Information processing device, information processing method, and information processing program

Abstract

[Problem] To solve the problem of appropriately extracting desired information using a trained AI model by inputting a ground-truth example, in view of the difficulty of accurately designating a ground-truth example from data in formats such as images, CAD files, or BIM models. [Solution] To solve the problem, an information processing device or the like is provided, the information processing device including: a first prompt setting unit that sets a desired-information acquisition prompt for acquiring desired information on a desired item from file data including target items; a first sample setting unit that sets ground-truth sample file data that is input to the trained AI model together with the set desired-information acquisition prompt and serves as a ground-truth example for acquiring the desired information; and a ground-truth information designation unit that designates specific information on the data as ground-truth information on a viewer that displays the set ground-truth sample file data.

Description

Information processing device, information processing method, and information processing program

[0001] The present invention relates to an information processing apparatus for acquiring desired information from a file data group containing various information.

[0002] In recent years, the use of large-scale language models (hereinafter sometimes abbreviated as LLM) to acquire predetermined information from a group of file data has become widespread. For example, Patent Literature 1 discloses a technology that uses LLM to collect character information related to predetermined items from a variety of printed materials with no fixed format, such as advertising flyers.

[0003] Patent No. 7430437

[0004] The technology of Patent Document 1 collects one or more pieces of character information relating to one or more specified items from a group of character strings printed on a flyer read by OCR by providing the LLM with prompts including instructions to specify the items to be collected and various other instructions related to the collection.

[0005] In the technology for collecting desired information as disclosed in Patent Document 1, inputting correct examples into the LLM may allow for more suitable extraction of desired information. However, there is a problem in that it is not easy to accurately specify the correct examples from data in formats such as images, CAD, and BIM.

[0006] Therefore, in order to solve the above-mentioned problems, the present invention provides the following information processing device etc. That is, the information processing device includes: a first prompt setting unit that sets a desired information acquisition prompt for acquiring desired information, which is information on a desired item, from file data including a target item, a first sample setting unit that sets correct sample file data that is input to a trained AI model together with the set desired information acquisition prompt and serves as a correct example for acquiring the desired information, and a correct information designation unit that designates specific information in the set correct sample file data as correct information on a viewer that displays the set correct sample file data.

[0007] In addition to the above features, the present invention provides an information processing device further having an incorrect information designation unit that designates specific information on the set correct sample file data as incorrect information on a viewer that displays the data.

[0008] In addition to the above features, the present invention provides an information processing device further comprising a correct answer information explanation prompt setting unit that sets a correct answer information explanation prompt that explains the specified correct answer information.

[0009] In addition to the above features, the present invention provides an information processing device that further has a first desired information acquisition unit that inputs the set desired information acquisition prompt, the set correct sample file data, and the specified correct information into the trained AI model, and acquires the desired information from the target file data.

[0010] In addition to the above features, the information processing device further provides that the first desired information acquisition unit further inputs the specified incorrect answer information and acquires the desired information from the target file data.

[0011] In addition to the above features, the first desired information acquisition unit further inputs the set correct information explanation prompt to acquire the desired information from the target file data.

[0012] In addition to the above features, the present invention provides an information processing device further comprising a detailed prompt setting unit that sets a detailed prompt that is a detailed prompt for acquiring desired information in the set desired information acquisition prompt.

[0013] In addition to the above features, the present invention provides an information processing device further comprising a verification unit that verifies the acquired desired information.

[0014] In addition to the above features, the present invention provides an information processing device that further has an acquisition script generation unit that inputs the set desired information acquisition prompt, the set correct sample file data, and the specified correct information into the trained AI model to generate an acquisition script for acquiring the desired information.

[0015] In addition to the above features, the present invention also provides an information processing device further comprising a second desired information acquisition unit that acquires the desired information from the target file data using the generated acquisition script.

[0016] In addition to the above features, the present invention provides an information processing device further comprising an acquisition script storage unit that stores the generated acquisition script.

[0017] In addition to the above features, the present invention provides an information processing device that further has an acquisition script evaluation unit that evaluates the desired information acquired from file data containing the target item by the generated acquisition script in comparison with the specified correct answer information.

[0018] In addition to the above features, the present invention provides an information processing device in which the acquisition script generation unit further inputs the evaluation result by the acquisition script evaluation unit to generate the acquisition script.

[0019] In addition to the above features, the present invention also provides an information processing device that further has an acquisition script correction unit that inputs the generated acquisition script and some or all of the file data including the target items into the trained AI model and corrects the generated acquisition script to adapt it to the target file data.

[0020] In addition to the above features, the information processing device further provides an information processing device in which the acquisition script correction unit inputs a desired information acquisition prompt that is the basis of the generated acquisition script into the trained AI model to correct it.

[0021] In addition to the above features, the present invention provides an information processing device further comprising a detailed prompt setting unit that sets a detailed prompt that is a detailed prompt for acquiring desired information in the set desired information acquisition prompt.

[0022] In addition to the above features, the present invention provides an information processing device further comprising a verification unit that verifies the acquired desired information.

[0023] In addition to the above features, the information processing device further provides an information processing device in which the acquisition script correction unit corrects the acquisition script by reflecting the verification result by the verification unit.

[0024] An information processing device is provided that has: a second prompt setting unit that sets a prompt generation prompt for generating an acquisition prompt for acquiring desired information, which is information related to a desired item, from file data including a target item; a second sample setting unit that sets correct sample file data that is input to a trained AI model together with the set prompt generation prompt and serves as a correct example for acquiring the desired information; and a correct information designation unit that designates specific information on the set correct sample file data as correct information on a viewer that displays the set correct sample file data.

[0025] In addition to the above features, the present invention provides an information processing device that further has an acquisition prompt generation unit that inputs the set prompt generation prompt, the set correct answer sample file data, and the specified correct answer information into the trained AI model to generate the acquisition prompt.

[0026] In addition to the above features, the present invention provides an information processing device that further has a third desired information acquisition unit that inputs the generated acquisition prompt into the trained AI model and acquires the desired information from the target file data.

[0027] In addition to the above features, the present invention provides an information processing device further comprising an acquisition prompt storage unit that stores the generated acquisition prompt.

[0028] In addition to the above features, the present invention provides an information processing device further having an acquisition prompt evaluation unit that compares and evaluates the results of acquiring the desired information for the correct sample file data using the generated acquisition prompt with the correct information.

[0029] In addition to the above features, the information processing device provides an information processing device in which the acquisition prompt generation unit generates an acquisition prompt by reflecting the evaluation result by the acquisition prompt evaluation unit.

[0030] In addition to the above features, the third desired information acquisition unit further inputs the prompt generation prompt, the correct answer sample file data, and the correct answer information input to generate the generated acquisition prompt, to acquire the desired information.

[0031] In addition to the above features, the present invention also provides an information processing device further comprising a detailed prompt setting unit that sets a detailed prompt, which is a detailed prompt for generating an acquisition prompt from the set prompt generating prompt.

[0032] In addition to the above features, the present invention provides an information processing device further comprising a verification unit that verifies the acquisition result acquired by the third desired information acquisition unit.

[0033] In addition to the above features, the information processing device provides an information processing device in which the acquisition prompt generation unit generates an acquisition prompt by reflecting a result of the verification by the verification unit.

[0034] Also provided is an information processing method executed by an information processing device, the information processing method comprising: a first prompt setting step of setting an acquisition prompt for acquiring desired information, which is information relating to a desired item, from file data including the target item; a first sample setting step of setting correct sample file data that is input to a trained AI model together with the set acquisition prompt and serves as a correct example for acquiring the desired information; and a correct information designation step of designating specific information on the set correct sample file data as correct information on a viewer that displays the set correct sample file data.

[0035] The present invention also provides an information processing program that causes an information processing device to execute the following steps: a first prompt setting step for setting an acquisition prompt for acquiring desired information, which is information relating to a desired item, from file data including the target item; a first sample setting step for setting correct sample file data that is input to a trained AI model together with the set acquisition prompt and serves as a correct example for acquiring the desired information; and a correct information designation step for designating specific information on the set correct sample file data as correct information on a viewer that displays the set correct sample file data.

[0036] Also provided is an information processing method executed by an information processing device, the information processing method comprising: a second prompt setting step of setting a prompt generation prompt for generating an acquisition prompt for acquiring desired information, which is information relating to a desired item, from file data including the target item; a second sample setting step of setting correct sample file data that is input to a trained AI model together with the set prompt generation prompt and serves as a correct example for acquiring the desired information; and a correct information designation step of designating specific information on the set correct sample file data as correct information on a viewer that displays the set correct sample file data.

[0037] In addition, an information processing program is provided that causes an information processing device to execute the following steps: a second prompt setting step of setting a prompt generation prompt to generate an acquisition prompt for acquiring desired information, which is information related to a desired item, from file data including the target item; a second sample setting step of setting correct sample file data that is input to a trained AI model together with the set prompt generation prompt and serves as a correct example for acquiring the desired information; and a correct information designation step of designating specific information on the set correct sample file data as correct information on a viewer that displays the set correct sample file data.

[0038] The present invention makes it easy to accurately specify preferred correct examples for outputting desired information from a trained AI model.

[0039] FIG. 1 is a block diagram showing an example of the functional configuration of an information processing device of Example 1; FIG. 2 is a conceptual diagram showing an example of a viewer that displays correct sample file data; FIG. 3 is a conceptual diagram showing an example of the configuration of hardware that realizes the information processing device of Example 1; FIG. 4 is a flow diagram that simply shows an example of the processing flow of the information processing device of Example 1; FIG. 5 is a block diagram showing an example of the functional configuration of an information processing device of Example 2; FIG. 6 is a block diagram showing an example of the functional configuration of an information processing device of Example 3;

[0040] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. However, the present invention should not be limited to these embodiments and can be embodied in various forms without departing from the spirit and scope of the present invention.

[0041] <Example 1> <Overview> The present invention targets various file data groups and, when acquiring desired information, displays sample file data in a viewer that serves as a correct example to be input to a trained AI model along with prompts, etc., and enables the correct information to be specified by specifying specific information on the data. This makes it possible to easily specify the correct information.

[0042] The following describes the functions and processing flow of the information processing device, as well as the hardware. The functional blocks of the system described below can be implemented as a combination of hardware and software. Specifically, if a computer is used, these include hardware components such as a central processing unit (CPU), main memory, bus, or secondary storage device (such as a hard disk drive, nonvolatile memory, or storage media such as CDs or DVDs, and their reader drives), input devices used for information input, printers, display devices, and other external peripheral devices, as well as interfaces for those external peripheral devices, communication interfaces, driver programs and other application programs for controlling the hardware, and user interface applications. The CPU processes data input from input devices and other interfaces and stored in memory or on a hard disk in accordance with a program loaded in main memory, processing and storing the data, and generating commands to control the hardware and software. Alternatively, the functional blocks of the system can be implemented using dedicated hardware.

[0043] Furthermore, this invention can be realized not only as a system but also as a method. Furthermore, a part of such an invention can be configured as software. Furthermore, a program used to cause a computer to execute such software, and a recording medium on which the program is fixed, are naturally included within the technical scope of this invention (the same applies throughout this specification).

[0044] 1 is a block diagram showing an example of the functional configuration of an information processing device according to this embodiment. As shown in FIG. 1, the information processing device 100 includes a first prompt setting unit 101, a first sample setting unit 102, a correct answer information designation unit 103, and a first desired information acquisition unit 104.

[0045] In this embodiment, the trained AI model is a machine learning model that has been trained in advance to handle general-purpose tasks. A typical example is the well-known LLM, but it may also be a different model such as a small language model or a multimodal language model that can handle images and other formats.

[0046] <First Prompt Setting Unit> The first prompt setting unit 101 has a function of setting a desired information acquisition prompt for acquiring desired information, which is information related to a desired item, from file data including the target item. The target file data may be stored in a storage device provided in the information processing device, or may be stored in a storage device connected to the information processing device via a communication line or the like. The file data may be in a variety of file formats, including, but not limited to, Word files, Portable Document Format (PDF) files, Excel files, image files, Comma Separated Values ​​(CSV) files, text files, Building Information Modeling (BIM) files, 2D Computer Aided Design (2D CAD) files, and 3D model files.

[0047] Furthermore, "desired information" means information on a desired item or information obtained by performing some kind of calculation or processing on the information on that item (total number, average information, maximum information, coordinate information, spatial relationships such as distance and collision, three-dimensional model information, three-dimensional graphic information, an image, a specific part of an image, the content of the image, text information, etc.).

[0048] For example, if the "total number of elevators" is desired from a group of file data containing CSV files related to the facilities of multiple buildings, the number of elevators in each building included in each file is extracted, and the total number of elevators is the desired information. In addition to numerical calculations, the desired information may also be the results of processing such as three-dimensional spatial processing or two-dimensional graphic processing. For example, if the "collision points between columns and beams" are desired from three-dimensional model files of multiple buildings, all column and beam models included in each file are extracted, the presence or absence of collisions is calculated, and the three-dimensional spatial information obtained by extracting the collision points is the desired information.

[0049] Furthermore, for example, if you are looking for "inconsistencies between elevations and floor plans" in two-dimensional drawing files of multiple buildings, you can extract all of the architectural design information for elevations and floor plans contained in each file, check whether there are any inconsistencies when they are reconstructed into three-dimensional solids, and then document the inconsistencies into text, which becomes the desired information.

[0050] The desired information acquisition prompt can include not only instructions to the trained AI model, but also conditions for guiding the model to an appropriate answer and examples to refer to. For example, when setting a desired information acquisition prompt that instructs the model to extract a "column object" that is part of a BIM model, the desired information acquisition prompt can be set to a sentence such as, "Please extract the value of 'column object'. 'Column object' is part of a building's BIM model and has a long, three-dimensional structure that extends vertically from the ground."

[0051] Furthermore, the desired information acquisition prompt is not limited to one, and multiple prompts can be set. For example, in addition to the instruction to extract "column objects" mentioned above, instructions to extract "beam objects" and "elevator type and number" can also be set. For "beam object" extraction, the prompt can be set using a sentence such as, "Please extract the value of 'beam object'. 'Beam object' is part of the building's BIM model, and has a three-dimensional structure that is horizontally long and located more than 2 meters from the ground." For "elevator type and number," the prompt can be set using a sentence such as, "Please extract the elevator type and number in the BIM model. If there are no applicable entries, leave blank."

[0052] <First sample setting unit> The first sample setting unit 102 has a function of setting correct sample file data that is input to the trained AI model together with the set desired information acquisition prompt and serves as a correct example for acquiring the desired information.

[0053] The correct sample file data is file data that serves as a model when obtaining desired information from the target file data, and contains correct information that serves as a correct example. For example, if the desired information is the illuminance distribution in a room, BIM data that represents the room where lighting and other equipment are installed would be the correct sample file data. Such correct sample file data and the lighting represented therein are the correct information.

[0054] In addition, not only may one correct answer information be set for multiple target correct answer sample file data, but also may individual correct answer information be set for multiple correct answer sample file data. For example, if the correct answer sample file data is a 3D model file for multiple architectural structures including information on columns and beams, and the desired information is the "collision points between columns and beams" in each architectural structure, the actual number of "collision points between columns and beams" in each architectural structure is not limited to "1," and each collision point between a column and a beam is specified as correct answer information for each correct answer sample file data. Note that the correct answer information in this case may be coordinate information on a 3D image showing the collision points, or may be image information that clarifies the collision points by, for example, showing them with a surrounding line.

[0055] <Correct answer information designation unit> The correct answer information designation unit 103 has a function of designating specific information in the data as correct answer information on a viewer that displays the set correct answer sample file data. The operation of a user designating a three-dimensional space model, two-dimensional vector data, image, table data, numerical values, text, etc. that will become correct answer information requires operational expertise, but by providing a viewer as in this configuration, it is advantageous to be able to easily designate a three-dimensional model, three-dimensional space, parts of a two-dimensional figure, an area of ​​an image, a cell in table data, etc. that will become correct answer information.

[0056] Figure 2 is a conceptual diagram showing an example of a viewer displaying correct sample file data. As shown in Figure 2(a), the viewer 201 displays BIM data representing a room in which lighting and other equipment are arranged as correct sample file data. This BIM data includes various information about the dimensions and materials of the displayed room, as well as objects such as each piece of equipment (ceiling light 202, bracket light 203, bench 204, counter 205), and each of these can be extracted.

[0057] Here, if the desired information is the illuminance distribution in a room and the ceiling light 202 is to be specified as the correct information from this correct sample file data, for example, as shown in FIG. 2( b), the ceiling light 202 can be selected by encircling it with an encircling line 206, thereby specifying it as the correct information. The bracket light 205 can also be similarly specified as the correct information. By specifying the correct information in this way, information such as the illuminance and position of each of the ceiling light and bracket light is identified, and this becomes the correct information for obtaining the illuminance distribution in the room. Note that the correct information in this case can be specified not only by encircling it, but also by directly selecting and specifying the 3D object.

[0058] The correct answer information can be specified in various ways depending on the correct answer sample file data. For example, if the correct answer sample file data is an image file, by selecting a specific range in a viewer that opens the image file, that range can be used as the correct answer information to be extracted. Also, by selecting a specific layer in a viewer that opens a CAD file, that layer can be used as the correct answer information to be extracted.

[0059] It is also possible to specify content that does not explicitly exist in the data. For example, by opening a BIM file in a viewer and selecting the space between specific objects in the viewer, the distance between those objects can be used as the correct information to be extracted. Also, by opening spreadsheet data such as Excel in a viewer and selecting a specific cell, that cell can be used as the correct information to be extracted.

[0060] As described above, the specified correct answer information may be in any format, such as numerical information, a three-dimensional spatial model, a two-dimensional figure, an image, or text. Furthermore, the result of a calculation process may be specified as the correct answer information. Furthermore, the correct answer information may not only provide information on the desired final result, but also correct answer information for the processing steps leading up to that result. For example, in the example of extracting "collision points between columns and beams" from the aforementioned three-dimensional model, the correct answer information for the processing steps may include (1) the column and beam models included in each file, (2) whether or not there is a collision between them, and (3) information on the three-dimensional space from which the collision points are extracted.

[0061] <Incorrect Information Designation Unit> The incorrect information designation unit has a function of designating specific information in the data as incorrect information on a viewer that displays the set correct sample file data. In 3D models, there are often objects that are confusable with objects in the correct information. For example, when extracting an "air conditioning duct," objects such as "water supply and drainage piping" and "wiring rack" may be placed near the object, which are similar objects. In this case, by designating not only "air conditioning duct" as the correct answer but also "water supply and drainage piping" and "wiring rack" as incorrect answers, the trained AI can avoid excessive detection of piping and accurately obtain the desired information.

[0062] It is also possible that a file containing only incorrect answer information without correct answer information is set as correct answer sample file data.

[0063] <Correct Answer Information Explanation Prompt Setting Unit> The correct answer information explanation prompt setting unit has a function to set a correct answer information explanation prompt that explains the specified correct answer information. 3D models often contain objects that are confusable with the correct answer information object. For example, if you want to extract an "air conditioning duct," objects such as "water supply and drainage piping" and "wiring racks" may be placed near the object, which are similar to each other. In this case, instead of simply specifying "air conditioning duct" as the correct answer information, set a prompt to explain the correct answer information such as, "In this correct answer sample file data, this "air conditioning duct" is the correct answer for the desired information. This is because it is a long, thin cylindrical model and its end is connected to an "air intake vent." This prevents the trained AI from detecting incorrect information and accurately acquires the desired information. Furthermore, if multiple pairs of correct answer sample file data and correct answer information are input to the trained AI model as samples, a correct answer information explanation prompt may be set for each pair.

[0064] Furthermore, when incorrect information is specified by the incorrect information specifying unit, an explanation for this incorrect information may be included in the correct information explanation prompt.

[0065] In many cases, the correct sample file data will naturally be a small number compared to the multiple file data to be acquired by the first desired information acquisition unit, but the opposite is also possible. The correct sample file data may be a part of the multiple file data to be acquired, or it may be separate file data. The correct sample file data and correct information may be set by the user, or may be set in advance by a person who provides the device to the user.

[0066] <First desired information acquisition unit> The first desired information acquisition unit 105 has a function of inputting the set desired information acquisition prompt, the set correct sample file data, and the specified correct information into the trained AI model, and acquiring the desired information from the target file data.

[0067] The target file data is input into the trained AI model, and the extraction target is extracted according to the instructions of the desired information acquisition prompt. Processing such as calculations is performed as necessary, and the desired information is finally acquired. It is also possible to extract and acquire the desired information by further inputting a detailed prompt, which will be described later. It is also possible to extract and acquire the desired information by inputting the accumulated desired information acquisition prompt and detailed prompt.

[0068] When a desired information acquisition prompt that instructs extraction of the above-mentioned "pillar object" is input, the "pillar object" is extracted and can be output together with its file name and a thumbnail image of the file data. The extracted information can then be tagged and entered and saved in an Excel file, or the extracted information can be registered in a database, thereby converting unstructured file data into structured data suitable for use in search, statistical processing, AI learning data, etc.

[0069] In addition to importing the extracted information into an Excel file or the like and creating a database, the extracted information can also be processed according to the format of the file data from which it was extracted. For example, in the case of a CSV file, the extracted information can be added to a new column (row) within the CSV file and saved. In the case of an Excel file, the extracted information can be added to an added column or sheet within the Excel file and saved. In the case of a BIM file, the extracted information can be added to the BIM file and saved.

[0070] In addition, by inputting correct sample file data and correct information in various formats, it is possible to appropriately acquire information from file data in each format. This also applies when acquiring desired information using an acquisition script or acquisition prompt, which will be described later.

[0071] The first desired information acquisition unit may be configured to further input the specified incorrect answer information and acquire the desired information from the target file data. With this configuration, the desired information can be acquired more accurately.

[0072] The first desired information acquisition unit may be further configured to input the correct information explanation prompt set above to acquire the desired information from the target file data. With this configuration, the desired information can be acquired more accurately.

[0073] <Hardware Configuration> Fig. 3 is a conceptual diagram showing an example of the configuration of hardware that realizes the information processing apparatus of Example 1. As shown in the figure, the information processing apparatus 300 has a CPU 301 that performs various arithmetic processing, a RAM 302 that is a volatile storage medium, a storage 303 such as a flash memory or HDD that is a non-volatile storage medium, a communication interface 304, and an input / output interface 305. The RAM 302 reads programs that perform various arithmetic processing to have the CPU 301 execute them, and also provides a work area for the programs. In addition, multiple addresses are assigned to the RAM 302, and programs executed by the CPU 301 can exchange data and perform processing by identifying and accessing these addresses (the same applies throughout this specification).

[0074] Here, the functions of the first prompt setting unit 101, the first sample setting unit 102, and the first desired information acquisition unit 104 of the information processing device 100 in Figure 1 are mainly realized by the CPU 301 and RAM 302 in Figure 3. The function of the correct answer information designation unit 103 is mainly realized by the CPU 301, RAM 302, and input / output interface 305 in Figure 3. When a database storing the target file data exists externally or when the first desired information acquisition unit uses an external trained AI model, the functions are realized by mutually exchanging signals and information via the communication interface 304 and the input / output interface 305. The functions of the incorrect answer information designation unit and the correct answer information explanation prompt setting unit, which are not shown in Figure 1, are realized by the CPU 301, RAM 302, and input / output interface 305.

[0075] <Processing Flow> FIG. 4 is a flow diagram illustrating a simplified example of the processing flow of the information processing device according to the first embodiment. First, a desired information acquisition prompt is set to acquire desired information, which is information related to a desired item, from file data containing a target item (S401: first prompt setting step). Then, correct sample file data is set as a correct example for acquiring the desired information, which is input to the trained AI model together with the set desired information acquisition prompt (S402: first sample setting step). Then, the set correct sample file data is displayed in a viewer (S403: correct sample file data display step). Then, specific information in the displayed data is designated as correct information (S404: correct information designation step). Then, the set desired information acquisition prompt, the set correct sample file data, and the designated correct information are input to the trained AI model to acquire the desired information from the target file data (S405: first desired information acquisition step). Note that the correct information designation step may include a correct sample file data display step. The method may also include an incorrect answer information specifying step and a correct answer information explanation prompt setting step.

[0076] <Effects> According to the information processing device of this embodiment, it is possible to easily and accurately specify a suitable correct example for outputting desired information from a trained AI model.

[0077] <Example 2> <Overview> In the information processing device of Example 1, instead of inputting a desired information acquisition prompt together with correct answer information, etc. into a trained AI model to acquire desired information, the present invention acquires desired information by executing a script (simple program) generated by inputting a desired information acquisition prompt and correct answer information, etc. into a trained AI model.

[0078] <Functional Configuration> Fig. 5 is a block diagram showing an example of the functional configuration of the information processing device of this embodiment. As shown in Fig. 5, the information processing device 500 has a first prompt setting unit 501, a first sample setting unit 502, a correct answer information designation unit 503, an acquisition script generation unit 504, an acquisition script storage unit 505, and a second desired information acquisition unit 506. Each component of the first prompt setting unit, the first sample setting unit, and the correct answer information designation unit has the same function as the components with the same names in Example 1, and therefore description thereof will be omitted.

[0079] <Acquisition script generation unit> The acquisition script generation unit 504 has the function of inputting the set desired information acquisition prompt, the set correct sample file data, and the specified correct information into the trained AI model, and generating an acquisition script for acquiring the desired information.

[0080] As explained in Example 1, the desired information acquisition prompt is a prompt for acquiring the desired information, but in this example, it is input into a trained AI model along with a correct answer example to generate an acquisition script, and this acquisition script is executed to become a prompt for acquiring the desired information. For example, "We will now provide multiple BIM files as samples, each showing a room in which various pieces of equipment are located. Please create a script that extracts various types of lighting in the room from these BIM files and calculates the resulting lighting distribution in each room," which instructs the generation of an acquisition script for calculating the desired information, i.e., the lighting distribution in the room.

[0081] Furthermore, if the process for acquiring desired information using an acquisition script becomes complicated, the desired information acquisition prompt may be set to generate an acquisition script divided into multiple steps. For example, in the example of extracting "collision points between columns and beams" from a three-dimensional model, the process may be divided into (1) a desired information acquisition prompt that generates an acquisition script that extracts the column and beam models included in each file, (2) a desired information acquisition prompt that generates an acquisition script that checks whether or not there is a collision between them, and (3) a desired information acquisition prompt that generates an acquisition script that acquires information about the three-dimensional space where the collision points have been extracted, and the desired information acquisition prompt may be set to generate an acquisition script for each step.

[0082] <Acquisition Script Storage Unit> The acquisition script storage unit 505 has the function of storing the generated acquisition script. By having an acquisition script storage unit, it is possible to reuse an acquisition script once generated, making it possible to reproducibly acquire values ​​from a large amount of target file data. In addition, the generated acquisition script may be stored in association with the acquisition results acquired by the second desired information acquisition unit (described later) using the generated acquisition script, as well as the original desired information acquisition prompt and correct sample file data that were input together.

[0083] <Second Desired Information Acquisition Unit> The second desired information acquisition unit 506 has a function of acquiring the desired information from the target file data using the generated acquisition script. The second desired information acquisition unit 506 executes the generated acquisition script on the target file data to acquire the desired information. It is preferable to save the acquired results. This is because, as will be described later, the acquired results can be used for post-processing, such as verifying the acquired results.

[0084] In addition, the desired information acquired by the second desired information acquisition unit may be passed to another acquisition script as new target file data. In other words, the generation of the acquisition script and the acquisition of desired information from the target file data in this invention can be linked, making it possible to configure more complex processing.

[0085] As described above, the process of obtaining the desired information is performed by executing the acquisition script, which reduces the time and cost required compared to calling up a trained AI model each time to obtain the desired information, and also enables the extraction of reproducible information with a certain level of extraction accuracy.

[0086] Furthermore, there was a problem that "even if we tried to prepare a static script to solve this problem, only engineers with specialized knowledge could create the script themselves." However, by generating an acquisition script using a trained AI model, it is possible to achieve the effect that anyone can create an acquisition script themselves.

[0087] <Hardware Configuration> The information processing apparatus of this embodiment can be realized by the hardware configuration shown in Fig. 3 according to the first embodiment. The functions of the acquisition script generation unit 504 and the second desired information acquisition unit 506, which are components unique to the information processing apparatus 500 of the second embodiment shown in Fig. 5, are mainly realized by the CPU 301 and RAM 302 in Fig. 3. Furthermore, the function of the acquisition script accumulation unit 505 is mainly realized by the storage 303 in Fig. 3.

[0088] <Processing Flow> The processing flow of the information processing device of this embodiment is basically the same as the processing flow of the information processing device of Example 1 up to the correct information specification step (S404). The processing flow further includes an acquisition script generation step of inputting the set desired information acquisition prompt, the set correct sample file data, and the specified correct information into the trained AI model to generate an acquisition script for acquiring the desired information, an acquisition script accumulation step of accumulating the generated acquisition script, and a second desired information acquisition step of acquiring the desired information from the target file data using the generated acquisition script. The acquisition script accumulation step may be performed after the second desired information acquisition step.

[0089] <Effects> The information processing device of this embodiment makes it easy to accurately specify correct examples, and reduces the time and cost required compared to calling up a trained AI model each time to obtain desired information.

[0090] <Example 3> <Overview> The present invention generates prompts (hereinafter referred to as acquisition prompts) to be input to a trained AI model to acquire desired information from various file data groups by inputting correct sample file data that is a correct example for acquiring the desired information and correct information together with the prompt into the trained AI model. This allows even an inexperienced user to appropriately generate acquisition prompts.

[0091] <Functional Configuration> Fig. 6 is a block diagram showing an example of the functional configuration of the information processing device of this embodiment. As shown in Fig. 6, the information processing device 600 has a second prompt setting unit 601, a second sample setting unit 602, a correct answer information designation unit 603, an acquisition prompt generation unit 604, an acquisition prompt storage unit 605, and a third desired information acquisition unit 606.

[0092] <Second Prompt Setting Unit> The second prompt setting unit 601 has a function of setting a prompt generation prompt for generating an acquisition prompt for acquiring desired information, which is information about a desired item, from file data including the target item. The desired information in this configuration is as described in the first and second embodiments.

[0093] An acquisition prompt is generated by inputting the prompt generation prompt set in the second prompt setting unit and the correct answer sample file data and correct answer information described later into the acquisition prompt generation unit.The prompt generation prompt set here may be set in a form that specifies the desired information and generates an acquisition prompt by referring to the correct answer sample file data, or it may be set in a form that generates a prompt that acquires correct answer information from the correct answer sample file data.

[0094] For example, if the desired information is "illuminance distribution in the room" as in the example above, a prompt generation prompt can be set such as, "Please refer to the sample file data and correct answer information below to extract all indoor lighting for each of the multiple BIM files, and generate a prompt (acquisition prompt) to obtain the indoor illuminance distribution from the information about each lighting."

[0095] Furthermore, if the process for acquiring desired information is complicated, the prompt generation prompt may be set to generate an acquisition prompt in multiple steps. For example, in the example of acquiring the indoor illuminance distribution described above, the process may be divided into (1) a prompt for generating an acquisition prompt that extracts information necessary for calculating the illuminance distribution from information about all lighting included in each file, (2) a prompt for generating an acquisition prompt that calculates the indoor illuminance distribution using the extracted information, and so on, and the prompt generation prompt may be set to generate an acquisition prompt for each step.

[0096] <Second Sample Setting Unit> The second sample setting unit 602 has a function of setting correct sample file data that is input to the trained AI model together with the set prompt generation prompt and serves as a correct example for acquiring the desired information. The correct sample file data in this configuration is the same as the correct sample file data in Examples 1 and 2.

[0097] <Correct answer information designation unit> The correct answer information designation unit 603 has a function of designating specific information on the data as correct answer information on the viewer that displays the correct answer sample file data that has been set. The function of this configuration is the same as that of the correct answer information designation unit in the first and second embodiments.

[0098] Furthermore, the system may be configured to include the incorrect answer information designation unit and correct answer information explanation prompt setting unit described in the first embodiment.

[0099] <Acquisition prompt generation unit> The acquisition prompt generation unit 604 has the function of inputting the set prompt generation prompt, the set correct answer sample file data, and the specified correct answer information into the trained AI model to generate the acquisition prompt.

[0100] For example, in the case of an acquisition prompt generated to determine the illuminance distribution described above, by inputting the above-mentioned prompt generation prompt, a BIM file such as that described in Example 1, and the correct answer example specified in that file into a trained AI model, an acquisition prompt such as, "Extract all indoor lighting for each of the multiple BIM files, and generate a prompt to determine the indoor illuminance distribution from the information about each lighting. Please set the height of the work surface 1 mm higher than the height of the objects in the room." is generated.

[0101] In addition, generating a capture prompt is not limited to generating a prompt alone, but may also involve generating a static script at the same time, in which case the static script can stably and quickly obtain the desired information from the target file data.In addition, the generated prompt may call a script, or the script may call a prompt.

[0102] In addition, the incorrect information specified by the incorrect information specification unit and the correct information explanation prompt set by the correct information explanation prompt setting unit may be further input into the trained AI model to generate an acquisition prompt.

[0103] <Acquisition Prompt Storage Unit> The acquisition prompt storage unit 605 has a function of storing the generated acquisition prompt. By having an acquisition prompt storage unit, it is possible to reuse an acquisition prompt once generated, making it possible to reproducibly acquire values ​​from a large amount of target file data. In addition, the generated acquisition prompt may be stored in association with the acquisition results acquired by the third desired information acquisition unit (described later), the original prompt generation prompt, correct sample file data, and correct answer information input for generation.

[0104] <Third Desired Information Acquisition Unit> The third desired information acquisition unit 606 has a function of inputting the generated acquisition prompt into the trained AI model and acquiring the desired information from the target file data. It is preferable to store the acquired results. This is to provide the acquired results for post-processing, such as verifying or reformatting the acquired results, as described below. The trained AI model used by the third desired information acquisition unit may be the same as or different from the trained AI model used by the acquisition prompt generation unit.

[0105] The third desired information acquisition unit may be configured to acquire the desired information by further inputting the prompt generation prompt, the correct sample file data, and the correct answer information that were input to generate the generated acquisition prompt. Alternatively, the third desired information acquisition unit may be configured to acquire the desired information by further inputting the specified incorrect answer information and the set correct answer information explanation prompt. This configuration makes it easy to generate an acquisition prompt suitable for acquiring the desired information.

[0106] In addition, the desired information acquired by the third desired information acquisition unit may be passed to another process using an acquisition prompt as new target file data. In other words, the generation of the acquisition prompt and the acquisition of desired information from the target file in this invention can be linked, making it possible to configure more complex processes.

[0107] <Hardware Configuration> The information processing apparatus of Example 3 can be realized by hardware that realizes the information processing apparatus of Example 1 shown in Figure 3. That is, the functions of the second prompt setting unit 601, second sample setting unit 602, correct answer information designation unit 603, acquisition prompt generation unit 604, and third desired information acquisition unit 606 of the information processing apparatus 600 of Figure 6 are mainly realized by the CPU 301 and RAM 302 of Figure 3. The function of the acquisition prompt accumulation unit 605 is mainly realized by the storage 303 of Figure 3. In addition, when a database that accumulates the target file data is located externally, or when the acquisition prompt generation unit or the third desired information acquisition unit uses a trained AI model that is located externally, each function is realized by mutual exchange of signals and information via the communication interface 304 and the input / output interface 305.

[0108] <Processing Flow> FIG. 7 is a flow diagram illustrating a simplified example of the processing flow of the information processing device according to the third embodiment. First, a prompt generation prompt is set to generate an acquisition prompt for acquiring desired information, which is information related to a desired item, from file data containing a target item (S701: second prompt setting step). Then, correct sample file data is set as a correct example for acquiring the desired information, which is input to the trained AI model together with the set prompt generation prompt (S702: second sample setting step). Then, the set correct sample file data is displayed in a viewer (S703: correct sample file data display step). Then, specific information in the displayed data is designated as correct information (S704: correct information designation step). Then, the set prompt generation prompt, the set correct sample file data, and the designated correct information are input to the trained AI model to acquire the desired information from the target file data (S705: third desired information acquisition step). Note that the correct information designation step may include a correct sample file data display step.

[0109] <Effects> According to the information processing device of this embodiment, it is easy to accurately specify correct examples and also makes it easy for even inexperienced users to create suitable prompts for extracting desired information.

[0110] <Additional Functions> Functions that can be added to any of the above-described Examples 1 to 3 will be described below. First, functions that can be added to Example 1 will be described. If the functions can also be added to Example 2 or Example 3, this will be explained.

[0111] <Detailed Prompt Setting Unit> The detailed prompt setting unit has a function of setting a detailed prompt, which is a detailed prompt for acquiring desired information in the set desired information acquisition prompt.

[0112] 8 is a diagram showing an example of a detailed prompt. As shown in the figure, in this example, the desired information acquisition prompt is "Please extract the value of 'Drawing Number'." The setting items for setting the detailed instructions of this prompt are "No.", "Name.", "Description.", "Necessity.", and "Example."

[0113] For example, setting item 1 is called "Assigning Multiple Tag Names," and is indicated as a setting item for "dealing with variations in character notation." The explanation for this setting is "explicitly assign multiple data item names to be extracted," with an example being "["Drawing Number," "Drawing ID"]." By setting this detailed prompt, an instruction is added to the instruction "Please extract the value of 'Drawing Number,'" to include not only "Drawing Number," but also "Drawing ID" as a target for extraction.

[0114] Other advanced prompt settings include "Tag Description," which adds a description of the data item name to be extracted to address inconsistencies in character notation; "Absolute Position Specification," which specifies the range on an image where a data item exists to narrow the extraction range; "Target Page Specification," which specifies the page number of a data item in a multi-page PDF; "Sample Specification," which specifies samples close to or far from the extracted value to prevent incorrect extraction; "Format Specification," which allows users to specify the format they want to extract to prevent incorrect extraction and standardize the format of output results; "Relative Position Specification," which allows users to specify the relative position of some content to support table formats; and "Condition Settings," which allows users to change the extracted content based on specific conditions if the desired content is not available in all files. Of course, users can also configure various settings themselves to configure advanced prompt settings.

[0115] Here, the application methods for each of the above-mentioned setting items can be broadly divided into three: Method 1, "narrowing down file contents," Method 2, "connecting prompts," and Method 3, "checking and reshaping detection results." Method 1 involves narrowing down the file data to be input before inputting it into the trained AI model, while Method 2 involves connecting the set detailed prompt to the main prompt. Method 3 involves reshaping the extraction results according to the specified sample, format, etc.

[0116] FIG. 9 shows how each setting item in the detailed prompt is applied. As shown in the figure, "Connecting prompts" of Method 2 is applied to the items "Assigning multiple tag names," "Assigning tag descriptions," and "Setting conditions." Furthermore, "Specifying absolute position" and "Specifying target page" are applied to "Narrowing file contents" of Method 1 and Method 2. Furthermore, "Specifying sample" is applied to "Confirming extracted results and reformatting" of Method 2 and Method 3. Furthermore, "Specifying format" and "Specifying relative position" are applied to all of Methods 1 to 3. Examples of application are described below.

[0117] FIG. 10 is a conceptual diagram showing an example of an input screen when setting a detailed prompt. As shown in the figure, a desired information acquisition prompt 1002 that has already been set is displayed in the upper left part of the screen 1001. Input fields 1003 for the setting items "tag name," "absolute position specification," "sample specification," and "relative position specification" are displayed on the left side of the screen. Furthermore, input fields for other setting items are displayed in response to a user operation. A detailed prompt is set by accepting a user's entry operation in each of these input fields. The first desired information acquisition unit then applies the setting of the detailed prompt and performs extraction. By setting and extracting a detailed prompt in this manner, extraction accuracy can be improved.

[0118] An example of applying the above detailed prompt is shown below. For example, if the "Sample Specification" field is set to "Similar content to D-00033, F-00102" and the "Relative Position Specification" field is set to "This item is located in the bottom left corner of the page," this detailed prompt is connected to the desired information acquisition prompt and input. Then, by applying "Relative Position Specification," only the bottom left corner of the page is input for the file data to be input into the trained AI model (Method 1: Narrowing down the file contents). Then, by applying "Sample Specification," the extracted "Drawing Number" is extracted, and drawing numbers similar to "D-00033, F-00102" are extracted (Method 3: Check the extracted results and reformat). This reformat can be performed using the trained AI model or by inputting a separate script that instructs the reformat.

[0119] Furthermore, when setting the detailed prompt, the detailed prompt setting unit can be configured to automatically set each setting item according to the set desired information acquisition prompt. Alternatively, the detailed prompt setting unit may be configured to accept the user's settings while recommending some setting for items that have not been set. Accepting the detailed prompt settings from the user allows information to be acquired flexibly and appropriately in accordance with the user's wishes, but it may be difficult for the user to set the optimal prompt. Therefore, functions such as automatic entry and suggestion of detailed prompts are effective. These functions can simplify the process and improve operability.

[0120] Furthermore, for simplification and improved operability, the desired information acquisition unit is preferably configured to perform a test acquisition using the set desired information acquisition prompt or using the set desired information acquisition prompt and detailed prompts before acquiring all of the target file data. A test acquisition is an acquisition of a portion of the target file data. Based on the results of this test acquisition, the user can determine whether the set desired information acquisition prompt and detailed prompt will achieve the desired acquisition results, and can reconfigure the prompts to obtain more desirable acquisition results.

[0121] 10, a "Test" button 1005 is displayed on the left side of a screen 1001, below an input field 1003 for a detailed prompt and a display field 1004 for a desired information acquisition prompt, and test extraction is performed by clicking this button. The results of the test extraction are then displayed in a dotted-line frame 1006, which is shown for explanatory purposes in the figure. Based on the extraction results, the user can reset the prompt or perform actual extraction using the prompt for which the test extraction results were obtained.

[0122] The detailed prompt setting unit can also be applied as an additional configuration in Example 2 and Example 3. For example, in Example 2, the detailed prompt setting unit can be applied as a configuration for setting the above-mentioned detailed prompt as a detailed prompt for acquiring desired information in the set desired information acquisition prompt. Also, in Example 3, the detailed prompt setting unit can be applied as a configuration for setting the above-mentioned detailed prompt as a detailed prompt for generating an acquisition prompt using the set prompt generation prompt.

[0123] <Verification Unit> The verification unit has the function of verifying the acquired desired information. There are various specific modes of verification by the verification unit. FIG. 11 is a diagram illustrating various settings when performing verification. As shown in the figure, four types of settings can be made, namely, "Use detailed prompt," "Verification prompt," "Verification script," and "Compare between files."

[0124] "Use advanced prompt" uses the extraction settings set in the advanced prompt for verification. For example, if a negative sample is specified in the "Specify sample" setting of the advanced prompt, the specified negative sample is reused to verify whether the extraction result is similar to the negative sample. Also, the setting in "Specify relative position" is used to verify whether the word "roof" is to the left of the extracted value.

[0125] The "verification prompt" is used to individually set a prompt for verification, for example, by setting and inputting a prompt such as "Please check whether this value indicates the material of the elevator door." The "verification script" is used to individually set a script for verification, for example, as shown in the figure, by inputting a command to return "Error" if the extraction result is less than "1" and "OK" otherwise.

[0126] Furthermore, when extracting from multiple files, "file comparison" compares the extracted results with other files to confirm their validity. As shown in the example, when comparing extracted files A to D, files A, C, and D all contain only numbers and symbols, while file B is primarily written in kanji. In such a case, the verification result is that "file B is not valid compared to the others."

[0127] Here, when the verification unit verifies the obtained result, the first desired information acquisition unit can be configured to reflect the verification result and perform re-acquisition. For example, if the value of the extracted result is an error as described above, the unit inputs an instruction such as "The value XXX was extracted in the previous extraction, but the verification result was an error. Please consider this error result and extract the value again," and then acquires the desired information again.

[0128] The verification unit can also be applied as an additional configuration in Example 2 and Example 3. In Example 2, it can be applied as a configuration for verifying the desired information acquired by the second desired information acquisition unit, and in Example 3, it can be applied as a configuration for verifying the desired information acquired by the third desired information acquisition unit.

[0129] Next, functions that can be added to the second embodiment will be described.

[0130] <Acquisition Script Evaluation Unit> The acquisition script evaluation unit has a function of evaluating the desired information acquired from file data including the target item by the generated acquisition script by comparing it with the specified correct answer information. This correct answer information is as described in the first sample setting unit and the correct answer information designation unit of the first embodiment. By actually acquiring the desired information from the correct answer sample file data using the generated acquisition script and comparing the result with the correct answer information, it is possible to evaluate how accurate the acquisition script is. In other words, if the acquired result matches the correct answer information, it can be evaluated as accurate.

[0131] Furthermore, by configuring the acquisition script generation unit to further input the evaluation results into the trained AI model, an acquisition script can be generated based on the evaluation results, which can be useful for generating a more accurate acquisition script. For example, it can be configured to continue regenerating the acquisition script until the acquired information matches the correct information. Note that the evaluation of the acquisition script may be performed on the acquisition script stored in the acquisition script storage unit.

[0132] <Acquisition script correction unit> The acquisition script correction unit has a function of inputting the generated acquisition script and some or all of the target file data into the trained AI model, and correcting the generated acquisition script to adapt it to the target file data. Note that the trained AI model used for the correction may be the same as the trained AI model used in the acquisition script generation unit described above, or may be a different one.

[0133] For example, if the target file data contains file data in a file format that was not included in the sample file data entered when the acquisition script was created, or if the target file data contains item names that should be acquired, the desired information may not be acquired accurately. To prevent this situation, the generated acquisition script is input into the trained AI model along with the target file data.

[0134] The acquisition script is then modified by providing a prompt like the one shown below: "This script was created with the following objective in mind: 'Select column objects from the list of 3D building objects in the BIM file and calculate their total number.' However, it may not function properly with the file data being processed. Identify the malfunctioning parts and modify the script." For example, if an acquisition script created to obtain the "total number of column objects" executes a script that extracts objects with a property value of "Column," and the target file data contains a column object with a property value of "Pillar," the information in this file data will be omitted from the acquisition target, making it impossible to obtain accurate desired information. Therefore, by inputting the created acquisition script and the target file data into the trained AI model along with the prompt shown above, an acquisition script adapted to the target file data can be generated.

[0135] The modification to adapt to the file data may be performed by inputting the file data one by one from the target file data into the trained AI model, or by inputting all the file data into the trained AI model. Furthermore, the desired information acquisition prompt that was the basis for creating the modified acquisition script may also be input.

[0136] Furthermore, the acquisition scripts stored in the acquisition script storage unit can also be modified in the same way as the generated acquisition scripts are modified. Since the stored acquisition scripts may need to be modified due to an increase, decrease, or change in the target file data, it is preferable to configure the system so that they can be modified.

[0137] Furthermore, when the verification unit described as one of the additional functions is applied to Example 2, the acquisition script correction unit can be configured to reflect the verification result by this verification unit and correct the acquisition script. For example, if the information in the acquisition result is an error as described above, a prompt such as "In the previous extraction, information XXX was extracted, but the verification result was an error. Please correct the script taking this error result into consideration" is input to the trained AI model to correct the acquisition script.

[0138] Next, functions that can be added to the third embodiment will be described.

[0139] <Acquisition prompt evaluation unit> The acquisition prompt evaluation unit has a function of comparing the result of acquiring the desired information for the correct sample file data using the generated acquisition prompt with the correct information and evaluating it. In the same way as the acquisition script evaluation unit, which is an additional function of Example 2, in this configuration, the generated acquisition script is evaluated. Similarly, the acquisition prompt generation unit can be configured to further input the evaluation result by the acquisition prompt evaluation unit into the trained AI model, and an acquisition prompt that reflects the evaluation result can be generated.

[0140] Furthermore, functions that can be added to any of the first to third embodiments will be described.

[0141] <Filtering Unit> The filtering unit has a function of removing predetermined file data from the target file data from the objects to be acquired by any of the first desired information acquisition unit to the third desired information acquisition unit. There are various filtering methods. For example, filtering is performed based on file attributes (extension, size, file name, data items associated with the file, etc.). Alternatively, individual keywords are set and filtering is performed based on these keywords. Specifically, filtering is performed by generating a script for filtering by performing semantic search, clustering, character search, etc., and executing the script. Filtering may also be performed by giving instructions to a trained AI model. For example, by inputting an instruction such as "Does this file contain elevator content?", filtering can be performed based on whether or not the file contains elevator content.

[0142] This filtering function can meet user requests, such as, for example, obtaining column objects from structural models among 500 target file data, but skipping equipment models.

[0143] <Conversion Unit> The conversion unit has a function of converting the format of the target file data according to the characteristics of the file data. When the conversion unit is included, any of the first desired information acquisition unit to the third desired information acquisition unit acquires desired information from the file data converted by the conversion unit. For example, if the file data is in PDF format, it is converted into a text file format. In addition, in the case of image data, character recognition is performed using OCR to convert it into a text file format. In addition, three-dimensional data may be converted into two-dimensional data or an image. Performing such conversion contributes to improving robustness.

[0144] Furthermore, when converting to a text file, it is also preferable to add location information to the text. Figure 12 is a conceptual diagram illustrating an example of converting PDF file data into a text file with location information added. Figure 12(a) shows a portion of a PDF file showing a building area table for a house. Figure 12(b) shows an example of converting the text into a JSON format, in which the text for "shape," "formula," and "area" are described along with their location information. Figure 12(c) shows an example of converting the text into an Array format, in which the text is similarly described along with location information. Adding location information in this way allows extraction to take into account the relationships between text, contributing to improved robustness. Figure 12(d) shows an example of converting text that, in the original file, shows a table showing the "area" and "tatami mats" for "Western-style Room 1," "Western-style Room 2," and "Western-style Room 3," into text with table information (within the dotted-line frame in the figure) instead of location information. This conversion allows the table structure to be retained and retrieved.

[0145] Figure 13 is a conceptual diagram showing another example of converting file data formats. As shown in Figure 13, in the case of a file containing graphics (such as a kitchen floor plan), converting it to SVG format and converting it into text containing graphic information makes it easier to retrieve using a script. Conversely, two-dimensional vector information can be converted into a raster image, or three-dimensional data can be converted into two-dimensional data.

[0146] <Data Reduction Unit> The data reduction unit has the function of reducing the target file data. Data reduction can be performed in various ways. For example, data reduction can be performed by narrowing down the target file data to a specific space, property, area, specific page or section, paragraph, table, etc. It is also possible to filter by file attributes (page number, location, etc.), by individual keywords (semantic search, clustering, character search), or by spatial location (x, y, z, coordinates). It is also possible to filter by inputting instructions to a trained AI model ("Does this file contain elevator contents?").

[0147] FIG. 14 is a conceptual diagram illustrating another aspect of data reduction. As shown in FIG. 14( a), a specific portion 1402 is cut out from one file data 1401, and a check is made to see if the target data is included therein. If the target data is included, the portion is subjected to an acquisition process by one of the first desired information acquisition unit to the third desired information acquisition unit. If the target data is not included, a portion 1403 shifted from the previous portion is cut out and similarly checked. By excluding from the acquisition target data in the pre-acquisition check in this way, processing efficiency can be improved. Note that this check may be performed by executing a script, using a trained AI model, or by other methods. Furthermore, such processing may be performed in a three-dimensional space.

[0148] 14B, the above-described check is performed while shifting the specific location, and the location containing the desired data is temporarily stored in memory 1404. Then, it is checked whether processing memory 1404 contains all the elements necessary to acquire the desired data, and if it does, the location stored in memory is subjected to acquisition processing by one of the first desired information acquisition unit to the third desired information acquisition unit. On the other hand, if it does not contain all the elements, the specific location is shifted further, and the series of processes starting from the above-described check are repeated.

[0149] <Basis Display Unit> The basis display unit has a function of displaying the basis for the acquisition result acquired by any of the first to third desired information acquisition units.

[0150] In the information processing device of each embodiment, the desired information is acquired through the respective processes. For example, in Example 2, an acquisition script is generated by inputting a desired information acquisition prompt, correct sample file data, and correct answer information into a trained AI model, and the generated script is input into the trained AI model to acquire the desired information. Here, it may be difficult to set a desired information acquisition prompt to generate an optimal acquisition script for obtaining the desired information. Therefore, by displaying the basis for the acquisition result (such as the intermediate process), it is possible to show the user "what can be done to improve the acquisition result."

[0151] FIG. 15 is a conceptual diagram showing an example of displaying an extraction range on a 3D model as evidence. The illustration shows a 3D model representing a building, and the position of information (pillar objects) acquired from this model is highlighted in 3D by a thick dotted frame 1501. This allows the user to understand where the acquired results came from. For example, in the case of an acquisition script that processes a 3D model, displaying the processing process of the acquisition script and highlighting the information used during that process can also be useful for confirming the evidence.

[0152] Furthermore, a display explaining the reason for the acquisition may be provided. For example, as a pre-processing step, before acquisition, an instruction such as "Please output the reason for outputting this information as the acquisition result for each detailed process" may be input to the trained AI model. Then, after acquisition, the acquired results are displayed, along with the rationale, for example, "Various indoor lights are placed near the ceiling, and the object names contain the character 'lighting,' so these objects were extracted as lighting." In this way, by displaying the desired information acquisition prompt set by the user and the effect of the resulting acquisition script on the acquisition results, the user can reflect on various settings, contributing to the optimization of desired information acquisition. Similarly, in Examples 1 and 3, it is useful to display the rationale and intermediate processes that led to the acquisition of the desired information.

[0153] 100, 500, 600: Information processing device 101, 501: First prompt setting unit 102, 502: First sample setting unit 103, 503, 603: Correct answer information designation unit 104: First desired information acquisition unit 504: Acquisition script generation unit 505: Acquisition script storage unit 506: Second desired information acquisition unit 601: Second prompt setting unit 602: Second sample setting unit 604: Acquisition prompt generation unit 605: Acquisition prompt storage unit 606: Third desired information acquisition unit 301: CPU 302: RAM 303: Storage 304: Communication interface 305: Input / output interface

Claims

a first prompt setting unit that sets a desired information acquisition prompt for acquiring desired information, which is information related to a desired item, from file data that includes the target item; a first sample setting unit that sets correct sample file data that is input to a trained AI model together with the set desired information acquisition prompt and serves as a correct example for acquiring the desired information; a correct answer information designation unit that designates specific information on the data as correct answer information on a viewer that displays the set correct answer sample file data; An information processing device having the above.

2. The information processing apparatus according to claim 1, further comprising an incorrect answer information designation unit that designates specific information on the set correct answer sample file data as incorrect answer information on a viewer that displays the data.

3. The information processing apparatus according to claim 1, further comprising a correct answer information explanation prompt setting unit that sets a correct answer information explanation prompt that explains the specified correct answer information.

2. The information processing device according to claim 1, further comprising a first desired information acquisition unit that inputs the set desired information acquisition prompt, the set correct sample file data, and the specified correct information into the trained AI model to acquire the desired information from the target file data. 　 The information processing apparatus according to claim 4 , wherein the first desired information acquiring unit further inputs the specified incorrect answer information and acquires the desired information from the target file data. 　 The information processing apparatus according to claim 4 , wherein the first desired information acquisition unit further inputs the set correct information explanation prompt to acquire the desired information from the target file data.

3. The information processing apparatus according to claim 1, further comprising a detail prompt setting unit for setting a detail prompt, which is a detailed prompt for acquiring the desired information in the set desired information acquisition prompt. 　 The information processing apparatus according to claim 4 or 5, further comprising a verification unit that verifies the acquired desired information.

2. The information processing device according to claim 1, further comprising: an acquisition script generation unit that inputs the set desired information acquisition prompt, the set correct sample file data, and the specified correct information into the trained AI model to generate an acquisition script for acquiring the desired information. 　 The information processing apparatus according to claim 9 , further comprising a second desired information acquisition unit that acquires the desired information from the target file data using the generated acquisition script.

11. The information providing device according to claim 9, further comprising an acquisition script storage unit that stores the generated acquisition script. 　 The information processing apparatus according to claim 10 , further comprising an acquisition script evaluation unit that evaluates desired information acquired from file data including a target item by the generated acquisition script in comparison with the specified correct answer information. 　 The information processing apparatus according to claim 12 , wherein the acquisition script generation unit further inputs the evaluation result by the acquisition script evaluation unit to generate the acquisition script.

11. The information processing device according to claim 9, further comprising an acquisition script correction unit that inputs the generated acquisition script and some or all of the file data including the target item into the trained AI model and corrects the generated acquisition script to adapt it to the target file data. 　 The information processing device according to claim 14 , wherein the acquisition script correction unit further inputs a desired information acquisition prompt that is the basis of the generated acquisition script into the trained AI model to correct it.

11. The information processing apparatus according to claim 9, further comprising a detail prompt setting unit that sets a detail prompt that is a detailed prompt for acquiring the desired information in the set desired information acquisition prompt. 　 The information processing apparatus according to claim 10 , further comprising a verification unit that verifies the acquired desired information. 　 The information processing apparatus according to claim 17 , wherein the acquisition script correction unit corrects the acquisition script by reflecting the verification result by the verification unit. 　 a second prompt setting unit that sets a prompt generating prompt for generating an acquisition prompt for acquiring desired information, which is information related to a desired item, from file data including the target item; a second sample setting unit that sets correct sample file data that is input to the trained AI model together with the set prompt generation prompt and serves as a correct example for acquiring the desired information; a correct answer information designation unit that designates specific information on the data as correct answer information on a viewer that displays the set correct answer sample file data; An information processing device having the above.

20. The information processing device according to claim 19, further comprising an acquisition prompt generation unit that inputs the set prompt generation prompt, the set correct answer sample file data, and the specified correct answer information into the trained AI model to generate the acquisition prompt. 　 The information processing device according to claim 20 , further comprising a third desired information acquisition unit that inputs the generated acquisition prompt into the trained AI model to acquire the desired information from the target file data.

22. The information processing apparatus according to claim 20, further comprising an acquisition prompt storage unit that stores the generated acquisition prompt.

22. The information processing apparatus according to claim 21, further comprising an acquisition prompt evaluation unit that compares a result of acquiring the desired information for the correct sample file data using the generated acquisition prompt with the correct information and evaluates the result. 　 The information processing apparatus according to claim 23 , wherein the acquisition prompt generation unit generates an acquisition prompt by reflecting a result of evaluation by the acquisition prompt evaluation unit. 　 The information processing device according to claim 21 , wherein the third desired information acquisition unit acquires the desired information by further inputting the prompt generation prompt, the correct sample file data, and the correct answer information that were input to generate the generated acquisition prompt.

21. The information processing apparatus according to claim 19, further comprising a detailed prompt setting unit that sets a detailed prompt that is a detailed prompt for generating an acquisition prompt from the set prompt generating prompt. 　 The information processing apparatus according to claim 21 , further comprising a verification unit that verifies the acquisition result acquired by the third desired information acquisition unit. 　 The information processing apparatus according to claim 27 , wherein the acquisition prompt generation unit generates an acquisition prompt by reflecting a result of the verification performed by the verification unit. 　 An information processing method executed by an information processing device, a first prompt setting step of setting a desired information acquisition prompt for acquiring desired information, which is information relating to a desired item, from file data including the target item; A first sample setting step of setting correct sample file data that is input to the trained AI model together with the set desired information acquisition prompt and serves as a correct example for acquiring the desired information; a correct answer information designation step of designating specific information on the data as correct answer information on a viewer that displays the set correct answer sample file data; An information processing method comprising: 　 a first prompt setting step of setting a desired information acquisition prompt for acquiring desired information, which is information relating to a desired item, from file data including the target item; A first sample setting step of setting correct sample file data that is input to the trained AI model together with the set desired information acquisition prompt and serves as a correct example for acquiring the desired information; a correct answer information designation step of designating specific information on the data as correct answer information on a viewer that displays the set correct answer sample file data; An information processing program that causes an information processing device to execute the above. 　 An information processing method executed by an information processing device, a second prompt setting step of setting a prompt generating prompt for generating an acquisition prompt for acquiring desired information, which is information related to a desired item, from file data including the target item; A second sample setting step of setting correct sample file data that is input to the trained AI model together with the set prompt generation prompt and serves as a correct example for acquiring the desired information; a correct answer information designation step of designating specific information on the data as correct answer information on a viewer that displays the set correct answer sample file data; An information processing method comprising: 　 a second prompt setting step of setting a prompt generating prompt for generating an acquisition prompt for acquiring desired information, which is information related to a desired item, from file data including the target item; A second sample setting step of setting correct sample file data that is input to the trained AI model together with the set prompt generation prompt and serves as a correct example for acquiring the desired information; a correct answer information designation step of designating specific information on the data as correct answer information on a viewer that displays the set correct answer sample file data; An information processing program that causes an information processing device to execute the above.

Images