Question answering device, question answering method, question answering program, question answering system, and industrial machine

Abstract

A question answering device 10 for answering questions about an industrial machine is provided with: an information acquisition unit 101 that acquires question information received by a question reception unit; a determination processing unit 102 that determines whether or not the acquired question information includes unique information about the industrial machine; and an answer generation unit 103 that, if it is determined that the question information includes the unique information, generates an answer to the question information on the basis of a closed learning model that has been trained about information that relates to the industrial machine and the unique information and that can be provided only to users of the industrial machine.

Description

Question and answering device, question and answering method, question and answering program, question and answering system, industrial machinery

[0001] TECHNICAL FIELD An embodiment of the present invention relates to a technique for answering questions from users, particularly users who handle industrial machinery.

[0002] User support, in which users' questions are received via email or telephone and answered by an operator, is widely used. In recent years, answering systems utilizing AI (Artificial Intelligence) have rapidly become popular in place of operators for such user support. Answering systems include AI-based chatbots such as Chat GPT (Chat Generative Pre-trained Transformer) provided by Open AI, in which trained AI answers questions in a dialogue format. Because such chatbots can comprehensively use publicly known information, they are able to generate reasonable answers to general user questions.

[0003] A related art technique is known that provides a communication and open-loop / closed-loop control system for at least one filling system, the filling system including a machine, the machine having a software communication robot, in particular a chatbot, configured to recognize voice and / or text input by an operator and / or output or display information related to an operating status of the machine, and an open-loop / closed-loop control device connected to the software communication robot for data communication and configured to control the machine of the filling system in an open-loop and / or closed-loop manner based on the voice and / or text input recognized by the software communication robot (see Patent Document 1 listed below).

[0004] Special table 2021-532467 publication

[0005] The problem to be solved by the embodiments of the present invention is to provide a technology that can generate good answers to various questions, including technical questions from users. Other problems and novel features will become apparent from the description of this specification and the accompanying drawings.

[0006] In one embodiment, when a question from a user of industrial machinery includes unique information about the industrial machinery, the question answering device generates an answer to the question based on a closed learning model that has been trained on information about the industrial machinery and information about the unique information, which can only be provided to the user of the industrial machinery.

[0007] 1 is a block diagram schematically showing the configuration of a question and answering system according to a first embodiment. FIG. 2 is a block diagram showing the hardware configuration of the question and answering device according to the first embodiment. FIG. 3 is a block diagram showing the hardware configuration of a control device provided in the injection molding machine according to the first embodiment. FIG. 4 is a block diagram showing the functional configuration of the question and answering device according to the first embodiment. FIG. 5 is a block diagram showing the functional configuration of a control device provided in the injection molding machine according to the first embodiment. FIG. 6 is a flowchart showing the question and answering process according to the first embodiment. FIG. 7 is a flowchart showing the question and answering process according to the first embodiment. FIG. 8 is a diagram for explaining an answer sentence to question information according to the first embodiment. FIG. 9 is a flowchart showing the question correction process according to the first embodiment. FIG. 10 is a diagram for explaining the process of correction of question information according to the first embodiment. FIG. 11 is a block diagram showing an application example of the question and answering device according to the first embodiment. FIG. 12 is a block diagram showing the functional configuration of a question and answering device according to a second embodiment. FIG. 13 is a block diagram showing the functional configuration of a control device provided in the injection molding machine according to the second embodiment. FIG. 14 is a flowchart showing the question and answering process according to the second embodiment.

[0008] Hereinafter, embodiments of the present invention will be described with reference to the drawings. However, the present invention is not limited to the following embodiments. For clarity of explanation, the following description and drawings have been appropriately simplified. In each drawing, the same elements are given the same reference numerals, and duplicate explanations are omitted as necessary. Furthermore, hatching has been omitted in some areas to avoid cluttering the drawings.

[0009] First Embodiment (Configuration of Question and Answering System) The configuration of a question and answering system according to this embodiment will be described below. Fig. 1 is a block diagram that schematically shows the configuration of a question and answering system according to this embodiment.

[0010] 1, a question and answering system 1 for industrial machinery according to this embodiment includes a question and answering device 10 and a plurality of injection molding machines 20A to 20N. When there is no need to distinguish between the injection molding machines 20A to 20N, they will be referred to as injection molding machine 20 in the following description.

[0011] The industrial machinery question answering device 10 is an information processing device such as a server communicatively connected to a plurality of injection molding machines 20 via a network NW such as the Internet. The question answering device 10 acquires questions from the injection molding machines 20, specifically, acquires questions input by users of the injection molding machines 20 using a hardware I / F 240 (see FIG. 3 ), which will be described later, generates answers to the questions, and provides them to the users. The question answering device 10 is implemented with an automatic conversation program such as a chatbot that utilizes interactive AI, and generates answers using the program. The answers generated by the question answering device 10 are generated based on a closed-type learning model 11 and / or an open-type learning model 12. The closed-type learning model 11 is a machine learning model trained using information stored in a closed database 160 as a dataset. The open-type learning model 12 is a machine learning model trained using information stored in an open database 170 as a dataset.

[0012] Injection molding machine 20 is an industrial machine that has a heating cylinder, a screw, a mold clamping device to which a mold is attached, etc., and melts and kneads resin pellets fed into a hopper connected to the heating cylinder, and then injects the molten resin into a mold to obtain a desired resin molded product. Injection molding machine 20 is equipped with control device 200 that controls the series of device operations described above. Control device 200 according to this embodiment further accepts questions from users, appropriately corrects the questions, etc., and transmits them to question and answering device 10. When an answer to the question is received from question and answering device 10, it notifies the user of the answer.

[0013] (Hardware Configuration) A detailed description will be given of the hardware configuration of the question and answering device 10 and the control device 200. Fig. 2 is a block diagram showing the hardware configuration of the question and answering device according to this embodiment. Fig. 3 is a block diagram showing the hardware configuration of a control device provided in the injection molding machine according to this embodiment.

[0014] As shown in FIG. 2 , the question and answering device 10 includes a CPU (Central Processing Unit) 110, a ROM (Read Only Memory) 120, a RAM (Random Access Memory) 130, a communication I / F (InterFace) 140, a storage device 150, a closed database 160, and an open database 170.

[0015] The CPU 110 executes a BIOS (Basic Input / Output System) read from a ROM 120, which is a non-volatile storage area, an OS (Operating System) read from a storage device 150, general-purpose applications, and various other programs by expanding them onto a RAM 130, which is a volatile storage area. Examples of the programs include the above-mentioned automatic conversation program, as well as a question and answer program that handles part of the question and answer process (a question and answer method for industrial machinery) described below. The communication I / F 140 has a gateway (not shown) and is used for communication with the injection molding machine 20 via a network NW.

[0016] The storage device 150 is, for example, a hard disk drive (HDD) or a solid state drive (SSD), and stores various information used in the question and answer process. The stored information includes a list showing the correspondence between a personal identification code and a closed learning model 11. The personal identification code will be described later. The closed database 160 and the open database 170 are storage areas containing different information (which may contain some overlapping information) used to answer questions. These databases are constructed using storage devices such as an HDD or an SSD.

[0017] The closed database 160 is a storage area containing information about industrial machinery, in this embodiment, the injection molding machine 20, and information about unique information, and is used as a data set for the closed learning model 11. The information about the injection molding machine stored in the closed database 160 includes, for example, the contents of the instruction manual for the injection molding machine 20 and information that can only be provided to the user of the injection molding machine 20 (including the company to which the user belongs), in other words, highly confidential information. Therefore, the closed learning model 11 trained based on the closed database 160 is linked to a PIN code (identification information) that uniquely identifies the user and / or the injection molding machine 20 to which the question was asked.

[0018] The personal identification code is information that uniquely identifies each of the injection molding machines 20A to 20N, in other words, is used to identify a user (an individual or an organization such as a company to which the individual belongs) who owns the injection molding machine 20. Therefore, when the same user owns multiple different types of injection molding machines 20, the personal identification codes for the multiple injection molding machines 20 may be the same, or the personal identification code may be linked to the type of injection molding machine 20 and stored in the storage device 150. The personal identification code also serves as information that indicates the closed-type learning model 11 specific to the user.

[0019] In this embodiment, a unique closed learning model 11 is prepared for each user, and a unique closed database 160 is associated with each closed learning model 11. As a result, a response based on the closed learning model 11 associated with the user is provided to the user indicated by the PIN code. Therefore, a response is provided within the scope that can be disclosed to the user, and a response based on another closed learning model 11, that is, a response within the scope that cannot be disclosed to the user, is not provided.

[0020] Particularly confidential information includes a response method for each error code, information on the molding conditions (operating conditions) of the injection molding machine 20 that will produce the desired molded product, specification information for the injection molding machine 20, and information on actual measurements made during molding in the injection molding machine 20. It is preferable that a plurality of response methods for each error code be prepared according to the molding condition information of the injection molding machine 20.

[0021] Specifically, the closed-loop learning model 11 learns the relationship between questions (molding condition information, specification information, error codes, information indicating an abnormality, information about molded products, and / or actual measurement information) and answers (molding condition information and / or answers in the instruction manual). Here, "information indicating an abnormality" refers primarily to questions about abnormalities not indicated by error codes, such as abnormal noises, that users ask because they suspect an abnormality. Here, "information about molded products" refers primarily to questions about users' inability to obtain the desired molded products. It is also possible for users to prioritize production, energy conservation, or machine life. Therefore, if a question contains a user's priorities, it is preferable that answer patterns (e.g., modified values ​​of molding conditions) be linked to the above-mentioned relationships so that molding condition information consistent with those priorities can be included in the answer document.

[0022] Examples of the specification information of the injection molding machine include the inner diameter of the heating cylinder, the screw type (L / D), the performance of the servo motor (rated output, rated torque, rated rotation speed, etc.), the performance of the mold clamping mechanism, the usage history of the injection molding machine, etc. Examples of the molding condition information of the injection molding machine 20 include the heating cylinder control temperature, the injection speed during injection, the screw rotation speed and back pressure during metering, the clamping force and mold opening / closing speed of the mold clamping device, the material type, the material supply amount, etc. The actual measurement value information includes the detection results of various sensors such as temperature sensors and timers possessed by the injection molding machine 20, and examples of the actual measurement value information include the measured temperature of the heating cylinder, operating information during molding, injection information such as the injection speed, metering information such as the metering completion position and the screw rotation speed during metering, and clamping information such as the mold clamping force and the mold opening / closing speed, etc. For example, when the specification information, molding condition information, and actual measurement value information (operation information) are acquired as a question together with an error code, the question and answering device 10 can accurately answer how to respond to the error code according to the specification information, molding condition information, and actual measurement value information by referring to the closed-loop learning model 11. The response method preferably includes information on how to change the molding condition values, etc. By providing such a response method, it becomes possible to provide an accurate answer according to the user's current situation.

[0023] The unique information is information for determining whether to use the closed-type learning model 11 to answer the question, and in this embodiment, examples of the unique information include an error code, specification information of the injection molding machine 20, molding condition information, and actual measurement value information. Since there is a plurality of molding condition information, the unique information may be limited to specific molding conditions. Note that the unique information is not limited to error codes, specification information, molding condition information, and actual measurement value information, and may be any information suitable for generating an answer using the closed-type learning model 11. For example, the unique information may be the type (model) of the injection molding machine 20 or wording related to the instruction manual, such as the word "instruction manual" or symbols described in the instruction manual.

[0024] When a user's evaluation is obtained after an answer, the closed learning model 11 is trained to improve the accuracy of the answer through machine learning such as reinforcement learning and supervised learning based on the evaluation, deep learning, additional learning, etc. Preferably, the closed database 160 is also updated each time through this training. It is preferable that the learning includes feedback not only of the evaluation but also of the process leading to the question and answer. Therefore, the closed learning model 11 and the closed database 160 are individually customized for each user. Note that by associating the closed learning model 11 and the closed database 160 individually with each user, the results of learning that differ for each user are not shared between users. This prevents information in the closed database 160 from leaking.

[0025] On the other hand, in this embodiment, the open database 170 is a storage area containing low-confidentiality information related to the injection molding machine, in other words, publicly known information that can be provided not only to the user of the injection molding machine 20 but also to various other users. The open database 170 is used as a data set for the open learning model 12. Therefore, the open learning model 12 has learned the relationship between publicly known information and is used when a publicly known question is asked. The open database 170 can be, for example, a database used in chat GPT provided by Open AI, Inc.

[0026] It is preferable that information be added to the closed database 160 and the open database 170 as needed. While the information added to the open database 170 is publicly known information, information about industrial machinery (here, the injection molding machine 20) is added to the closed database 160. For example, the closed database 160 is automatically updated by the question and answering device 10 periodically accessing external information and collecting equipment information from industrial machinery manufacturers via the network NW. This is preferably done not only by equipment manufacturers, but also when new knowledge about parts manufacturers or industrial machinery is discovered. For example, when new knowledge is discovered, the maintenance and operation company of the question and answering device 10 can add the new knowledge to the closed database 160 as needed. Note that the closed database 160 may also be automatically updated by accessing a dedicated server owned by the maintenance and operation company that collects new knowledge. When information is added to the closed database 160 and the open database 170 in this way, the closed learning model 11 and the open learning model 12 are re-learned.

[0027] As shown in FIG. 3 , the control device 200 includes a CPU 210 , a ROM 220 , a RAM 230 , a hardware I / F 240 , a user interface consisting of a display 241 , an input device 242 , a microphone 243 and a speaker 244 , a communication I / F 250 and a storage device 260 .

[0028] The CPU 210 executes the BIOS read from the ROM 220, which is a non-volatile storage area, the OS read from the storage device 260, general-purpose applications, and various other programs by expanding them on the RAM 230, which is a volatile storage area. The programs include a program that handles part of the question and answer process described below.

[0029] The hardware I / F 240 is used for communication with user interfaces such as a display 241, an input device 242, a microphone 243, and a speaker 244. Specifically, the hardware I / F 240 is used to receive questions from the user and notify the user of answers. The display 241 displays questions and answers to the questions as character strings. The input device 242 includes a keyboard, a mouse, etc., and receives input operations from the user. The microphone 243 receives questions and the like by voice from the user. The speaker 244 leaves answers to questions and the like by voice.

[0030] These user interfaces of the injection molding machine 20 are the question receiving unit of the present invention that receives questions from the user via voice or key input. The user interface also serves as the question answering unit of the present invention that conveys answers generated by the answer generation unit 103 of the question answering apparatus 10 (described later) to the user as voice from the display 241 or speaker 244. The user interface of the injection molding machine 20 also contributes to solving the problem of the invention. In this embodiment, the user interfaces, such as the display 241, input device 242, microphone 243, and speaker 244, that constitute the question receiving unit and question answering unit of the question answering apparatus of the present invention are provided in each injection molding machine 20. However, the user interface may be a portable device that is shared by multiple injection molding machines 20. A user interface dedicated to the injection molding machine 20, which is attached to the injection molding machine 20 and has a portable question receiving unit and question answering unit, can be considered to be part of the injection molding machine 20. Examples of such a user interface that is a portable device include a tablet terminal such as a smartphone, a notebook PC (Personal Computer), etc. Furthermore, a desktop PC or the like installed in a user's factory may be provided with the functions of the question receiving unit and the question answering unit.

[0031] The communication I / F 250 has a gateway (not shown) and is used for communication with the question and answering apparatus 10 via the network NW. The storage device 260 is, for example, a hard disk drive (HDD) or a solid state drive (SSD), and stores various information used in the question and answering process. The stored information includes a personal identification code, additional saved information, and answer documents to questions sent from the question and answering apparatus 10. The additional saved information is information added in the question correction determination process (described later), and details of this will be described later.

[0032] (Functional Configuration) A detailed description will be given of the functional configuration of the question and answering device 10 and the control device 200. Fig. 4 is a block diagram showing the functional configuration of the question and answering device according to this embodiment. Fig. 5 is a block diagram showing the functional configuration of the control device provided in the injection molding machine according to this embodiment.

[0033] 4, the question and answering apparatus 10 includes, as its functions, an information acquisition unit 101, a determination processing unit 102, an answer generation unit 103, a verification processing unit 104, an answer transmission unit 105, and a learning update unit 106. These functions are realized by cooperation of the hardware of the question and answering apparatus 10 described above.

[0034] The information acquisition unit 101 acquires various information transmitted from the injection molding machine 20. The acquired information includes question information indicating a user's question, evaluation information indicating a user's evaluation of an answer to the question, etc. The judgment processing unit 102 makes various judgments in the answer generation process. The judgments include, for example, whether or not unique information is present in the question information, and whether or not correction of the question information is necessary.

[0035] The answer generation unit 103 is a so-called AI chatbot that generates answers to question information based on the closed-type learning model 11 and / or the open-type learning model 12. Note that the answer generation unit 103 is not limited to an AI chatbot, and any automatic response (automatic conversation) program that can generate answers to questions using each learning model will suffice. The verification processing unit 104 verifies the generated answer to determine whether it should be corrected.

[0036] The answer sending unit 105 sends the generated, and in some cases modified, answer to a user interface installed in the injection molding machine 20 that sent the question information. When the learning update unit 106 acquires information for machine learning, it performs machine learning based on the information and updates the closed learning model 11 and the closed database 160. The information for machine learning includes evaluation information and condition change information. The condition change information includes actual measurement values ​​obtained when settings are changed from those included in the answer in the injection molding machine 20 that received the answer to the question, the changed settings, and the like.

[0037] 5, the control device 200 includes, as its functions, an information acquisition unit 201, a conversion processing unit 202, an information transmission unit 203, a determination processing unit 204, a notification unit 205, a monitoring processing unit 206, and a question correction unit 207. These functions are realized by cooperation between the hardware of the control device 200 described above.

[0038] The information acquiring unit 201 acquires input from the user and various information transmitted from the question answering apparatus 10. Input from the user includes questions, evaluations, etc. The acquired information includes a personal identification code and answers generated by the question answering apparatus 10. The conversion processing unit 202 converts the voice-input question into text.

[0039] The information sending unit 203 sends question information and evaluation information to the question and answering apparatus 10. The judgment processing unit 204 makes various judgments in the answer generation process. These judgments include, for example, whether evaluation information has been input or whether monitoring processing is ON or not. The monitoring processing is a process for monitoring the behavior of the injection molding machine 20 after an answer to a question has been obtained, and collects information such as various settings and detection results obtained from various sensors. The notification unit 205 notifies the user of the answer to the question by voice or text. The monitoring processing unit 206 executes the monitoring processing. The question correction unit 207 corrects the question information in the question correction determination processing described below.

[0040] (Question Answering Process) The question answering process performed by the question answering system according to this embodiment will be described in detail. FIGS. 6 to 8 are flowcharts showing the question answering process according to this embodiment. This process is triggered by a question button being displayed on various injection molding setting screens, etc., on the display 241 (question receiving unit) of the injection molding machine 20, and the button being selected. Note that this is not a limitation, and for example, the control device 200 may constantly collect sound using the microphone 243 (question receiving unit), and this process may be executed when a specific voice command is acquired.

[0041] First, as shown in FIG. 6 , the information acquisition unit 201 acquires a voice question from a user as voice information (S101). Specifically, the information acquisition unit 201 collects voice uttered by a person using the microphone 243, and when a certain time has passed since the voice collection started and the voice stopped, the information acquisition unit 201 acquires the collected voice as voice information indicating a question. Note that if the voice does not stop within a preset collection time after the voice collection started or if the voice cannot be picked up, the notification unit 205 may notify the user of an error by voice using the speaker 244 or by displaying a text message on the display 241. In this case, the question and answer process ends.

[0042] After acquiring the voice information, the conversion processing unit 202 converts the acquired voice information into text to generate question information (S102). The process of converting voice into text may use an existing voice-to-text conversion program, such as AI-based voice recognition. At this time, it may be advisable to modify the question information into a sentence using existing natural language processing. The question may be acquired as text input by the user using the input device 242. In this case, the input text can be used as the question information as is. After generating the question information, the information acquisition unit 201 acquires the PIN code stored in the storage device 260 (S103).

[0043] After acquiring the personal identification code, a determination is made as to whether the generated question information needs to be corrected, and if so, a question correction determination process is executed in which correction is made (S104). The question correction determination process will be described in detail later. After the question correction determination process, information sending unit 203 sends the question information together with the personal identification code to question and answering device 10 (S105), and control device 200 enters a standby state in the question and answering process.

[0044] The question answering apparatus 10 is always in a standby state where it can acquire various types of information. When question information and a secret code are transmitted in this state, the information acquiring unit 101 of the question answering apparatus 10 acquires (receives) them (S106). After acquisition, the determination processing unit 102 determines whether the unique information is included in the question information (S107).

[0045] If it is determined that the unique information is included in the question information (S107, YES), the information acquisition unit 101 acquires the secret code received together with the question information (S108). After acquisition, the answer generation unit 103 reads the list stored in the storage device 150 and selects the closed-type learning model 11 corresponding to the acquired secret code (S109). After selection, the answer generation unit 103 generates an answer document corresponding to the question information based on the selected closed-type learning model 11 (S110).

[0046] For example, if the question information includes an error code, the answer document generated using the closed-loop learning model 11 preferably includes a more detailed explanation of the countermeasures for the error code, particularly the countermeasures described in the instruction manual. Furthermore, if the question information includes molding condition information, actual measurement value information, and specification information, an answer derived from the conditions and actual measurement values ​​is generated. The answer includes appropriate changes to the molding conditions, etc.

[0047] In particular, when an error code is included in the question information, it is preferable to present multiple changes to the molding conditions that can avoid the abnormal state indicated by the error code. Furthermore, when the above-mentioned priority items are included in the question information, molding conditions that are in line with the priority items are included in the answer document. Note that even when such priority items are not included in the question document, multiple answer patterns may be presented. Furthermore, it is also preferable to present candidate molding conditions in order of the likelihood that they will not cause any problems in the operation of the injection molding machine.

[0048] On the other hand, if it is determined that the question information does not contain unique information (S107, NO), the answer generation unit 103 selects the open-type learning model 12 (S111), and in step S110, an answer document corresponding to the question information is generated based on the open-type learning model 12. If unique information is not included, it can be determined that a general question with low confidentiality has been asked, and therefore a sufficient answer can be obtained from the open-type learning model 12. Note that even if there are unclear parts of the question that are not sufficient to obtain an answer that is expected to fully satisfy the user, the question information is corrected by a question correction determination process, the details of which will be described later, so that a highly accurate answer that satisfies the user can be obtained.

[0049] As shown in FIG. 7 , after generating the question document, the determination processing unit 102 determines whether the generated answer document is verifiable (S112). This determination is made based on whether the answer document contains enough molding conditions to enable the verification process. The verification process is a so-called simulation process using the molding conditions. If the answer document contains the molding conditions necessary for the simulation process, the generated answer document is determined to be verifiable. In addition to determining whether the answer document contains enough molding conditions to enable the verification process, it may also be determined whether the question information corresponding to the answer document contains specification information that enables the verification process. In this case, if both determination results are positive, it is determined that the verification process is possible, and a simulation process using the molding conditions and specification information is performed.

[0050] If the generated response document is verifiable (S112, YES), the verification processing unit 104 executes the verification process (S113) and obtains the verification result. After obtaining the verification result, the judgment processing unit 102 determines whether the response document needs to be corrected (S114). This determination is made based on whether the verification result contains any defects, such as molded products generally not being of good quality or errors (e.g., inconsistencies in modified molding conditions, physical damage to the equipment, etc.). Information on the quality of the molded products may be pre-stored in the storage device 150 as data to be used in the verification process, or may be obtained by the closed-type learning model 11 or the open-type learning model 12. The response document may also include information on the quality of the molded products. The obtained verification result is preferably used as training data for training the closed-type learning model 11 at a predetermined timing. At this time, the weights of the neural network of the closed-type learning model 11 may be adjusted using a technique such as backpropagation.

[0051] If it is determined that the answer document needs to be corrected (S114, YES), the answer generation unit 103 corrects the answer document (S115), and the verification process of step S113 is performed again. It is preferable that the correction content, i.e., the changes to the molding conditions, be determined based on the verification results. The correspondence between the verification results and the changes may be stored in advance in the storage device 150. Furthermore, a priority order for the types of molding conditions to be changed may be determined in advance. This priority order may be changed depending on the type of resin or the type of injection molding machine 20. Furthermore, an answer may be regenerated using the closed-type learning model 11 or the open-type learning model 12 again based on the question information including the verification results.

[0052] On the other hand, if it is determined that the answer document does not require correction (S114, NO), or if it is determined that the generated answer document is not verifiable (S112, NO), the answer sending unit 105 sends the answer document to the injection molding machine 20 that sent the question information (S116). After sending, the question and answering device 10 enters a standby state. Note that, if the sent answer document was generated using the closed-type learning model 11, it is preferable that it be stored in the storage device 150 together with the secret code. This is because the answer document will be used for later learning of the closed-type learning model 11.

[0053] After transmission, the information acquisition unit 201 of the injection molding machine 20 acquires (receives) the response document (S117). After acquisition, the notification unit 205 notifies the user of the response document (S118). This notification may be performed by displaying the response document using the display 241 or by audibly transmitting the response document using the speaker 244. Therefore, the user interfaces such as the display 241 and the speaker 244 contribute to solving the problems according to the present invention.

[0054] After the notification, the determination processing unit 204 determines whether the user has a follow-up question (S119). This determination is made, for example, by displaying a window on the display 241 as a GUI (Graphical User Interface) that displays in text whether there is a follow-up question and has selectable "YES" and "NO" buttons, and by checking whether each button is selected. Alternatively, the speaker 244 may be used to audibly ask whether there is a follow-up question, and the determination may be made by checking whether the user's voice response (e.g., a voice similar to "yes" or "no," or a follow-up question) is acquired using the microphone 243 within a predetermined period.

[0055] If it is determined that the user has a follow-up question (S119, YES), the voice question input process of step S101 shown in FIG. 6 is performed. On the other hand, if it is determined that the user has not a follow-up question (S119, NO), as shown in FIG. 8, the determination processing unit 204 determines whether the user has provided a rating (S120). This determination is made based on whether a rating has been acquired by the information acquisition unit 201 within a predetermined time period after the determination of step S119. For example, a window having a "Rate" button is displayed on the display 241, and depending on the selection of the button, a window having a "Resolved" (good rating) button and a "Not Resolved" (bad rating) button is displayed on the display 241 as the rating. Depending on the selection of the "Resolved" or "Not Resolved" button, a rating is acquired, and it is determined that a rating has been provided.

[0056] Alternatively, after selecting the "Evaluate" button, a written or voice-input evaluation may be acquired. Instead of "Resolved" or "Not Resolved," evaluations may be made on a multi-level scale, such as a five-level scale. The predetermined time for this evaluation is preferably set long enough for the user to operate the injection molding machine 20 in accordance with the response document and recognize the evaluation results. Such a time may range from several minutes to several hours, for example. The time available for evaluation varies depending on the content of the question. Therefore, it is preferable to store the acquired response document in the storage device 260 so that evaluation can be performed at any time. In this case, steps S120 and S121, as well as steps S126 to S129 (described later) that are executed as a result of these steps, are preferably separated from the question-answering process and can be performed as evaluation processing at any time.

[0057] If it is determined that the user has made an evaluation (S120, YES), the information sending unit 203 sends evaluation information indicating the evaluation result together with the personal identification code to the question and answering device 10 (S121), and the process proceeds to step S122. On the other hand, if it is determined that the user has not made an evaluation (S120, NO), the process proceeds to step S122 without performing the processing of step S121.

[0058] In step S122, the determination processing unit 204 determines whether the monitoring process is set to ON (S122). The monitoring process is a process in which the monitoring processing unit 206 monitors the operation of the injection molding machine 20, and collects molding conditions and actual measured values ​​during operation (during injection molding). The monitoring process can be switched ON / OFF using a selectable "ON" or "OFF" button included in a screen displayed on the display 241 while the injection molding machine 20 is operating.

[0059] If it is determined that the monitoring process is ON (S122, YES), the determination processing unit 204 determines whether the molding conditions have been changed by the user (S123). This determination is made based on whether the molding conditions have been changed within a predetermined change determination time after the response document is acquired. If it is determined that the molding conditions have been changed by the user (S123, YES), the determination processing unit 204 determines whether the changed molding conditions are OK, i.e., whether there are no problems with the molding conditions (S124). This determination is made based on whether molding has been performed a predetermined number of times consecutively under the changed molding conditions. Alternatively, the user may input that the molding conditions are OK by selecting a button, etc., and the determination processing unit 204 may determine whether such input has been made.

[0060] If the changed molding conditions are determined to be NG, i.e., there is a problem with the molding conditions (S124, NO), the user is likely to change the molding conditions again. Therefore, the judgment processing unit 204 performs the judgment process of step S124 again, and this judgment is repeated until it is determined that the molding conditions are OK. Note that, since the monitoring process continues, the judgment process of step S124 may be triggered again by another change in the molding conditions. Also, at this stage, the user may ask another question. If a second question is received, this process may be interrupted or terminated, and question information may be generated and transmitted. In this case, the question and answering device 10 preferably manages the received question as related to the previous question.

[0061] If it is determined that the changed molding conditions are OK (S124, YES), the information sending unit 203 sends condition change information including the changed molding conditions and the secret code to the question and answering device 10 (S125), and the question and answering process on the injection molding machine 20 side is terminated. The condition change information preferably includes all information that can be collected in the monitoring process, along with the changed conditions. On the other hand, if it is determined that the monitoring process is set to OFF (S122, NO), or if it is determined that the molding conditions have not been changed by the user (S123, NO), the question and answering process on the injection molding machine 20 side is terminated. Note that if it is determined that the molding conditions have not been changed by the user, this means that the elapsed time since the answer document was obtained exceeds the change determination time described above.

[0062] When the evaluation information and the secret code are transmitted to the question answering device 10, the information acquiring unit 101 of the question answering device 10 acquires (receives) the evaluation information and also acquires the secret code (S126, S127). After acquisition, the determination processing unit 102 determines whether the answer document corresponding to the evaluation information was generated using the closed-type learning model 11 (S128). The secret code is stored in association with the answer document in the storage device 150. Therefore, it is possible to determine whether the answer document corresponding to the evaluation information was generated using the closed-type learning model 11 from the secret code acquired together with the evaluation information.

[0063] If it is determined that the answer document corresponding to the evaluation information was created using the closed learning model 11 (S128, YES), the learning update unit 106 updates the closed learning model 11 based on the evaluation information (S129). Examples of the update include additional learning of the closed learning model 11 based on the evaluation information and the answer document. At this time, it is preferable that the closed database 160 is also updated. After the update, the question answering device 10 enters a standby state (S130). On the other hand, if it is determined that the answer document corresponding to the evaluation information was not created using the closed database 160, for example, because no answer document is stored in the storage device 150 (S128, NO), the device transitions to the standby state of step S130.

[0064] When the condition change information and the secret code are transmitted to the question answering device 10, the information acquisition unit 101 of the question answering device 10 acquires (receives) the condition change information and also acquires the secret code (S131, S132). After acquisition, the determination processing unit 102 determines whether the answer document corresponding to the condition change information was generated using the closed-type learning model 11 (S133). The secret code is stored in association with the answer document in the storage device 150. Therefore, it is possible to determine whether the answer document corresponding to the condition change information was generated using the closed-type learning model 11 from the secret code acquired together with the condition change information.

[0065] If it is determined that the answer document corresponding to the condition change information was created using the closed learning model 11 (S133, YES), the learning update unit 106 updates the closed learning model 11 based on the condition change information (S134). Examples of the update include additional learning of the closed learning model 11 based on the condition change information and the answer document. At this time, it is preferable that the closed database 160 is also updated. After the update, the question answering process on the question and answering device 10 side is terminated. On the other hand, if it is determined that the answer document corresponding to the condition change information was not created using the closed learning model 11, for example, because no answer document is stored in the storage device 150 (S133, NO), the question and answering process on the question and answering device 10 side is terminated.

[0066] An example of a response document to a question will be described. FIG. 9 is a diagram illustrating a response sentence to a question according to this embodiment. The symbol Qu in FIG. 9 indicates question information, and the symbols An1 to An3 indicate different response documents to the question Qu. The question Qu includes unique information such as an error code A1, specification information, molding condition information, and actual measurement value information. Therefore, for such question Qu, a closed-loop learning model 11 is used to generate the response document. In this example, the response documents An1 to An3 are generated because the information concealed in the question Qu is different.

[0067] Answer document An1 provides an appropriate solution to the servo motor overload error indicated by error code A1, including information about molding conditions. Answer document An2 indicates that the user is forcing the injection molding machine 20, which is currently in use and cannot resolve the overload error, based on the model of injection molding machine 20 indicated in the question. Answer document An2 also indicates that the user's personal identification code identifies the injection molding machine 20 owned by the user. Therefore, the answer includes information that the problem will not occur if the user uses a different type of injection molding machine 20 owned by the user. Answer document An3 also indicates that the injection molding machine 20 requires inspection. These answers are accurate and appropriate because each unique information, particularly specification information, molding condition information, and actual measurement value information, is included in the question information Qu. However, even if the question information Qu does not contain specification information, molding condition information, or actual measurement value information, the question answering device 10 can at least answer by referring to the contents of the instruction manual, so that an accurate answer can be obtained and the answer can be obtained much faster than the time it takes for the user to refer to the instruction manual himself.

[0068] The question correction determination process will be described in detail. Questions from users may be subjective, and the relationship between subjects and predicates may be unclear. In such cases, even if the user's question is directly input into the closed learning model 11 of the question answering device 10, an appropriate answer may not be obtained. In such cases, the question is organized using an open learning model with high processing capabilities of the language model, such as Chat GPT, or is converted by a dedicated conversion processing unit 202 before being input into the closed learning model 11. These processes are also called prompt engineering. Figure 10 is a flowchart showing the question correction determination process according to this embodiment.

[0069] 10 , first, the determination processing unit 204 of the injection molding machine 20 determines whether additional storage information is necessary (S201). The additional storage information here includes detailed explanations of various error codes, specification information of the injection molding machine 20 that is the source of the question, such as the type, model, and product name of the injection molding machine 20, explanations of technical terms related to the injection molding machine 20, and explanations of the instruction manual for the error code. Therefore, if the question information includes an error code, technical terms related to the injection molding machine 20, or technical terms or symbols found in the instruction manual, or if basic information about the injection molding machine 20 is not included, it is determined that additional storage information is necessary.

[0070] If it is determined that additional saved information is necessary (S201, YES), the question correction unit 207 acquires the necessary information from the storage device 260 (S202) and corrects the question information (S203). For example, the third paragraph from the top of the question information Qu shown in FIG. 9 , which reads, "The model of this injection molding machine is ____...," is the specification information corrected as additional saved information. In other words, this portion is a sentence automatically added in response to the user's oral question, and it is desirable to replace it with linguistic information like this sentence and input it into the conversational closed-loop learning model. After correction, the determination processing unit 204 determines whether the question information requires molding conditions and / or actual measured values ​​(S204). This determination is made by determining whether specific keywords are included in the question information. Examples of keywords include error codes. While it is preferable to include all error codes, only some of the error codes may be included. Examples of some error codes include those that have a very wide range of answer options when molding conditions or actual measured values ​​are not available, and those that make it difficult to obtain accurate answers.

[0071] If it is determined that the question information requires molding conditions and / or actual measurement values ​​(S204, YES), the determination processing unit 204 determines whether the molding conditions and / or actual measurement values ​​are set to be transmittable to the question answering device 10 (S205). The molding conditions and / or actual measurement values ​​are often confidential information of the company when transmitted to an external device. Therefore, even when transmitting information to the question answering device 10, where confidential information is managed, it is preferable that the transmission setting be switchable between ON and OFF. For example, to change the settings of the injection molding machine 20, a selectable switch button may be displayed on the setting screen displayed on the display 241. The switch button may be individually switchable for each molding condition and each actual measurement value. The transmission setting can be switched ON or OFF depending on the selection of these buttons. Therefore, the determination in step S205 is made by determining whether the transmission setting is set to ON or OFF.

[0072] If it is determined that the molding conditions and / or actual measured values ​​are set to be transmittable to the question answering device 10 (S205, YES), the question correction unit 207 acquires the current molding conditions and actual measured values ​​that are set to be transmittable (S206). After acquisition, the question correction unit 207 adds the acquired molding conditions and / or actual measured values ​​to the question information (S207), and the question correction determination process ends. For example, the fourth paragraph from the top of the question information Qu shown in FIG. 9 , which reads, "The operating status of the injection molding machine before the question was asked...," is the added and corrected molding conditions and actual measured values. It is desirable that this portion be replaced with linguistic information as shown in FIG. 9 and input into a conversational closed-loop learning model.

[0073] On the other hand, if it is determined that the molding conditions and / or actual measurement values ​​are not set to be sendable to the question and answering device 10 (S205, NO), or if it is determined that the molding conditions and / or actual measurement values ​​are not required for the question information (S204, NO), the question correction determination process ends. Also, if it is determined that additional saved information is not required (S201, NO), the process proceeds to the determination process of step S204.

[0074] FIG. 11 is a diagram illustrating the process of correcting question information according to this embodiment. The symbols Qu1 to Qu3 in FIG. 11 each represent question information, with the correction status increasing as the number at the end of the number increases. Question information Qu1 is uncorrected question information when it was first generated, and although the error code is known, the specific error cannot be identified. Therefore, if the error code is not set in the unique information, an answer will be obtained using the open database 170, but there is a high possibility that no answer will be obtained, or that an incorrect answer will be obtained.

[0075] On the other hand, the question information Qu2 has had the error code content added as stored additional information through the question correction determination process. In this state, even if an answer is obtained using the open database 170, a reasonable answer can be obtained, but it is unlikely that an answer that satisfies the user will be obtained. In contrast, the question information Qu3 has had the content of the instruction manual corresponding to the error code added as stored additional information through the question correction determination process. In this state, it can be said that an answer that satisfies the user to a certain extent can be obtained, even if an answer is obtained using the open learning model 12. Furthermore, when the molding conditions and actual measured values ​​are added to the question information through the question correction determination process, the question information Qu shown in FIG. 9 is obtained. In this state, a highly accurate answer that satisfies the user can be obtained.

[0076] If the error code is included in the question information and set in the unique information, the closed learning model 11 has already learned the error code information and the instruction manual information (stored in the closed database 160), and therefore, even for the question information Qu1, it is possible to generate an answer equivalent to that for the question information Qu3 without correction. Therefore, in the question correction determination process, the feature of being able to correct the molding conditions and / or actual measured values ​​is particularly useful.

[0077] According to the present embodiment described above, it is possible to provide a good answer to a user's technical questions. Furthermore, compared to when the user refers to the instruction manual by themselves, the answer can be obtained much more quickly. Furthermore, for questions that include molding conditions, actual measured values, etc., answers are generated using the closed-type learning model 11, which has already learned specialized knowledge, so more specific and accurate answers can be obtained. Furthermore, because the closed-type learning model 11 is uniquely prepared for each user, it is possible to prevent confidential information from leaking to the outside, such as the leaking of know-how to other users. Similarly, because feedback on questions, answers, and evaluations of answers is provided only to the closed-type learning model 11, it is possible to constantly improve the answer accuracy of the learning model while preventing confidential information from leaking to the outside.

[0078] In addition, the generated answers are verified as much as possible, preventing the sending of incorrect answers to the user. Questions are appropriately corrected and evaluated, allowing the content of the questions to be enriched so that highly accurate answers can be obtained.

[0079] Furthermore, users can ask questions orally and receive answers in voice or text, making it extremely convenient as they can easily get answers to their questions as if they had a skilled technician nearby.

[0080] In this embodiment, the question and answering system 1 has been described, which answers questions from users of the injection molding machine 20. However, the users are not limited to users of the injection molding machine 20. For example, the system can be applied to users of all kinds of industrial machinery, such as kneading machines, extruders, presses, and semiconductor manufacturing equipment. Industrial machinery such as the injection molding machine 20 also includes its peripheral devices. Furthermore, plant equipment including at least one or more devices, such as a kneading machine, an extruder, and a press, is also included in the industrial machinery of the present invention. In this case, naturally, information about various industrial machines is stored in the closed database 160 of the question and answering device 10 of the question and answering system 1, and the closed learning model 11 is trained based on the information about the various industrial machines.

[0081] In this embodiment, the question correction determination process is described as being executed by the control device 200 of the injection molding machine 20. However, the question correction determination process may be executed by the question answering device 10. In this case, the question correcting unit 207 is provided as a function in the question answering device 10, and the additional saved information is pre-stored in the storage device 150. In step S205, the question answering device 10 checks with the injection molding machine 20 that has sent the question whether it is possible to transmit the molding conditions and / or actual measured values. If it is possible to transmit the molding conditions and / or actual measured values, the injection molding machine 20 transmits the molding conditions and / or actual measured values ​​to the question answering device 10 as a response, and the question answering device 10 acquires them. The question correcting unit 207 provided in the question answering device 10 constitutes a part of the question answering device according to the present invention. Furthermore, the question correcting unit 207 may be provided in an intermediate portion connecting the information acquiring unit 101 and the judgment processing unit 102 of the question answering device 10 in FIG. 5 .

[0082] In the present embodiment, once a question is received, question information is generated, corrected, and then transmitted to the question answering device 10. However, depending on the content of the question, a question may be returned to the user to enhance the question information. For example, when a question such as "An abnormality has occurred during weighing. What should I do?" is received, the determination processing unit 204 determines whether an error code is included based on the word "abnormal." If it is determined that an error code is not included, the notification unit 205 may notify the user with a message such as, "Is an abnormality number displayed on the screen? If so, please tell me the number (number of digits)." This allows the error code to be included in the question. This may also prompt the user to respond regarding other conditions of the injection molding machine 20. Therefore, the task of the question correction unit 207 of the question answering device to return a question to the user, enhance the question information, and send it to the answer generation unit of the question answering device corresponds to the prompt engineering unit.

[0083] Furthermore, in this embodiment, the question answering device 10 has been described as having the open-type learning model 12 and the open-type database 170, but it may also be configured to have only the closed-type learning model 11 and the closed-type database 160 without these. In other words, the closed-type database 160 may have various publicly known information so that general questions can also be answered by the closed-type learning model 11; in other words, the open-type database 170 may be incorporated into the closed-type database 160.

[0084] In the question correction determination process of this embodiment, it has been explained that in order to acquire molding conditions and actual measurement values, these must be set so that they can be transmitted to the question answering device 10. However, molding conditions and actual measurement values ​​may be automatically acquired simultaneously with the worker's oral question. Furthermore, as described above, when the question answering device 10 has the question correction unit 207 and the question correction determination process is executed on the question answering device 10 side, it is preferable to perform a process of confirming the molding conditions and actual measurement values ​​of the injection molding machine 20 after acquiring the question information. For example, the question correction unit 207 of the question answering device 10 requests the molding conditions and actual measurement values ​​from the injection molding machine 20, and the injection molding machine 20 that receives the request transmits the molding conditions and actual measurement values ​​set at that time to the question answering device 10.

[0085] 12 , the question answering device 10 may be configured such that a first question answering device 10-1 comprising a closed database 160 and hardware such as a CPU 110, and a second question answering device 10-2 comprising an open database 170 and hardware such as a CPU 110 are installed in different locations and are communicatively connected to each other via a communication I / F over the Internet or the like. For example, a chatbot such as ChatGPT, which is the open database 170, may be stored on a server of a management company that operates the conversational AI, causing the server to function as the second question answering device 10-2. On the other hand, the first question answering device 10-1, such as a server comprising the closed database 160 and hardware such as a CPU 110, may be located at a manufacturer that produces the injection molding machine 20. Furthermore, the first question answering device 10-1, such as a server comprising the closed database 160 and hardware such as a CPU 110, may be located at a factory of a user that operates the injection molding machine 20.

[0086] Furthermore, the functions of the information acquisition unit 101, judgment processing unit 102, answer generation unit 103, verification processing unit 104, answer transmission unit 105, and learning update unit 106 of the question and answering apparatus 10, as well as each learning model, may be incorporated into the injection molding machine 20. In this case, the various functions of the question and answering apparatus 10 are realized by cooperation of the hardware resources of the injection molding machine 20, and the injection molding machine 20 is constructed as a question and answering apparatus according to the present invention. Naturally, only a part of the functions of the question and answering apparatus 10, such as only the first question and answering apparatus 10-1, such as a server constituting the closed database 160 and the hardware of the CPU 110, may be incorporated into the injection molding machine 20.

[0087] Second Embodiment Fig. 13 is a block diagram showing the functional configuration of a question and answering device according to this embodiment. Fig. 14 is a block diagram showing the functional configuration of a control device provided in an injection molding machine according to this embodiment.

[0088] This embodiment differs from the first embodiment in that the question and answering system 1 has a question and answering device 10A instead of the question and answering device 10. Also, this embodiment differs from the first embodiment in that the injection molding machine 20 has a control device 200A instead of the control device 200.

[0089] 13, the question answering device 10A differs from the question answering device 10 according to the first embodiment in that it includes an open-type answer generation unit 103A and a closed-type answer generation unit 103B instead of the answer generation unit 103. As shown in FIG. 14, the control device 200A differs from the control device 200 according to the first embodiment in that it further includes an information extraction unit 208.

[0090] Open-ended answer generation unit 103A in question and answering device 10A is an AI chatbot that generates answers to publicly known question information (described later) based on open-ended learning model 12. On the other hand, closed-ended answer generation unit 103B is an AI chatbot that generates answers to question information based on closed-ended learning model 11. Closed-ended answer generation unit 103B also integrates the two obtained answers and generates an answer document to be sent to injection molding machine 20.

[0091] The information extraction unit 208 in the control device 200A extracts publicly known information from the question information and generates publicly known question information.

[0092] (Question Answering Process) The question answering process performed by the question answering system according to this embodiment will be described in detail. Fig. 15 is a flowchart showing the question answering process according to this embodiment. The question answering process according to this embodiment is a process in which some processes have been added or changed from the question answering process according to the first embodiment. Here, only the added or changed processes will be picked up and explained.

[0093] As shown in FIG. 15 , after the question correction determination process, the information extraction unit 208 of the control device 200A extracts publicly known information from the question information to generate publicly known question information (S301). Here, publicly known information refers to information obtained by subtracting unique information from the question information. In other words, the portion of the question information from which highly confidential information such as error codes, molding condition information, actual measurement value information, and specification information has been removed is extracted as publicly known information. According to the question correction determination process, detailed explanations are added to the error codes in the question information. It is preferable that these detailed explanations be treated as publicly known information, and that information such as "(error code A1)" included in the question information Qu2 shown in FIG. 11 be treated as publicly known information and not extracted. After generating the publicly known question information, the information transmission unit 203 transmits the question information and the publicly known question information together with a personal identification code to the question and answering device 10A (S302).

[0094] The question answering device 10A is always in a standby state where it can acquire various information. When question information, publicly known question information, and a secret code are transmitted in this state, the information acquisition unit 101 of the question answering device 10 acquires (receives) them (S303). After acquisition, the open-type answer generation unit 103A generates an open-type answer document (first answer) corresponding to the publicly known question information based on the open-type learning model 12 (S304). After selecting the closed-type learning model 11, the closed-type answer generation unit 103B generates a closed-type answer document (second answer) corresponding to the question information based on the closed-type learning model 11 (S305). After generating the closed-type answer document, the closed-type answer generation unit 103B integrates the obtained open-type answer document and the closed-type answer document to generate one answer document (S306). Note that the question information from the closed-type learning model 11 may be not only a closed-type answer document (second answer) corresponding to conversational question information, but also closed-type answer data corresponding to the question data. In the above, "question data" refers to actual measurement value information (including historical information) and refers to a question that is not in a conversational format. Also, "answer data" refers to a response that is not in a conversational format and refers to numerical values ​​such as forming conditions. In the above case, the closed-type answer generation unit 103B also integrates the obtained open-type answer document and closed-type answer data to generate one answer document (S306).

[0095] Here, integration may mean adding a closed-type answer document (including closed-type answer data) to an open-type answer document, or may mean appropriately organizing the information. Examples of organizing the information include modifying the format of the answer document, deleting overlapping content, and deleting answers from the open-type answer document when conflicting answers are included. In particular, the format of an open-type answer document is likely to be easy for users to understand. Therefore, it is preferable to organize the information so as to supplement the answer content of the closed-type answer document while maintaining the format of the open-type answer document.

[0096] The subsequent processing is the same as the question answering processing according to the first embodiment, and therefore will not be described here. Furthermore, if it is determined in step S107 that the unique information is not included in the question information (S107, NO), the process proceeds to step S306. In this case, since there is no closed-type answer document to be integrated, an open-type answer document is the answer document to be sent to the injection molding machine 20.

[0097] According to this embodiment, since it is possible to generate both open-type response documents and closed-type response documents, it is possible to make the response documents more accurate.

[0098] It is also conceivable that a confidentiality agreement may be concluded with the company operating question and answering device 10A, stipulating that highly confidential information will not be leaked to the outside. In such cases, answers may be generated for question information based on open-type learning model 12, even if the question information contains unique information, without using publicly known question information. Even if an open-type learning model 12 from an external company is used, answers may be generated for question information containing unique information based on open-type learning model 12, as long as a confidentiality agreement has been concluded with the company. In the above case, highly confidential question information posed about injection molding machine 20, etc., is not imported into open database 170, and open database 170 can be used only for answers to question information.

[0099] In the above-described embodiments, the programs (question and answer programs related to industrial machinery) that realize the various functions of the question and answering apparatus 10 and the control device 200 have been described as being pre-installed in the question and answering apparatus 10 and the control device 200. However, the programs may be provided by recording them on a computer-readable recording medium in an installable or executable format. Here, the term "storage medium" refers to any medium that can be read or executed by the computer serving as the question and answering apparatus 10 and the control device 200, such as magnetic tape, magnetic disks (hard disk drives, etc.), optical disks (CD-ROMs, DVDs, etc.), magneto-optical disks (MOs, etc.), and flash memories, which are detachable from the control device 7, as well as media that can be transmitted via a network. Examples of media that can be transmitted via a network include media stored on a computer, such as an external server connected to a network such as the Internet, and provided via the network. At least some of the functions of the question and answering apparatus 10 and the control device 200 may be implemented by a server or the like external to the question and answering system 1, and data from the server may be provided to the question and answering apparatus 10 and the control device 200 via communication via the network.

[0100] Although several embodiments of the invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These novel embodiments may be embodied in various other forms, and various omissions, substitutions, and modifications may be made without departing from the spirit of the invention. These embodiments and their modifications are included within the scope and spirit of the invention, and are also included in the scope of the invention and its equivalents as defined in the claims.

[0101] 1 Question and answering system 10 Question and answering device 11 Closed learning model 12 Open learning model 16 Closed database 17 Open database 101 Information acquisition unit (question information acquisition unit, identification information acquisition unit, evaluation acquisition unit, operation information acquisition unit) 102 Determination processing unit (unique information determination unit, correction determination unit, verification determination unit) 103 Answer generation unit 103A Open answer generation unit 103B Closed answer generation unit 104 Verification processing unit (verification unit) 105 Learning update unit (update unit) 20 Injection molding machine 200 Control device 207 Question correction unit 241 Display (question and answering unit) 242 Input device (question reception unit) 243 Microphone (question reception unit) 244 Speaker (question and answering unit)

Claims

1. A question and answering device that answers questions about industrial machinery, comprising: a question information acquisition unit that acquires question information accepted by a question acceptance unit; a unique information determination unit that determines whether the acquired question information includes unique information about the industrial machinery; and an answer generation unit that, when it is determined that the question information includes the unique information, generates an answer to the question information based on a closed learning model that has been trained on information about the industrial machinery and information about the unique information, which can be provided only to a user of the industrial machinery.

2. A question and answer device as described in claim 1, further comprising an identification information acquisition unit that acquires identification information relating to a user of the industrial machine, wherein the answer generation unit generates answers to the extent that they can be disclosed to the user indicated in the acquired identification information.

3. A question answering device as described in claim 2, which has a plurality of the closed learning models, each of the plurality of closed learning models being individually associated with a user's identification information, and the answer generation unit selects a closed learning model associated with the acquired identification information and generates an answer based on the selected closed learning model.

4. The question and answering device described in claim 1, wherein the answer generation unit generates answers to the acquired question information based on an open-type learning model that has learned about publicly known information that can be provided not only to users of the industrial machinery but also to other users, and the closed-type learning model.

5. The question answering device described in claim 4, wherein the question information acquisition unit acquires, together with the question information, publicly known question information consisting of publicly known information extracted from the question, and the answer generation unit generates a first answer to the publicly known question information based on the open-type learning model, generates a second answer to the question information based on the closed-type learning model, and generates an answer to the question information by integrating the first and second answers.

6. A question answering device as described in claim 1, further comprising: a correction determination unit that determines whether or not to make a correction to the acquired question information; and a question correction unit that makes the correction to the question information when it is determined that a correction should be made to the acquired question information.

7. The question and answering device according to claim 1, further comprising: a verification determination unit that determines whether the answer generated by the answer generation unit is verifiable or not; and a verification unit that verifies the answer when it is determined that the answer is verifiable.

8. The question and answering device according to claim 1, further comprising: an evaluation acquisition unit that acquires an evaluation for the answer; and an update unit that updates the closed learning model based on the acquired evaluation.

9. A question and answering device as described in claim 1, further comprising an operation information acquisition unit that acquires operation information including operating conditions of the industrial machine for which the question was asked, and the answer generation unit generates the answer using the closed learning model and the operation information.

10. A question answering method for answering a question about industrial machinery, comprising: a question answering device acquires question information indicating the question; determines whether the acquired question information includes unique information about the industrial machinery; and if it is determined that the question information includes the unique information, generates an answer to the question information based on a closed learning model that has been trained on information about the industrial machinery and information about the unique information, which can be provided only to a user of the industrial machinery.

11. A question and answering program for answering questions about industrial machinery, which causes a computer to function as: a question information acquisition unit that acquires question information accepted by a question acceptance unit; a unique information determination unit that determines whether the acquired question information includes unique information about the industrial machinery; and an answer generation unit that, if it is determined that the question information includes the unique information, generates an answer to the question information based on a closed learning model that has been trained on information about the industrial machinery and information about the unique information that can be provided only to a user of the industrial machinery.

12. A question and answering system including industrial machinery which is at least one of an injection molding machine, an extruder, a press machine, and a semiconductor manufacturing machine, and a question and answering device which answers questions about the industrial machinery, wherein the industrial machinery comprises: a question receiving unit which receives questions from a user; and a transmitting unit which transmits the questions to the question and answering device as question information, and the question and answering device comprises: a question information acquisition unit which acquires the question information transmitted by the transmitting unit; a unique information determination unit which determines whether the acquired question information includes unique information about the industrial machinery; and an answer generation unit which, when it is determined that the question information includes the unique information, generates an answer to the question information based on a closed learning model which has been trained on information about the industrial machinery and information about the unique information, which can be provided only to a user of the industrial machinery.

13. An industrial machine that answers questions related to the industrial machine, the industrial machine being at least one of an injection molding machine, an extruder, a press machine, and a semiconductor manufacturing machine, comprising: a question receiving unit that receives questions from a user; a question information acquisition unit that acquires the question received by the question receiving unit as question information; and a question answering unit that provides an answer to the question to the user, wherein it is determined whether the question information includes unique information related to the industrial machine, and if it is determined that the question information includes the unique information, an answer to the question information is generated based on a closed learning model that is learned about information related to the industrial machine and information related to the unique information that can be provided only to a user of the industrial machine, and the question answering unit provides an answer to the question information.

Images