Information processing device, information processing method, and computer program

Abstract

This technology promotes the effective use of visualized information such as graphs. [Solution] The information processing device comprises an input receiving unit that receives input of visualization information including information attributes, an acquisition unit that acquires interpretation information for interpreting the visualization information using the attributes of the input visualization information, a generation unit that generates advice regarding the input visualization information using a generated AI model and the acquired interpretation information, and an output unit that outputs the generated advice.

Description

Technical Field

[0001] The present disclosure relates to an information processing apparatus, an information processing method, and a computer program.

Background Art

[0002] An apparatus for graphing the quantity of work processed by production equipment, the time required for processing, etc. is known (for example, Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] Conventionally, visualization of information such as graphing has been performed to facilitate understanding of information, and the visualized information has been used for extracting problems in information sources and improving problems. However, since operations such as extracting problems require a high level of expertise and knowledge, even if it is visualized, it is not easy to understand and utilize the information, and the visualized information may not be effectively utilized.

Means for Solving the Problems

[0005] The present disclosure can be realized in the following forms.

[0006] (1) According to a first embodiment of the present disclosure, an information processing device for utilizing visualization information is provided. This information processing device includes an input receiving unit that receives visualization information including information attributes, an acquisition unit that acquires interpretation information for interpreting the visualization information using the attributes of the input visualization information, a generation unit that generates advice regarding the input visualization information using a generated AI model and the acquired interpretation information, and an output unit that outputs the generated advice. This type of information processing device can generate and output advice regarding the input visualization information. Therefore, the visualization information can be effectively utilized according to the advice. (2) In the above-described information processing device, the interpretation information may include a definition of a procedure for identifying problems related to the visualization information, and a definition of events that cause the problems, according to the attributes of the visualization information. This type of information processing device can improve the accuracy of the analysis of generational AI models regarding problems related to visualized information and the events that cause those problems. (3) The information processing device of the above form may further include a storage unit that stores a plurality of pre-prepared prompts for each attribute of the visualization information as interpretation information, and the acquisition unit may acquire the prompts corresponding to the attributes of the input visualization information from the storage unit as interpretation information. According to this type of information processing device, the acquisition unit can acquire appropriate interpretation information according to the attributes of the input visualization information. (4) The above-described information processing device further comprises a storage unit that stores the interpretation information and the attributes of the visualization information in association, and the acquisition unit may acquire the interpretation information corresponding to the attributes of the input visualization information from the storage unit. According to this type of information processing device, the acquisition unit can acquire appropriate interpretation information according to the attributes of the input visualization information. (5) In the above-described information processing apparatus, the generation unit may generate the advice including problems related to the visualization information and methods for improving those problems. This type of information processing device makes it easy for users to understand problems related to visualized information and methods for improving those problems, even if they do not possess high levels of expertise or knowledge. (6) According to a second embodiment of the present disclosure, an information processing method is provided. This information processing method receives input visualization information including attributes of information, obtains interpretation information for interpreting the visualization information using the attributes of the input visualization information, generates advice regarding the input visualization information using a generating AI model and the obtained interpretation information, and outputs the generated advice. This type of information processing method can generate and output advice regarding the input visualization information. Therefore, the visualization information can be effectively utilized according to the advice. (7) According to a third form of the present disclosure, a computer program is provided. This computer program enables the computer to perform the following functions: to accept input visualization information including attributes of information; to acquire interpretation information for interpreting the visualization information using the attributes of the input visualization information; to generate advice regarding the input visualization information using a generating AI model and the acquired interpretation information; and to output the generated advice. This type of computer program can generate and output advice regarding the input visualization information. Therefore, the visualization information can be effectively utilized according to the advice. This disclosure can also be implemented in various forms other than information processing devices, information processing methods, and computer programs. For example, it can be implemented in the form of a recording medium on which a computer program is stored. [Brief explanation of the drawing]

[0007] [Figure 1] An explanatory diagram showing the configuration of the information processing device according to the first embodiment. [Figure 2] A schematic diagram illustrating the production line. [Figure 3] A flowchart illustrating the process for generating advice using an information processing device. [Figure 4] An explanatory diagram showing the first example of visualized information. [Figure 5] An explanatory diagram showing the first example of advice regarding visualized information. [Figure 6] An explanatory diagram showing a second example of visualized information. [Figure 7] An explanatory diagram showing a second example of advice regarding visualized information. [Figure 8] An explanatory diagram showing the configuration of the information processing device according to the second embodiment. [Modes for carrying out the invention]

[0008] A. First Embodiment: Figure 1 is an explanatory diagram showing the configuration of the information processing device 100 in the first embodiment. Figure 2 is an explanatory diagram schematically showing a factory production line LN. In this embodiment, the information processing device 100 supports the effective utilization of visualization information DS related to the factory production line LN. Visualization information DS is information in which quantities, etc., are visually represented. Visualization information DS can represent statistical charts such as bar graphs, pie charts, line graphs, tables, histograms, scatter plots, box plots, and heat maps.

[0009] As shown in Figure 1, the information processing device 100 is comprised of a computer comprising a processing device 101, a storage device 102, an input / output interface 103, an internal bus 104, an input device 105, a display device 106, and a communication device 107. The processing device 101, the storage device 102, and the input / output interface 103 are connected via the internal bus 104 to enable bidirectional communication. The processing device 101 includes, for example, a CPU and a GPU. The storage device 102 includes, for example, RAM, ROM, a hard disk drive (HDD), and a solid-state drive (SSD). The input / output interface 103 is connected to the input device 105, the display device 106, and the communication device 107. The input device 105 is, for example, a keyboard and a mouse. The display device 106 is, for example, a display and a projector. The input device 105 and the display device 106 may be integrated as a touch panel. The communication device 107 communicates with external devices via wired or wireless communication.

[0010] The processing unit 101 functions as an input receiving unit 110 that receives input of visualization information DS including information attributes by executing an advice generation program PG1 pre-stored in the storage device 102, an acquisition unit 120 that acquires interpretation information DK for interpreting the visualization information DS using the attributes of the input visualization information DS, an advice generation unit 130 that generates advice regarding the input visualization information DS using the generated AI model MD and the acquired interpretation information DK, and an output unit 140 that outputs the generated advice. In this disclosure, the storage device 102 may be referred to as the storage unit, the advice generation program PG1 as the computer program, and the advice generation unit 130 as simply the generation unit.

[0011] The attributes of a visualization data system (DS) refer to the types of statistical charts and graphs included in the DS, as well as the parameters represented in those charts and graphs. Attributes can be represented, for example, as text alongside the statistical charts and graphs within the DS. They may also be represented by tags assigned to the DS.

[0012] The interpretation information DK and the generation AI model MD are pre-stored in the storage device 102. In this embodiment, the interpretation information DK includes the definition of a procedure for identifying problems related to the visualization information DS according to the attributes of the visualization information DS, and the definition of events that cause these problems. For example, ChatGPT® and Gemini® can be used for the generation AI model MD. The generation AI model MD is pre-trained to generate and output appropriate advice regarding the visualization information DS when the visualization information DS and interpretation information DK are input. In this embodiment, the generation AI model MD is a multimodal generation AI model that can accept images and text as input. The visualization information DS is input to the generation AI model MD as an image, and the interpretation information DK is input to the generation AI model MD as text. The advice includes problems related to the visualization information DS and methods for improving these problems.

[0013] As shown in FIG. 2, the production line LN includes a conveying device 20 for conveying the product PD being processed and various production facilities 31 to 34 for producing the product PD. The conveying device 20 is, for example, a belt conveyor, a transporter moving on a track, or the like. The production facilities 31 to 34 are arranged on the conveying device 20 or around the conveying device 20. The production facilities 31 to 34 are, for example, robot arms, processing machines, welding machines, painting machines, inspection machines, and the like. An operator WK is arranged on the production line LN. The operator WK performs operations such as operating the production facilities 31 to 34, assembling the product PD, and inspecting the product PD. Sensors are attached to each of the production facilities 31 to 34. The sensors attached to each of the production facilities 31 to 34 are used, for example, to measure the cycle time (CT) in the production line LN or to detect the state of each of the production facilities 31 to 34. The sensors attached to each of the production facilities 31 to 34 are connected to the production management device 40 by wired communication or wireless communication. In the present embodiment, the production management device 40 generates visualization information DS regarding the production line LN using the measurement information DT received from the sensors of each of the production facilities 31 to 34. The production management device 40 is connected to the information processing device 100 by wired communication or wireless communication. The visualization information DS generated by the production management device 40 is transmitted to the information processing device 100.

[0014] In the visualization information DS regarding the production line LN, for example, cycle time, stop time, operation rate, production rate, etc. are represented in statistical charts. The cycle time is the length of time until the product PD is completed in the production line LN. In the following description, the measured value of the cycle time is referred to as the actual CT, the target value of the cycle time is referred to as the set CT, and the threshold value that serves as a warning criterion when the actual CT exceeds the allowable range is referred to as the warning CT. The stop time is the length of time during which the production line LN stops. The operation rate is the ratio at which the production line LN operates normally. The production rate is the ratio of the actual production quantity to the production quantity when the product PD is produced at a 100% operation rate in accordance with the set CT.

[0015] Figure 3 is a flowchart showing the processing procedure for advice generation by the information processing apparatus 100. In step S110, the input reception unit 110 receives the input of the visualization information DS. In the present embodiment, the input reception unit 110 causes the display device 106 to display an input screen for inputting the visualization information DS. The user can input the visualization information DS received from the production management apparatus 40 to the input reception unit 110 via the input screen by operating the input device 105. Note that the input reception unit 110 may be configured to automatically acquire the visualization information DS without depending on the user's operation.

[0016] In step S120, the acquisition unit 120 acquires the interpretation information DK using the attributes of the visualization information DS. In the present embodiment, a plurality of prompts prepared in advance for each attribute of the visualization information DS are stored in the storage device 102 as the interpretation information DK, and the acquisition unit 120 acquires the prompt corresponding to the attribute of the input visualization information DS from the storage device 102 as the interpretation information DK. Therefore, the acquisition unit 120 can acquire appropriate interpretation information DK according to the attributes of the input visualization information DS. However, tabular data such as Excel (registered trademark) in which the attributes of the visualization information DS and the interpretation information DK are associated may be stored in the storage device 102, and the acquisition unit 120 may acquire the interpretation information DK corresponding to the attribute of the input visualization information DS from the tabular data stored in the storage device 102. Even in this form, the acquisition unit 120 can acquire appropriate interpretation information DK according to the attributes of the input visualization information DS.

[0017] In the present embodiment, the interpretation information DK includes a definition of a procedure for specifying problems regarding the visualization information DS according to the attributes of the visualization information DS, and a definition of events that are the causes of these problems.

[0018] <Interpretation information for interpreting the CT histogram> When the visualization information DS is a CT histogram representing CT by a histogram, problems include the deviation between the mode value of the actual CT and the set CT, and the large variation in the actual CT. Events that can cause the deviation between the mode value of the actual CT and the set CT include process addition and insufficient compliance with standard operations. Events that can cause large variation in the actual CT include work difficulty and insufficient operator proficiency. Therefore, the definition of the procedure for identifying problems includes checking the degree of deviation between the mode value of the actual CT and the set CT, and the degree of variation in the actual CT. The definition of events that can cause problems includes process addition, insufficient compliance with standard operations, work difficulty, and insufficient operator proficiency.

[0019] <Interpretation information for interpreting the CT time series graph> When the visualization information DS is a CT time series graph representing the change of CT in time series, a problem is that the actual CT varies with time zones and deviates from the set CT. Events that can cause the actual CT to vary with time zones and deviate from the set CT include the fatigue of the worker WK, the inappropriateness of the part number switching procedure, and the states of the production facilities 31 to 34. Therefore, the definition of the procedure for identifying problems includes checking the presence or absence of a time zone in which the actual CT is longer than a predetermined ratio with respect to the set CT. The definition of events that can cause problems includes the fatigue of the worker WK, the inappropriateness of the part number switching procedure, and the states of the production facilities 31 to 34.

[0020] <Interpretation information for interpreting the operation rate and production rate graphs> If the visualized information DS is a time-series graph of the utilization rate and the volume rate, problems can be identified when the volume rate is lower than the utilization rate or when the volume rate exceeds 100%. One possible factor that causes the volume rate to be lower than the utilization rate is that the actual turnaround time (CT) is delayed. One possible factor that causes the volume rate to exceed 100% is that the set CT is too long. Therefore, the definition of the procedure for identifying problems includes checking whether there are any time periods when the volume rate is lower than the utilization rate, and checking whether there are any time periods when the volume rate exceeds 100%. The definition of the factors that cause problems includes whether the actual CT is delayed and whether the set CT is too long.

[0021] <Interpretation information for interpreting detailed graphs of factors by type of stoppage> When the visualized information DS is a detailed graph of stop-related factors, representing the length of stop time for each stop-related factor as a bar graph, a problem arises in that it is not possible to identify stop-related factors that negatively impact productivity. One possible reason why stop-related factors cannot be identified is insufficient classification of stop-related factors. Therefore, the definition of the procedure for identifying the problem includes identifying stop-related factors with long total time and evaluating the impact of those factors on productivity. The definition of events that contribute to the problem includes insufficient classification of stop-related factors.

[0022] In step S130, the advice generation unit 130 generates advice using the generated AI model MD and the interpretation information DK. In this embodiment, the advice includes problems related to the visualization information DS and methods for improving these problems. For example, if the visualization information DS is a CT time series graph, problems related to the visualization information DS include, for example, that the actual CT fluctuates depending on the time of day, causing a discrepancy between the actual CT and the set CT. Methods for improving the problems include, for example, that if the actual CT is lagging behind the set CT, it is necessary to check the processes and production equipment 31-34; that attention should be paid to the delay in the actual CT at a specific time of day, and that the fatigue of the workers WK and the work procedures should be re-examined; and that fluctuations in the actual CT after breaks and shift changes should be checked to evaluate the impact of the workers WK's proficiency.

[0023] In step S140, the output unit 140 outputs the advice generated by the advice generation unit 130. In this embodiment, the output unit 140 displays the generated advice on the display device 106. Therefore, the user can refer to the advice displayed on the display device 106.

[0024] Figure 4 is an explanatory diagram showing the first example of Visualized Information DS. Figure 5 is an explanatory diagram showing the first example of advice regarding Visualized Information DS. Figure 4 shows a CT time series graph as an example of Visualized Information DS. In the CT time series graph, the horizontal axis represents time, and the vertical axis represents CT. In the CT time series graph, the solid line represents actual CT, the dashed line represents set CT, and the dotted line represents warning CT.

[0025] Figure 5 shows advice regarding the CT time series graph shown in Figure 4. The advice regarding the CT time series graph includes problems with the CT time series graph and methods for improving these problems. Problems include the fact that the actual CT exceeds the set CT in many time periods, with particularly large peaks observed around 10:00, 21:00, and 4:00 the following day. Suggested improvement methods include identifying process bottlenecks, focusing on time periods when the actual CT significantly exceeds the set CT to confirm what kind of work was being performed during those times, checking the status of production equipment 31-34 and the skill level of workers WK, and revising work procedures or adjusting production equipment 31-34 if necessary, and examining detailed logs related to the time periods when peaks occur to determine whether specific problems were occurring.

[0026] Figure 6 is an explanatory diagram showing a second example of Visualized Information DS. Figure 7 is an explanatory diagram showing a second example of advice regarding Visualized Information DS. Figure 6 shows a graph of cumulative downtime by time as an example of Visualized Information DS. The graph of cumulative downtime by time shows the cumulative time of CT delay and the cumulative downtime for each length of production line LN downtime as bar graphs. The bar graph shows the downtime with different types of hatching for each downtime factor. Below the bar graph, the number of downtimes and the cumulative downtime are shown numerically.

[0027] Figure 7 shows advice regarding the hourly cumulative downtime graph shown in Figure 6. The advice regarding the hourly cumulative downtime graph includes problems with the hourly cumulative downtime graph and methods for improving these problems. The problems identified are frequent delays in CT, frequent short-duration downtimes, and downtime due to quality abnormalities. As for improvement methods, for the problem of frequent short-duration downtimes, adjustments to production equipment 31-34 and improvements to operations are proposed. For the problem of downtime due to quality abnormalities, strengthening quality control and strengthening the training of workers (WK) are proposed. For the problem of frequent delays in CT, it is proposed to re-evaluate the entire process and identify the bottlenecks.

[0028] As described above, the information processing device 100 in this embodiment can generate and output advice regarding visualization information DS using the generated AI model MD and interpretation information DK. The interpretation information DK includes definitions of procedures for identifying problems related to visualization information DS according to the attributes of visualization information DS, and definitions of events that are causes of the problems. Therefore, the generated AI model MD can be made to correctly analyze problems related to visualization information DS and events that are causes of the problems. The output advice includes problems related to visualization information DS and methods for improving these problems. Therefore, by referring to the output advice, the user can easily grasp the problems from visualization information DS and easily work on improving the problems. In contrast, conventionally, even if visualization information DS was generated from measurement information DT related to the production line LN, it was necessary to grasp the problems and work on improving them from visualization information DS alone, and a high level of expertise and knowledge was required to effectively utilize the measurement information DT. Therefore, the measurement information DT that was obtained was not fully utilized. In response to this problem, the information processing device 100 in this embodiment makes it possible for users to identify problems in the production line LN and work to improve them, even without advanced expertise or knowledge. Therefore, it is possible to promote the effective utilization of measurement information DT related to the production line LN.

[0029] B. Second Embodiment: Figure 11 is an explanatory diagram showing the configuration of the information processing device 100b in the second embodiment. In the second embodiment, the information processing device 100b generates the visualization information DS itself, rather than acquiring the visualization information DS generated externally, which is the difference from the first embodiment. The other configurations are the same as in the first embodiment unless otherwise specified.

[0030] In this embodiment, the storage device 102 pre-stores the advice generation program PG1, the generated AI model MD, and the interpretation information DK, as well as the visualization information generation program PG2. The processing device 101 functions as a visualization information generation unit 150 by executing the visualization information generation program PG2. The visualization information generation unit 150 acquires measurement information DT from, for example, a production management device 40, and generates visualization information DS using the measurement information DT. The input receiving unit 110 accepts the visualization information DS generated by the visualization information generation unit 150 as input. The acquisition unit 120 acquires interpretation information DK from the storage device 102 according to the attributes of the visualization information DS input to the input receiving unit 110. The advice generation unit 130 generates advice regarding the visualization information DS using the interpretation information DK and the generated AI model MD. The output unit 140 outputs the visualization information DS and the advice.

[0031] As described above, the information processing device 100b in this embodiment can generate visualization information DS using measurement information DT itself. Therefore, the user does not need to generate visualization information DS from measurement information DT and provide it to the information processing device 100b, thus improving user convenience.

[0032] C. Other embodiments: In each of the above embodiments, the information processing devices 100 and 100b are used to generate and output advice regarding visualization information DS related to the production line LN, in other words, regarding visualization information DS related to the manufacturing industry. In contrast, the information processing devices 100 and 100b may also be used to generate and output advice regarding visualization information DS related to service industries such as retail and medical care, for example, as shown below.

[0033] <Examples of application in the retail sector> As an example of application in the retail sector, let's consider the analysis of the relationship between retail store sales data and weather data. For example, daily sales data for one year is collected as sales data, and weather data (temperature, precipitation, weather) for the same period as the sales data is obtained as weather data. Using the sales data and weather data, daily sales are divided by time of day, and a visualization information DS is generated in which the sales pattern for each weather condition (sunny, cloudy, rainy) is represented as a heatmap. The generating AI model MD, which receives the above visualization information DS and interpretation information DK for interpreting the above visualization information DS as input, analyzes the relationship between sales and temperature, for example, using regression analysis, and analyzes the impact of temperature on sales. If the analysis results show patterns in customer traffic fluctuations due to weather, such as "sales decrease in the afternoon on rainy days" or "sales increase in the evening on sunny days," the generating AI model MD generates advice that includes improvement methods such as adjusting the number of staff to match rainy days or strengthening promotions in the evening on sunny days. This enables efficient store operations. Furthermore, if the analysis reveals a tendency for sales of cold beverages to surge on hotter days, the Generative AI Model MD will generate advice that includes improvement measures such as increasing cold beverage inventory on days when rising temperatures are forecast, and conversely, promoting hot beverages and hot food on days when falling temperatures are forecast. In this way, by appropriately adjusting staffing and promotions based on weather and temperature, it becomes possible to maximize sales and optimize operating costs.

[0034] <Examples of application in the medical field> As an example of application in the medical field, we will describe the optimization of resource allocation based on patient visit patterns and consultation times. For example, visit data is collected for one year, showing the number of patients visiting each department per day, broken down by time of day. Consultation time data is recorded, showing the average consultation time for each department and patient waiting times. Using the visit data and consultation time data, visualization information DS is generated, representing the number of patients visiting each department by time of day using a heatmap, and visualization information DS is generated, graphing the utilization rate of doctors and nurses in comparison with the consultation time data. By analyzing the heatmap, if patterns such as "internal medicine patients are concentrated on Monday mornings" or "orthopedic patients increase in the afternoon" are observed, the generated AI model MD generates advice that includes adjusting staffing and examination room allocation for each department as improvement methods. This makes it possible to reduce patient waiting times and provide more efficient medical care. Furthermore, by analyzing utilization rate graphs, if it is found that there are times when doctors and nurses are excessively busy or, conversely, times when they are idle, the generating AI model MD will generate advice that includes improvement methods such as adjusting shifts to match those times or obtaining support from other departments. This can reduce the burden on healthcare professionals and improve the efficiency of medical care. Also, by analyzing consultation time data, if it is found that consultation times tend to be long in certain departments or time slots, the generating AI model MD will analyze the cause and generate advice that includes improvement methods such as having patients fill out questionnaires in advance or conducting preliminary examinations online as measures to improve the consultation process. In this way, data analysis based on patient visit patterns and consultation times makes it possible to improve the quality of medical services by optimizing medical resources, reducing patient waiting times, and reducing the burden on healthcare professionals.

[0035] This disclosure is not limited to the embodiments described above, and can be implemented in various configurations without departing from its spirit. For example, the technical features in the embodiments corresponding to the technical features in each form described in the summary of the invention can be replaced or combined as appropriate in order to solve some or all of the above-described problems, or to achieve some or all of the above-described effects. Furthermore, if a technical feature is not described as essential in this specification, it can be deleted as appropriate. [Explanation of symbols]

[0036] 20...Conveying equipment, 31-34...Production equipment, 40...Production management equipment, 100, 100b...Information processing equipment, 101...Processor, 102...Memory, 103...Input / output interface, 104...Internal bus, 105...Input device, 106...Display device, 107...Communication device, 110...Input receiving unit, 120...Acquisition unit, 130...Advice generation unit, 140...Output unit, 150...Visualization information generation unit, PG1...Advice generation program, PG2...Visualization information generation program, DS...Visualization information, DK...Interpretation information, MD...Generated AI model, LN...Production line, DT...Measurement information, PD...Product, WK...Worker

Claims

1. An information processing device, An input receiving unit that accepts input of visualization information including information attributes, An acquisition unit that acquires interpretation information for interpreting the visualization information using the attributes of the input visualization information, A generation unit that generates advice regarding the input visualization information using the generated AI model and the acquired interpretation information, An output unit that outputs the generated advice, An information processing device equipped with the following features.

2. An information processing apparatus according to claim 1, The interpretation information includes a definition of a procedure for identifying problems related to the visualization information, and a definition of events that cause the problems, according to the attributes of the visualization information, in an information processing device.

3. An information processing apparatus according to claim 2, The system further includes a storage unit that stores multiple pre-prepared prompts as interpretation information for each attribute of the visualization information, The acquisition unit is an information processing device that acquires the prompt corresponding to the attributes of the input visualization information from the storage unit as the interpretation information.

4. In the information processing apparatus according to claim 2, The system further includes a storage unit that stores the interpretation information and the attributes of the visualization information in association with each other. The acquisition unit is an information processing device that acquires the interpretation information corresponding to the attributes of the input visualization information from the storage unit.

5. In the information processing apparatus according to claim 1, The generation unit is an information processing device that generates advice including problems related to the visualization information and methods for improving those problems.

6. Information processing method, It accepts input of visualization information, including the attributes of the information. Using the attributes of the input visualization information, interpretation information for interpreting the visualization information is obtained. Using the generated AI model and the acquired interpretation information, advice regarding the input visualization information is generated. Output the generated advice. Information processing methods.

7. It is a computer program, A function that accepts input of visualization information including information attributes, A function to obtain interpretation information for interpreting the visualization information using the attributes of the input visualization information, A function that generates advice regarding the input visualization information using the generated AI model and the acquired interpretation information, A function to output the generated advice, A computer program that enables a computer to realize something.

Images