectivity, etc.

[0066] Specifically, energy efficiency includes the energy consumption level of the building, usually measured in kilowatt-hours (kWh) or kilowatt-hours per square meter (kWh / m 2 ) per year; the energy saving rate of the building after adopting energy-saving technologies and equipment such as solar power systems, energy-saving lamps, and high-efficiency insulation materials; and the proportion of solar or other renewable energy in the total energy consumption.

[0067] Environmental impact includes the environmental performance of building materials, such as the use of renewable or recycled materials; carbon emissions during the building's life cycle, usually measured in carbon dioxide (CO2) equivalents; the efficiency of rainwater collection systems and wastewater treatment equipment, and the proper management of waste.

[0068] Indoor environmental quality includes the control range of indoor temperature and humidity, and its impact on employee productivity; the effect of ventilation design, such as fresh air volume and air flow rate; and indoor air quality indicators such as PM2.5 and formaldehyde concentrations.

[0069] By selecting and replacing the building materials used in the initial building scheme based on the above sustainable data, the initial building scheme is optimized to obtain a target building scheme that better meets the actual application requirements.

[0070] Step S103, output the target building scheme.

[0071] In this embodiment, the architect can choose the output form of the target building scheme according to his actual needs, and the output form includes but is not limited to text, picture, and text plus picture. The target building scheme is a professional building scheme that meets the sustainable conditions.

[0072] Specifically, in the case where the architect inputs text information and requires text information as output, when implementing text-to-text conversion, natural language processing technology (NLP) is usually required. In some embodiments, natural language processing technology includes but is not limited to recurrent neural network (RNN), long short-term memory recurrent neural network architecture (LSTM), and Transformer, etc., which are not limited here. These models can capture the semantic and contextual information of the text, thereby realizing complex building text conversion tasks.

[0073] Specifically, in the case where the architect inputs text information and requires picture information as output, when implementing text-to-picture conversion, the specific steps required include but are not limited to text embedding, image generation, and image optimization.

[0074] wherein the text embedding is converting the input text into a fixed-size vector representation so that the model can understand and process it. Some possible embodiments complete the image generation through natural language processing models, such as a pre-trained language model based on Transformer (Bidirectional Encoder Representations from Transformers, BERT), OpenAI API, DALL-E 2 or GPT, etc. Then a generative model is used to generate new data samples, such as Generative Adversarial Networks (GANs) or diffusion models. GANs is a deep learning model that can generate corresponding architectural images according to the text embedding. The target large language model in this embodiment generates architectural design schemes that match the text description by learning the mapping relationship from text to image.

[0075] Further, the architectural design images generated by the target large language model may need to be optimized and adjusted to better meet the user's expectations. Optimization and adjustment include color adjustment, detail enhancement, or style transfer operations, etc.

[0076] Specifically, in the case where the architect inputs picture information and requires output of text information and picture information, the terminal device uses deep learning techniques, including Convolutional Neural Network (CNN) and generative models, to realize the conversion of pictures when realizing the conversion of pictures. The specific various network models mentioned in the embodiments of the present application are only one way to realize the embodiments and are not limiting. In actual application, the actual needs are the standard.

[0077] The embodiments of the present application interact with the large language model by inputting description information from the architect, which includes functional requirements and site conditions of the building, etc. The large language model generates an initial architectural design scheme that meets the requirements according to these information, combined with the architectural field knowledge it has learned. The terminal device can further optimize the preliminary architectural design scheme according to the sustainable data corresponding to the architectural application scene, such as window-wall ratio, carbon emission of building materials, and light reflectivity parameters, to obtain a building design scheme that better meets the requirements of the building industry. These architectural design schemes can serve as a reference and selection for architects, helping architects quickly sort out design ideas and stimulate their creative inspiration, so as to better complete the architectural design work.

[0078] Please refer to Figure 2 , Figure 2is a subflowchart of a building design method based on a large language model provided in the present application. In the present embodiment, before the description information is input into the target large language model to obtain an initial building scheme, the following steps are further included:

[0079] In step S201, the application scenario of the input information is analyzed according to the input information of the user.

[0080] In the present embodiment, the terminal device can first analyze the semantics, context, tone, and keywords contained in the input information of the user, and extract the vocabulary used for the building design scheme. Then, the terminal device can query the application scenario of the data with the above-mentioned vocabulary. Since there are recognition errors in data analysis, the terminal device can output the information of the application scenario obtained by the query to the display screen, so that the architect can confirm whether the application scenario obtained by the current analysis is correct. In the case where no positive answer is received from the architect, the terminal device can re-analyze the information input by the architect or obtain more input information of the architect.

[0081] In the case where a positive answer is received from the architect, the input information of the architect is further analyzed to obtain more information or data about the scene.

[0082] In step S202, the description information of the user for the application scenario is obtained.

[0083] In the present embodiment, the description information includes the size, material, structure, energy efficiency, etc. of the building. Specifically, a building model is obtained using 3D modeling and rendering technology, which can create a highly realistic building model, so that the architect can more intuitively see the influence of the parameter change on the building scheme.

[0084] In other embodiments, the architect can view the appearance, internal structure, and performance parameters of the building in real time on the display.

[0085] In other embodiments, the architect can use virtual reality (VR) or augmented reality (AR) technology. Virtual reality technology can create an immersive building environment and allow the architect to freely walk and view the parameters in the building through a head-mounted device or other interactive device, so that the architect can more intuitively see the influence of the parameter change on the building scheme. Augmented reality technology enhances the user's perception of the real world by superimposing computer-generated virtual information onto the real world. This technology uses specific devices such as smartphones, tablets, or specialized AR glasses to seamlessly integrate virtual content with the real environment, creating a hybrid reality experience. This allows the architect to more intuitively see the influence of the parameter change on the building scheme.

[0086] Referring to Figure 3 , Figure 3 is a flowchart of step S201 in an embodiment of the present application. In step S201, the application scenario of the input information is analyzed according to the input information of the user, including the following steps:

[0087] In step S2011, the text information and / or image information input by the user are obtained.

[0088] In this embodiment, the user interface or application programming interface of EarlyArch is used to receive the text information and / or various picture information input by the user about the construction industry.

[0089] In step S2012, the building information related to the construction industry in the text information and / or image information is analyzed.

[0090] In this embodiment, the processing of the text information by the terminal device includes decomposing the text into individual words or phrases and determining the part of speech of each word (such as noun or verb, etc.), using named entity recognition to identify building-related entities in the text, such as building names, locations, architects, and building types, etc. The terminal device analyzes the sentence structure in the text and extracts the relationships between building entities, such as the location, age, style, and function of the building, etc.

[0091] The processing of the image information includes necessary preprocessing operations on the image, such as scaling, cropping, and denoising, etc., and then using a convolutional neural network or other image feature extraction method to extract building-related features from the image, such as shape, texture, and color, etc. Then, through target detection algorithm, the building objects in the image are identified, and image segmentation technology is used to separate the building from the background. Finally, according to the extracted features and detected building objects, the attributes of the building are inferred, such as building style, material, and structure, etc.

[0092] In step S2013, the application scenario is analyzed according to the building information.

[0093] In this embodiment, the application scenario of the building scheme is analyzed according to the parameters of the location, age, style, function, attribute, style, material, and structure of the building.

[0094] Referring to Figure 4 , Figure 4 is a sub-flowchart of a building design method based on a large language model provided in an embodiment of the present application. In this embodiment, before the description information is input into the target large language model to obtain the initial building scheme, the following steps are further included:

[0095] In step S401, the design label is obtained according to the design semantics and design language in the construction industry.

[0096] In the construction industry, design semantics refers to the meaning and information conveyed by a design, while design language is the specific tools and methods to achieve these semantics. Design labels are a summary and description of design features and styles, which help people quickly understand and identify design works. Design semantics and design language play a crucial role in forming design labels.

[0097] Design semantics includes functionality, aesthetics, culture, and sustainability, etc. For example, the design semantics of a library may emphasize a quiet reading environment, a rich collection of space, and a place for cultural exchange.

[0098] Design language includes architectural style, material selection, spatial layout, and color use, etc. For example, in the case of a library, the design language may be manifested in the use of traditional architectural style to reflect cultural heritage, the selection of natural materials to create a quiet atmosphere, the rational planning of space to meet different functional needs, and the use of soft colors to enhance the reading experience.

[0099] In the embodiments of the present application, the extraction of design labels needs to consider design semantics, design language, and the needs and preferences of architects. All labels mentioned in this embodiment are only examples and not limited.

[0100] Step S402, using design labels to annotate the historical data set stored in the database to obtain an architectural data set.

[0101] In the construction industry, the historical data stored in the database includes various architectural text data and image data. The terminal device uses the design labels analyzed in the above steps to re-label the existing data set and obtains an architectural data set specifically labeled for the construction industry.

[0102] Step S403, training the initial large language model using the architectural data set to obtain a target large language model.

[0103] In this embodiment, the terminal device first selects a pre-trained initial large language model as a basis, such as GPT or BERT, etc. These models have been trained on a large amount of text data and image data and have certain understanding and generation capabilities. Then the initial large language model is trained for domain adaptation to enable the initial large language model to have professional analysis capabilities in the construction industry.

[0104] In some embodiments, some domain-specific layers or modules are added to the top layer of the large language model to better capture specific information in the construction field, thereby obtaining a target large language model more suitable for the construction field.

[0105] Please refer to Figure 5 , Figure 5is a sub-flowchart of a building design method based on a large language model provided by an embodiment of the present application. In this embodiment, the building design method further includes the following steps:

[0106] In step S501, feedback data about the target building scheme input by the user is obtained.

[0107] In some embodiments, a questionnaire containing multiple-choice questions and open-ended questions can be designed to cover various aspects of the scheme, such as design, functionality, sustainability, and cost, and then be sent to users of different industries and identities to obtain feedback data about the target building scheme. Analyze the feedback data to get the satisfied aspects and points for improvement of the target building scheme by different users.

[0108] In step S502, the target large language model is optimized according to the feedback data.

[0109] In some embodiments, the terminal device analyzes the problems of the target large language model according to the above feedback data, such as syntax errors, semantic understanding deviations, and incoherent output. The target large language model is optimized for the above problems.

[0110] Please refer to Figure 6 , Figure 6 is a sub-flowchart of a building design method based on a large language model provided by an embodiment of the present application. In this embodiment, the building design method further includes the following steps:

[0111] In step S601, in the case where the number of target building schemes is more than two, the professional scores of each target building scheme are calculated according to the sustainable data standards in the application scenario.

[0112] In this embodiment, sustainable data standards applicable to different scenarios are determined. The sustainable data standards may include indicators such as environmental impact, resource utilization efficiency, energy use, carbon emissions, and social impact of buildings. Existing standards are derived from international or domestic green building certification systems, or are self-defined standards based on local specific environmental and cultural backgrounds.

[0113] According to the determined sustainable data standards, the scoring standards are formulated. Each standard can be assigned different weights according to its importance and impact. For example, energy use efficiency may be considered as a very important standard and thus be assigned a higher weight; while some social impact standards may be assigned a lower weight according to specific circumstances.

[0114] According to the scoring standards, the professional scores are calculated in combination with the performance of the building scheme in each standard. The sustainable data standards and scoring standards mentioned in this embodiment are only examples and not limiting.

[0115] Step S602, screening the final target building scheme according to the professional score.

[0116] In this embodiment, the architect can combine the professional score with other non-quantitative factors, and make trade-offs and comparisons. Based on the comprehensive evaluation result, the target building scheme that best meets the needs and expectations is selected.

[0117] Please refer to Figure 7 , Figure 7 is a flowchart of step S102 in an embodiment of the present application, wherein step S102, according to the sustainable data corresponding to the application scenario, optimizes the initial building scheme to obtain the target building scheme, includes the following steps:

[0118] Step S1021, determine the parameter type of the sustainable data under the application scenario.

[0119] In some embodiments, the construction industry includes three values related to sustainable development, namely window-wall ratio, carbon emissions, and visual light reflectance. In different application scenarios, the use of the above three sustainable data is different.

[0120] Step S1022, according to the preset parameter in the parameter type, optimize the initial building scheme to obtain the target building scheme.

[0121] In some embodiments, after confirming the preset sustainable data parameter type under the application scenario, the target building scheme can be optimized according to the sustainable data parameters. For example, reducing energy consumption during construction, improving renewable energy utilization rate during construction, and improving indoor environmental quality.

[0122] This embodiment introduces sustainable data under the application scenario, so that the building design scheme is more in line with the actual application, provides reasonable and effective scheme for architects, and improves the efficiency of architects in scheme design.

[0123] The embodiment of the present application provides a building design system 800 based on a large language model, which includes an input module 810, an optimization module 820, and an output module 830.

[0124] The input module 810 is configured to input description information into a target large language model to obtain an initial building scheme; the optimization module 820 is configured to optimize the initial building scheme according to sustainable data corresponding to an application scenario to obtain a target building scheme; and the output module 830 is configured to output the target building scheme.

[0125] In some embodiments, the building design system further comprises an analysis module and an acquisition module. The analysis module is configured to analyze an application scenario of the input information according to the input information of the user, and the description information is configured to describe the application scenario. The acquisition module is configured to acquire the description information of the application scenario provided by the user.

[0126] In some embodiments, the building design system further comprises a collection module, a labeling module and a training module. The collection module is configured to obtain design labels according to design semantics and design languages in the building industry. The labeling module is configured to label historical data sets stored in a database to obtain a building data set by using the design labels. The training module is configured to train an initial large language model to obtain a target large language model by using the building data set.

[0127] In some embodiments, the analysis module comprises an acquisition unit, a first analysis unit and a second analysis unit. The acquisition unit is configured to acquire text information and / or image information input by the user. The first analysis unit is configured to analyze building information related to the building industry in the text information and / or the image information. The second analysis unit is configured to analyze the application scenario according to the building information.

[0128] In some embodiments, the building design system further comprises a feedback module and an optimization model module. The feedback module is configured to acquire feedback data about the target building scheme input by the user. The optimization model module is configured to optimize the target large language model according to the feedback data.

[0129] In some embodiments, the optimization module in the building design system comprises a determination unit and an output unit. The determination unit is configured to determine a parameter type of the sustainable data in the application scenario. The output unit is configured to optimize the initial building scheme to obtain a final scheme according to a preset parameter in the parameter type.

[0130] In some embodiments, the building design system further comprises a scoring module and a screening module. The scoring module is configured to calculate professional scores of each target building scheme according to sustainable data standards in the application scenario when the number of the target building schemes is more than two. The screening module is configured to screen a final target building scheme according to the professional scores.

[0131] The embodiments of the present application also provide a terminal device, which comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the method of any one of the above building design methods based on a large language model is implemented.

[0132] The terminal device provided in the embodiments of the present application can be a mobile phone, a tablet computer, an augmented reality (AR) / virtual reality (VR) device, a notebook computer, an ultra-mobile personal computer (UMPC), a netbook, a personal digital assistant (PDA), or the like. The embodiments of the present application do not limit the specific type of the terminal device.

[0133] The terminal device 9 in the embodiments of the present application includes a memory 901, one or more processors 902 (only one is shown) and a computer program stored in the memory 901 and executable on the processor. The memory 901 is configured to store software programs and units, and the processor 902 is configured to execute the building design method based on a large language model by running the software programs and units stored in the memory 901. Figure 9

[0134] It should be understood that, in the embodiments of the present application, the processor 902 can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, or the like. The general-purpose processor can be a microprocessor or any conventional processor, or the like.

[0135] The memory 901 can include read-only memory and random access memory, and provide instructions and data for the processor 902. Part or all of the memory 901 can also include non-volatile random access memory. For example, the memory 901 can also store device category information.

[0136] ​The embodiment of the application interacts with a large language model by an architect inputting description information, the description information including functional requirements and site conditions of the building, etc. The large language model generates an initial building design scheme according to the information, in combination with the learned knowledge in the field of architecture. The terminal device can further optimize the preliminary building design scheme according to the sustainable data corresponding to the building application scene, such as the window-wall ratio, carbon emissions of building materials, and light reflectivity, etc. to obtain a building design scheme that is more in line with the building industry. The building design schemes can serve as a reference and selection for the architect, help the architect quickly sort out the design ideas, stimulate creative inspiration, and thus better complete the building design work.

[0137] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional units and modules is exemplified, and in actual application, the above functions can be completed by different functional units and modules according to needs, that is, the internal structure of the above device is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit or module in the embodiment can be integrated in one processing unit, or each unit can exist physically independently, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit. In addition, the specific name of each functional unit or module is only for convenient distinction, and does not limit the protection scope of the application. The specific working process of the unit or module in the above system can refer to the corresponding process in the foregoing method embodiments, which will not be repeated here.

[0138] The embodiments of the present application further provide a computer readable storage medium storing a computer program. The computer program, when executed by a processor, causes the processor to perform the steps of any of the above methods. Specifically, the program can be stored in a non-volatile computer readable storage medium. When executed, the program can include the procedures of the embodiments of the above methods. Any reference in the embodiments of the present application to a memory, a storage, a database, or another medium can include a non-volatile and / or volatile memory. The non-volatile memory can include a read-only memory (ROM), a programmable ROM (PROM), an electrically programmable ROM (EPROM), an electrically erasable programmable ROM (EEPROM), or a flash memory. The volatile memory can include a random access memory (RAM) or an external cache memory. As an illustration but not a limitation, the RAM is available in many forms such as a static RAM (SRAM), a dynamic RAM (DRAM), a synchronous DRAM (SDRAM), a double data rate SDRAM (DDR SDRAM), an enhanced SDRAM (ESDRAM), a Synchlink DRAM (SLDRAM), a Rambus direct RAM (RDRAM), a direct Rambus dynamic RAM (DRDRAM), and a Rambus dynamic RAM (RDRAM), etc.

[0139] In the above embodiments, the description of each embodiment has its own focus, and the parts not described or recorded in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0140] Those skilled in the art can understand that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solutions. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0141] In the embodiments provided by the present application, it should be understood that the disclosed apparatus / network device and method can be implemented in other ways. For example, the above-described apparatus / network device embodiments are merely schematic. For example, the division of the modules or units is merely a logical function division. There can be another division manner in actual implementation. For example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the shown or discussed units can be indirect coupling or communication connection through some interfaces, devices or units, and can be electrical, mechanical or other forms.

[0142] The units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment scheme.

[0143] The above embodiments are only used to illustrate the technical solutions of the present application, but not limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can still be modified, or some technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A building design method based on a large language model, characterized by, The building design method comprises: inputting description information into a target large language model to obtain an initial building scheme, the description information being used to describe an application scenario; optimizing the initial building scheme according to sustainable data corresponding to the application scenario to obtain a target building scheme; outputting the target building scheme.

2. The architectural design method of claim 1, wherein, Before inputting the description information into the target large language model to obtain the initial building scheme, the method further comprises: analyzing an application scenario of input information of a user according to the input information of the user; obtaining description information of the user for the application scenario.

3. The architectural design method of claim 2, wherein, The method of analyzing the application scenario of the input information of the user according to the input information of the user comprises obtaining text information and / or image information input by the user; analyzing building information related to the building industry in the text information and / or the image information; analyzing the application scenario according to the building information.

4. The architectural design method of claim 1, wherein Before inputting the description information into the target large language model to obtain the initial building scheme, the method further comprises: obtaining design labels according to design semantics and design languages in the building industry; annotating historical data sets stored in a database to obtain building data sets by using the design labels; training the initial large language model by using the building data sets to obtain the target large language model.

5. The architectural design method of claim 1, wherein The building design method further comprises: obtaining feedback data input by the user about the target building scheme; optimizing the target large language model according to the feedback data.

6. The architectural design method of claim 1, wherein Before outputting the target building scheme, the method further comprises: in a case where the number of the target building schemes is more than two, calculating professional scores of the target building schemes according to sustainable data standards in the application scenario; selecting a final target building scheme according to the professional scores.

7. The architectural design method according to any one of claims 1 to 6, wherein The method of optimizing the initial building scheme according to sustainable data corresponding to the application scenario to obtain the target building scheme comprises: determining parameter types of the sustainable data in the application scenario; optimizing the initial building scheme according to preset parameters in the parameter types to obtain the target building scheme.

8. A large language model-based architectural design system, characterized by, The building design system comprises: an input module, configured to input description information into a target large language model to obtain an initial building scheme, the description information being used to describe an application scenario; an optimization module, configured to optimize the initial building scheme according to sustainable data corresponding to the application scenario to obtain a target building scheme; an output module, configured to output the target building scheme.

9. A terminal device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the method of any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 9. The computer program is executed by the processor to implement the method of any one of claims 1 to 7.

Images