Auxiliary systems, auxiliary devices, and programs

Abstract

It provides a mechanism that can assist users' thinking based on existing technical information. [Solution] A system to assist a user's thinking, wherein the system has one or more processors, the processors receive technical information which is information about technology, display a graph which represents the technical information with multiple nodes and multiple edges, and when one of the multiple nodes in the graph is selected as a selected node, the system generates an extended node which is a new node that is technically related to the selected node, and an extended edge which is a new edge that shows the relationship between the selected node and the extended node, and displays an extended graph which adds the generated extended nodes and extended edges to the graph.

Description

Technical Field

[0006] , ,

[0005] ,

[0001] The present invention relates to an auxiliary system, an auxiliary device, and a program.

Background Art

[0002] In recent years, there has been a shortage of engineers including researchers, and the shrinkage of the human network has made it difficult to transmit and integrate knowledge. That is, the decrease in the number of engineers has led to a decrease in the number of people participating in discussions, resulting in a decline in awareness from the knowledge and statements of others. In addition, there is a concern that useful past technologies cannot be effectively utilized in current technology development because there are no personnel who know the past knowledge (for example, technical information, etc.) accumulated in the organization.

[0003] In addition, by generating a knowledge graph based on information collected from knowledge data including knowledge related to a specific field, even when not having knowledge related to the specific field, it is possible to grasp related matters by the connection of nodes in the knowledge graph and consider countermeasures and make decisions considering the contribution degree given to the nodes. A mechanism has been proposed (see, for example, Patent Document 1).

Prior Art Documents

Patent Documents

[0007] To this end, the present disclosure provides an auxiliary system for assisting a user's thinking, comprising one or more processors, wherein the processors receive technical information, which is information relating to a technology, display a graph representing the technical information with a plurality of nodes and a plurality of edges, and, in response to one of the plurality of nodes being selected as a selected node in the graph, generate an extended node, which is a new node that is technically related to the selected node, and an extended edge, which is a new edge that shows the relationship between the selected node and the extended node, and display an extended graph in which the generated extended node and extended edge are added to the graph. [Effects of the Invention]

[0008] According to the present invention, it is possible to provide a mechanism that can assist the user's thinking based on existing technical information. [Brief explanation of the drawing]

[0009] [Figure 1] This is a schematic diagram showing the configuration of the thinking assistance system according to this embodiment. [Figure 2] This is a block diagram showing the hardware configuration of the device included in the thinking assistance system according to this embodiment. [Figure 3] This is a functional block diagram showing an example of the functional configuration of a thinking assistance device. [Figure 4] This flowchart illustrates the basic operation of the extended graph display process in the thinking assistance device. [Figure 5] This flowchart illustrates other operations of the extended graph display process in the thinking assistance device. [Figure 6] This flowchart illustrates other operations of the extended graph display process in the thinking assistance device. [Figure 7] This is a flowchart illustrating the operation of node selection processing in a thinking assistance device. [Figure 8] This flowchart illustrates the operation of node selection processing and node explanation display processing in a thinking assistance device. [Figure 9] This is a flowchart illustrating the operation of the extended node switching display process in a thinking assistance device. [Figure 10] This is a flowchart illustrating the operation of the unknown knowledge data search process in a thinking assistance device. [Figure 11] This is a flowchart illustrating the operation of a thinking assistance device in which multiple pieces of knowledge are combined. [Figure 12] This is a schematic diagram showing an example of the configuration of the technical information input screen displayed on the engineer's terminal. [Figure 13] This is a schematic diagram illustrating an example of the configuration of the knowledge graph display screen shown on the engineer's terminal. [Figure 14] This is a schematic diagram showing an example configuration of the extended graph display screen shown on the engineer's terminal. [Figure 15-1] This schematic diagram shows an example of the configuration of an extended graph display screen, illustrating the steps for displaying the extended graph shown on the engineer's terminal as an educational aid for thinking. [Figure 15-2] This schematic diagram shows an example of the configuration of an extended graph display screen, illustrating the steps for displaying the extended graph shown on the engineer's terminal as an educational aid for thinking. [Figure 15-3] This schematic diagram shows an example of the configuration of an extended graph display screen, illustrating the steps for displaying the extended graph shown on the engineer's terminal as an educational aid for thinking. [Figure 15-4] This schematic diagram shows an example of the configuration of an extended graph display screen, illustrating the steps for displaying the extended graph shown on the engineer's terminal as an educational aid for thinking. [Figure 15-5] This schematic diagram shows an example of the configuration of an extended graph display screen, illustrating the steps for displaying the extended graph shown on the engineer's terminal as an educational aid for thinking. [Figure 16]It is a schematic diagram showing a configuration example of an extended graph display screen for explaining extended nodes and extended edges generated by selecting nodes in a graph displayed on an engineer terminal. [Figure 17-1] It is a schematic diagram showing a configuration example of an extended graph display screen for explaining a plurality of extended node candidates and extended edge candidates generated by selecting nodes in a knowledge graph displayed on an engineer terminal. [Figure 17-2] It is a schematic diagram showing a configuration example of an extended graph display screen for explaining a state in which one extended node candidate is selected in the knowledge graph shown in FIG. 17-1. [Figure 18] It is a schematic diagram showing a configuration example of an extended graph display screen for explaining technical explanations displayed when one extended node candidate is selected in the knowledge graph shown in FIG. 17-1. [Figure 19-1] It is a schematic diagram showing a configuration example of an extended graph display screen for explaining the switching of extended node candidates generated by selecting nodes in a knowledge graph displayed on an engineer terminal. [Figure 19-2] It is a schematic diagram showing a configuration example of an extended graph display screen for explaining the switching of extended node candidates generated by selecting nodes in a knowledge graph displayed on an engineer terminal. [Figure 19-3] It is a schematic diagram showing a configuration example of an extended graph display screen for explaining the switching of extended node candidates generated by selecting nodes in a knowledge graph displayed on an engineer terminal. [Figure 20-1] It is a schematic diagram showing a configuration example of an extended graph display screen for explaining the switching of extended node candidates generated by selecting edges in an extended graph displayed on an engineer terminal. [Figure 20-2] It is a schematic diagram showing a configuration example of an extended graph display screen for explaining the switching of extended node candidates generated by selecting edges in an extended graph displayed on an engineer terminal. [Figure 21] (A) A schematic diagram showing a graph of first technical information and (B) a schematic diagram showing a graph of second technical information used for explaining the generation of an extended graph. [Figure 22-1]This schematic diagram shows an example of the configuration of the extended graph display screen, illustrating the steps for displaying the extended graph on the engineer's terminal during the search for unknown knowledge data. [Figure 22-2] This schematic diagram shows an example of the configuration of the extended graph display screen, illustrating the steps for displaying the extended graph on the engineer's terminal during the search for unknown knowledge data. [Figure 22-3] This schematic diagram shows an example of the configuration of the extended graph display screen, illustrating the steps for displaying the extended graph on the engineer's terminal during the search for unknown knowledge data. [Figure 22-4] This schematic diagram shows an example of the configuration of the extended graph display screen, illustrating the steps for displaying the extended graph on the engineer's terminal during the search for unknown knowledge data. [Figure 22-5] This schematic diagram shows an example of the configuration of the extended graph display screen, illustrating the steps for displaying the extended graph on the engineer's terminal during the search for unknown knowledge data. [Figure 23-1] This schematic diagram shows an example of the configuration of the extended graph display screen, sequentially explaining other display procedures for the extended graph displayed on the engineer's terminal during the search for unknown knowledge data. [Figure 23-2] This schematic diagram shows an example of the configuration of the extended graph display screen, sequentially explaining other display procedures for the extended graph displayed on the engineer's terminal during the search for unknown knowledge data. [Figure 23-3] This schematic diagram shows an example of the configuration of the extended graph display screen, sequentially explaining other display procedures for the extended graph displayed on the engineer's terminal during the search for unknown knowledge data. [Figure 23-4] This schematic diagram shows an example of the configuration of the extended graph display screen, sequentially explaining other display procedures for the extended graph displayed on the engineer's terminal during the search for unknown knowledge data. [Figure 23-5] This schematic diagram shows an example of the configuration of the extended graph display screen, sequentially explaining other display procedures for the extended graph displayed on the engineer's terminal during the search for unknown knowledge data. [Figure 24-1] This schematic diagram shows an example of the configuration of the extended graph display screen, illustrating the steps for displaying the extended graph on the engineer's terminal when linking two pieces of knowledge. [Figure 24-2] This schematic diagram shows an example of the configuration of the extended graph display screen, illustrating the steps for displaying the extended graph on the engineer's terminal when linking two pieces of knowledge. [Figure 24-3] This schematic diagram shows an example of the configuration of the extended graph display screen, illustrating the steps for displaying the extended graph on the engineer's terminal when linking two pieces of knowledge. [Figure 24-4] This schematic diagram shows an example of the configuration of the extended graph display screen, illustrating the steps for displaying the extended graph on the engineer's terminal when linking two pieces of knowledge. [Modes for carrying out the invention]

[0010] Below, as an example of this embodiment, we will explain, with reference to the drawings, how it expands the scope based on existing knowledge (technical information) to provide clues for insight, assisting the thinking of engineers (an example of a user) in areas such as knowledge training for new engineers, searching for unknown knowledge data, and connecting two pieces of knowledge.

[0011] The embodiments described below are preferred examples of the present invention and are subject to various technical limitations. However, the scope of the present invention is not limited to these forms unless otherwise specified in the following description.

[0012] (System Configuration) Figure 1 is a diagram showing an overview of the configuration of the thinking assistance system according to this embodiment.

[0013] As shown in Figure 1, the thinking assistance system 1 in this embodiment (hereinafter sometimes referred to as "the System 1") comprises a thinking assistance device 10 that assists the user's thinking and a technician terminal 30. In the System 1, the thinking assistance device 10 and the technician terminal 30 are interconnected via a network N such as the Internet.

[0014] System 1, upon receiving technical information (hereinafter simply referred to as "technical information"), has its AI generate and display a graph (hereinafter referred to as "knowledge graph") representing that technical information using multiple nodes and multiple edges. Furthermore, System 1 has its AI generate and display a new knowledge graph (hereinafter referred to as "extended graph") by adding new nodes and edges to the existing knowledge graph. In this way, System 1 can provide engineers with clues to solutions in solving technical problems and assist their thinking.

[0015] In other words, System 1 is a thought-search assistance tool that does not have the generative AI process the problem all at once, but rather uses RAG (Retrieval Augmented Generation) with an AI agent to process the problem in multiple stages using a Large Language Model (LLM) with the generative AI at its core. More specifically, System 1 can use prompt engineering methods such as CoT (Chain of Thoughts) and ReAct (Reasoning and Action) in a knowledge graph that represents technical information with nodes (points) and edges (lines). Then, System 1 generates (displays) an extended graph with expanded nodes and edges in the knowledge graph based on the technical information stored in the database by having the LLM perform an inference step that extends the nodes and edges.

[0016] Specifically, for example, several nodes that can be explored in depth in the knowledge graph are identified, and depending on the selection of one of these nodes, a large concept is classified into several systems to generate new nodes, while new edges are extended to expand the knowledge graph. The extension of nodes and edges can be done using taxonomy extension techniques and ontology extension techniques, based on the relationship between node types and edges, such as causal relationships and parent-child relationships (higher-level concepts and lower-level concepts), using the internal knowledge of LLMs trained on scientific benchmarks such as GPQA. In this way, System 1 compensates for the weaknesses of the generation AI with RAG, thereby linking the generation AI with an external database to create highly accurate answers.

[0017] More specifically, basic RAG performs a hybrid search combining full-text search and vector (similarity) search as an unstructured data search. In LLM, the amount of information that can be processed at once is determined by the size of the context window, so the information (context) to be searched in RAG is divided into fixed-length chunks. Furthermore, these chunks are embedded and converted into vectors (numerical values) to make them easier for the generating AI to search, and then registered in an external database.

[0018] Therefore, full-text search finds chunks of text that exactly or partially match the keywords as the target information, while vector search uses an embedding model to convert chunks into numbers and finds chunks with similar numbers as the target information.

[0019] Furthermore, when searching for context to input into LLM from an unstructured database, vector searches may retrieve irrelevant contexts, potentially missing valuable ones. Therefore, by performing RAG using not only full-text and vector searches but also queries such as SQL and NoSQL against structured data (searching structured data), we can minimize the loss of valuable contexts and generate more accurate and reliable answers.

[0020] In short, unstructured databases are not suitable for data models that represent the complex structure of text context. Therefore, we use structured databases to represent the meaning of text with graph data, enabling a hybrid search RAG using keywords, vectors, and graphs. In this way, RAG generates answers by supplementing information that the LLM has not learned by referring to an external database.

[0021] Furthermore, structured databases represent the relationships between data using nodes and edges, and are databases that efficiently store and search information from multiple documents. They decompose input information into elements called nodes and edges using a method called GraphRAG, and display the relationships between nodes as graph data (knowledge graph). Therefore, GraphRAG utilizes structured databases to achieve output that reflects a richer context.

[0022] In this embodiment, for example, when creating a knowledge graph of technical information in the field of chemistry, technical elements such as substances and materials, reactions and processes, conditions and parameters, physical properties and performance, equipment and analytical methods are represented by individual "nodes". Furthermore, relationships between technical elements (nodes), such as reaction relationships, causal relationships, structural relationships, condition dependencies, and comparison relationships, are represented by "edges". By arranging these multiple nodes and edges in a graph, the target technical information can be represented as a knowledge graph. At this time, the target technical information can also be represented differently depending on the positional relationships of the nodes, such as the arrangement of nodes in the graph and the distance between nodes.

[0023] The thinking assistance device 10 has computational processing and communication functions and functions as a server in this system 1. In this embodiment, the thinking assistance device 10 is implemented by an electronic device such as a personal computer, such as a desktop or laptop computer, or a tablet computer. Therefore, in the following description, the thinking assistance device 10 may be referred to as a "server".

[0024] The engineer terminal 30 is a communication terminal used by engineers such as researchers and technology developers who use this system 1 as users, and is a device that has computational processing functions and communication functions. This engineer terminal 30 is connected to the thinking assistance device 10 via a communication network (network N), and is implemented by an electronic device such as a personal computer like a desktop or laptop, a tablet computer, or a smartphone.

[0025] Furthermore, while network N is comprised of the internet, it is not particularly limited and may be a communication network such as a LAN (Local Area Network) or a WAN (Wide Area Network). Also, the communication lines of network N may be wired or wireless, or both may be used in combination.

[0026] In the example shown in Figure 1, only one engineer terminal 30 is shown, but the number is not particularly limited. Furthermore, although only one thinking assistance device 10 is shown, the thinking assistance device 10 is not limited to a standalone device; it could also be a distributed server system or cloud server that operates cooperatively by communicating via a network N such as the internet.

[0027] (Hardware configuration) Figure 2 is a block diagram showing the hardware configuration of the thinking assistance device 10 according to this embodiment.

[0028] The thinking assistance device 10 includes a CPU (Central Processing Unit) 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, a bus 14, an input / output interface 15, an output unit 16, an input unit 17, a storage unit 18, and a communication unit 19.

[0029] The CPU 11 executes various processes according to the program stored in the ROM 12 or the program loaded from the memory unit 18 into the RAM 13. In other words, the CPU 11 is a processor that controls the operation of the entire thinking assistance device 10, executing various processes according to the program. The program is, for example, an operating system or firmware for operating each part of the thinking assistance device 10, or application software for realizing the functional blocks described later.

[0030] RAM13 also stores data and other information necessary for the CPU11 to perform various processes. The CPU 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output interface 15 is also connected to this bus 14. The input / output interface 15 is connected to an output unit 16, an input unit 17, a storage unit 18, and a communication unit 19.

[0031] The output unit 16 consists of a display, speakers, etc., and outputs various information as images and sounds. The input unit 17 consists of a keyboard, mouse, microphone, camera, etc., and inputs various information according to the received instructions and operations. For example, the output unit 16 and the input unit 17 may be integrated and implemented using a touch panel.

[0032] The memory unit 18 stores programs necessary for the thinking assistance device 10 to perform various information processing tasks, as well as various data used by those programs. It is composed of, for example, a hard disk or DRAM (Dynamic Random Access Memory). The memory unit 18 also functions to store LLMs. These LLMs are not limited to being stored in the thinking assistance device 10, but may be stored in external devices or systems. An LLM is an AI model that generates content such as text, images, audio, and video based on input instructions (prompts). The communication unit 19 communicates with other devices via a network N, including the Internet.

[0033] The thinking assistance device 10 may also be equipped with a drive that can be appropriately fitted with removable media such as magnetic disks, optical disks, magneto-optical disks, or semiconductor memory. This drive allows programs read from the removable media to be installed in the storage unit 18. Furthermore, the removable media can store various types of data stored in the storage unit 18, just as the storage unit 18 does.

[0034] Furthermore, a large-scale language model can be used in the generative AI used in this embodiment. The generative AI is obtained by training a deep neural network. The generative AI receives prompts containing instructions, as well as inference data such as audio data representing speech, text data representing text, and image data representing images. The generative AI infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference can be exemplified by analysis, classification, prediction, and / or summarization.

[0035] In this embodiment, prompts are input to specifically specify the information to be obtained so that the generation AI can output (generate) the desired information. Furthermore, in this embodiment, depending on the conditions, the prompts may include a specification of the information that will be used as material for generating the desired information. For example, the reception unit 110, generation unit 120, display unit 130, selection unit 140, and confirmation unit 150 input prompts to the functional unit that functions as the generation AI, referring to predetermined information and including content that will allow each component to obtain the desired information.

[0036] Furthermore, the generation AI of this embodiment can refer to organizational information limited to within an organization, such as a large amount of business documents and regulations accumulated within an organization, such as a company. In addition, the generation AI of this embodiment can also refer to the latest information from outside the organization that can be obtained via a network connected to outside the organization. The generation AI of this embodiment can generate information such as text based on organizational information while incorporating the latest information through external information retrieval. In other words, in this system 1, the generation AI can be used via RAG, a search-enhanced generation technology, to generate the desired information.

[0037] Furthermore, the hardware configuration of the engineer terminal 30 in this embodiment is the same as that of the thinking assistance device 10 described above.

[0038] (Thinking aid 10) Figure 3 is a block diagram illustrating an example configuration of the thinking assistance device 10, which functions as a server in this embodiment.

[0039] The thinking assistance device (server) 10 includes a reception unit 110, a generation unit 120, a display unit 130, a selection unit 140, and a confirmation unit 150. The thinking assistance device 10 also includes a technical information DB 210, a structured information DB 220, and an unstructured information DB 230.

[0040] The reception unit 110 is a means that has the function of receiving technical information. Specifically, the reception unit 110 receives information such as research and development materials such as monthly reports, technical papers, academic conference materials, and patent application materials as technical information. The technical information received by the reception unit 110 is stored in the technical information database (DB) 210. Furthermore, the reception unit 110 is a means that also has the function of receiving technical questions as technical information.

[0041] Furthermore, the reception unit 110 also has the function of receiving information from the user (engineer) regarding the selection of one node from among multiple nodes on the knowledge graph as the selected node. In other words, when an engineer makes a selection of a selected node in this system 1, the reception unit 110 receives one of the nodes selected by the engineer as the selected node.

[0042] Furthermore, the reception unit 110 also has the function of receiving (detecting) that a selected node has been selected (for example, by clicking or tapping). Furthermore, the reception unit 110 also has the function of receiving (detecting) that an extended edge has been selected (for example, by clicking or tapping).

[0043] Furthermore, the reception unit 110 also has the function of receiving information from the user (engineer) to select one as an extension node from among multiple extension node candidates on the knowledge graph. In other words, when an engineer selects an extension node from among multiple extension node candidates, the reception unit 110 receives one of the extension node candidates selected by the engineer as an extension node.

[0044] Furthermore, the reception unit 110 is a means that has the function of receiving multiple pieces of technical information. That is, the reception unit 110 receives multiple pieces of technical information. The reception of technical information may be newly received technical information, or it may be information to select arbitrary technical information from the technical information DB 210.

[0045] The generation unit 120 is a means that has the function of generating a knowledge graph in which technical information is represented by multiple nodes and multiple edges. That is, the generation unit 120 analyzes the technical information received by the reception unit 110 and generates a knowledge graph. The knowledge graph generated by the generation unit 120 is stored as structured data in the structured information database (DB) 220. The generation unit 120 also divides the technical information received by the reception unit 110 to generate fixed-length sentence fragments (chunks) and generates vector information by numerically converting (transforming) these chunks. The chunks and vector information generated by the generation unit 120 are stored as unstructured data in the unstructured information database (DB) 230.

[0046] Furthermore, the generation unit 120 also has the function of generating new nodes that are technically related to the selected node as extension nodes, and generating new edges that show the relationship between the selected node and the extension node as extension edges, in response to the selection of one node from among multiple nodes in the knowledge graph as the selected node. In other words, the generation unit 120 performs a hybrid search using RAG that combines full-text search using unstructured data and vector (similarity) search, as well as a graph search using structured data, in response to the selection of the selected node, and generates extension nodes and extension edges. In this process, when generating extension nodes, it is advisable to quantify the strength of the relationship with the selected node as a confidence level to prevent the generation of an unlimited number of nodes (extension nodes), and then set a threshold to adjust the number of nodes generated.

[0047] Furthermore, the generation unit 120 also has the function of repeatedly generating new extension nodes and new extension edges in response to selecting an extension node as a selected node in the extension graph. The generation unit 120 then generates the extended nodes and extended edges, for example, in relation to the selected node and other nodes that are located close to it. In other words, the generation unit 120 of this embodiment generates the extended nodes and extended edges based on the technical elements represented by the selected node and other nodes that are located close to it in the knowledge graph.

[0048] Here, a node that is close to the selected node can be defined as the node that is closest to the selected node among multiple nodes placed in the graph. Furthermore, a node that is close to the selected node can be defined as a node that is edge-connected to the selected node, a node that can be connected to the selected node by an edge (line) without intersecting with other edges, or a node that can be connected to a newly generated node by an edge without intersecting with other edges.

[0049] The generation unit 120 then causes the LLM to generate concepts that are predicted to be related to the selected node and the concepts (technical elements) contained in nodes that are located close to the selected node. In this way, the generation unit 120 of this embodiment generates new nodes as extension nodes that are associated with the technical elements of the selected node and other nodes.

[0050] Furthermore, the generation unit 120 is also a means that has the function of generating technically related extension nodes and extension edges based on the selected node selected by the selection unit 140. Furthermore, the generation unit 120 is also a means that has the function of generating extension nodes and extension edges that are technically related to the selected node received by the reception unit 110. Furthermore, in response to the reception of a selected node by the reception unit 110, the generation unit 120 generates several new nodes that are technically related to the selected node as candidates for extended nodes.

[0051] Furthermore, the generation unit 120 is also a means that detects a so-called mouseover, where a user selects one of the multiple extension node candidates displayed by the display unit 130 by, for example, placing (aligning) the mouse pointer over it, and generates technical explanatory information for that extension node candidate. Furthermore, the generation unit 120 also has the function of detecting a so-called mouseover, which occurs when a user selects one of the multiple extension node candidates displayed by the display unit 130 by, for example, placing (aligning) the mouse pointer over it, and then generating a new edge as an extension edge that indicates the relationship between the selected node and the extension node, with the selected extension node candidate being treated as an extension node.

[0052] Furthermore, the generation unit 120 also has the function of sequentially generating extension node candidates one by one in response to the selection of a selected node by the reception unit 110. Specifically, the generation unit 120 in this embodiment generates multiple extension node candidates that are technically related to the selected node. Then, the generation unit 120 displays the generated multiple extension node candidates one by one in response to the selection of a selected node.

[0053] Furthermore, the generation unit 120 generates a new extension node that is in a polar or parallel relationship with the displayed extension node, in response to the reception unit 110's acceptance of the selection of an extension edge. Here, a polar relationship refers to a technical relationship, such as "oxidation" versus "reduction," or "generation (synthesis)" versus "decomposition." A parallel relationship refers to a technical relationship, such as "chemical bond" versus "ionic bond," "covalent bond," or "metallic bond."

[0054] Furthermore, the generation unit 120 displays new edges connected to selected nodes in the knowledge graph according to the selection of selected nodes, and displays extended nodes according to operations performed on the new edges. Here, the generation unit 120 can display multiple new edges connected to the selected node in response to the selection of the selected node, and can display an extended node in response to the selection of one of the multiple edges. Furthermore, the generation unit 120 can accept move operations for new edges and display extended nodes according to the position of the new edge in the knowledge graph resulting from the move operation.

[0055] Furthermore, the generation unit 120 also has the function of generating additional extension nodes and extension edges, and generating a new extension graph with an increased number of extension nodes and extension edges, if the verification unit 150 determines that there are no other technical information knowledge graphs.

[0056] The display unit 130 is a means that has the function of displaying the knowledge graph generated by the generation unit 120. Furthermore, the display unit 130 also has the function of displaying a knowledge graph selected from the technical information DB 210. Furthermore, the display unit 130 also has the function of displaying a new graph, including the new extended nodes and new extended edges generated by the generation unit 120, as an extended graph. That is, each time the generation unit 120 generates new extended nodes and new extended edges, the display unit 130 repeatedly displays a new extended graph that includes both the newly generated extended nodes and extended edges.

[0057] Furthermore, the display unit 130 also has the function of displaying multiple new nodes that are technically related to the selected node as extension node candidates in response to the receipt of information for selecting a node. That is, the display unit 130 displays multiple extension node candidates generated by the generation unit 120. Furthermore, the display unit 130 also has the function of repeatedly displaying an extended graph in which the extended node (for example, the selected extended node candidate) and extended edges are added to the knowledge graph, depending on the selection of one extended node candidate from among a plurality of extended node candidates.

[0058] Furthermore, the display unit 130 also has the function of displaying a technical explanation of the selected extension node candidate when one of several extension node candidates is selected as an extension node. That is, the display unit 130 displays the technical explanation information generated by the generation unit 120 next to the selected extension node candidate. When the mouse pointer is moved away from the extension node (for example, when the selection is deselected), the display of the technical explanation disappears and the system returns to its original state.

[0059] Furthermore, the display unit 130 also has a function to hide other extension node candidates that were not selected when one of the multiple extension node candidates is selected as an extension node. That is, the display unit 130 detects that one of the multiple extension node candidates has been selected (for example, by clicking or tapping), maintains the display of only the selected extension node candidate, and cancels the display of the other extension node candidates that were not selected.

[0060] Furthermore, the display unit 130 also has a function of switching and displaying new nodes that are technically related to the selected node as extension node candidates, in response to the receipt of information for selecting a selected node. In other words, the display unit 130 sequentially switches and displays the extension node candidates generated by the generation unit 120.

[0061] Furthermore, the display unit 130 also has the function of displaying new extension nodes that are technically related to the selected node in the extended graph, in place of the selected extension node, depending on the selection of the extension edge. That is, the display unit 130 reverses the orientation of the extension edge and displays it according to the extension node generated by the generation unit 120.

[0062] Furthermore, the display unit 130 also has a function to display the knowledge graph and the extended graph of other technical information such that, if there is another knowledge graph representing other technical information, the technically related or matching nodes and the extended nodes of the extended graph overlap in at least a portion of the other technical information. In other words, if the confirmation unit 150 determines that there is another knowledge graph representing other technical information, the display unit 130 displays the two graphs together such that the technically related or matching nodes in that knowledge graph and the extended nodes of the extended graph overlap in at least a portion of the other knowledge graph. Furthermore, the display unit 130 also has the function of displaying a new extended graph with additional extended nodes and extended edges if there are no other knowledge graphs representing other technical information. In other words, the display unit 130 displays a new extended graph with additional extended nodes and extended edges generated by the generation unit 120.

[0063] The selection unit 140 is a means that has the function of selecting a node containing a more abstract concept (technological element) as a selected node from among multiple nodes in the knowledge graph. In other words, when the control unit in this system 1 automatically selects a selected node, the selection unit 140 selects a node with a more abstract concept that can be explored in depth, that is, a node that can generate more extension nodes, as a selected node.

[0064] Here, "abstract" is contrasted with "concrete," such as proper nouns like chemical names. "Abstract" can be understood as an expression that encompasses multiple things. "Abstract" can be seen as a higher-level concept that brings together many things. On the other hand, "concrete" can be seen as a lower-level concept that is individualized.

[0065] Furthermore, the selection unit 140 of this embodiment uses prompts to the LLM(AI) that include instructions such as "compare the concepts (technical elements) assigned to each of the multiple nodes and select the node with the most abstract concept as the selected node." In addition, using the selected node selected by the selection unit 140, the generation unit 120 generates an extended node that has subordinate concepts that more concretely represent the concept possessed by the selected node.

[0066] The verification unit 150 is a means that, each time the extended graph is displayed, checks for the existence of other technical information that has nodes that are technically related to or match the newly generated extended nodes. In other words, each time an extended graph including new extended nodes and new extended edges generated by the generation unit 120 is generated (displayed), the verification unit 150 repeatedly checks for the existence of knowledge graphs of other technical information that have nodes that are technically related to or match the newly generated extended nodes by referring to the structured information DB 220.

[0067] The verification unit 150 in this embodiment receives multiple pieces of technical information relating to the technology, generates and displays knowledge graphs (for example, one graph and another graph) for each piece of information. Whenever a new extended graph is displayed in the first graph, with further extended nodes and extended edges generated, the verification unit 150 identifies in the other graph whether there are any nodes that are technically related to or match the newly generated extended nodes. The verification unit 150 displays one graph and the other graph such that, if there are technically related or matching nodes in the other graph, at least a portion of those nodes and the extended nodes overlap. On the other hand, if there are no technically related or matching nodes in the other graph, the verification unit 150 displays a new extended graph in which further extended nodes and extended edges have been generated in the first graph, and then re-identifies in the other graph whether there are any nodes that are technically related or matching to the newly generated extended nodes.

[0068] Technical Information DB210 is a means of storing unstructured technical information, such as research and development materials like monthly reports, technical papers, conference materials, and patent application documents. In other words, Technical Information DB210 stores the source information for the information stored in Structured Information DB220 and Unstructured Information DB230.

[0069] The structured information DB220 is a means of extracting technical information stored in the technical information DB210 and storing knowledge graph information represented by nodes and edges for each piece of technical information. The knowledge graph information is stored as graph data that directly represents the relationships between data using node information, edge information, and properties (attribute information).

[0070] The unstructured information DB230 is a means of extracting technical information stored in the technical information DB210 and storing unstructured data, which is divided into fixed-length chunks for each piece of technical information, and unstructured data obtained by converting the chunks into vectors (numerical values), in association with each other. Alternatively, the technical information stored in the technical information DB210 and the chunk information and vector information stored in the unstructured information DB230 may be stored together as unstructured information without being separated.

[0071] As described above, the thinking assistance device 10 according to this embodiment functions as a device including a reception unit 110, a generation unit 120, a display unit 130, a selection unit 140, and a confirmation unit 150, when various programs (OS, applications, etc.) stored in the auxiliary storage device are loaded into the main storage device and executed by the CPU 11.

[0072] Next, we will describe the flow of the extended graph display processing in the thinking assistance system 1 according to one embodiment of the present invention. Figure 4 is a flowchart illustrating the basic operation of the extended graph display process in the thinking assistance device 10.

[0073] As shown in Figure 4, in step S301, the reception unit 110 performs a process to determine whether or not it has received the technical information. If the reception unit 110 receives technical information (YES in S301), the process proceeds to step S302. On the other hand, if the reception unit 110 does not receive technical information (NO in S301), the process in step S301 is repeated.

[0074] In step S302, the generation unit 120 performs a process to generate a graph in which the technical information received by the reception unit 110 is represented by multiple nodes and multiple edges. Next, in step S303, the display unit 130 performs the process of displaying the graph generated by the generation unit 120. Furthermore, in step S304, the reception unit 110 performs a process to determine whether or not it has received an instruction to expand the graph displayed by the display unit 130. If the reception unit 110 receives a command to expand the graph (YES in S304), the process proceeds to step S305. On the other hand, if the reception unit 110 does not receive a command to expand the graph (NO in S304), the operation of the thinking assistance device 10 ends (END).

[0075] In step S305, the selection unit 140 performs the process of selecting (identifying) one node that can be expanded (deepened) from among the multiple nodes that make up the graph as the selected node. In step S306, the generation unit 120 performs a process to generate extension nodes that are technically related to the node (selected node) selected (identified) by the selection unit 140. Furthermore, in step S307, the generation unit 120 performs a process to generate extended edges that indicate relationships with extended nodes. Then, in step S308, the display unit 130 performs the process of displaying a new extended graph that includes the extended nodes and extended edges generated by the generation unit 120. After that, the operation of the thinking assistance device 10 ends (END).

[0076] Next, we will explain the processing flow when the extension process that generates extension nodes and extension edges in this system 1 is performed not just once, but repeatedly. Figure 5 is a flowchart illustrating other operations of the extended graph display process in the thinking assistance device 10.

[0077] As shown in Figure 5, the processes from step S401 to step S408 are the same as those described above from step S301 to step S308, so their explanation will be omitted. In step S409, the reception unit 110 performs a process to determine whether or not it has received an instruction to terminate the further expansion of the extended graph displayed by the display unit 130. If the reception unit 110 receives a command to terminate the extension of the extension graph (YES in S409), the operation of the thinking assistance device 10 ends (END). On the other hand, if the reception unit 110 does not receive a command to terminate the extension of the extension graph (NO in S409), the process returns to step S406 and repeats the subsequent processes of generating extension nodes (step S406), generating extension edges (step S407), displaying a new extension graph with the generated extension nodes and extension edges added (step S408), and determining whether the extension has ended (step S409).

[0078] Next, we will explain the processing flow in System 1 when it receives instructions at each expansion stage and automatically performs a series of expansion processes. Figure 6 is a flowchart illustrating other operations of the extended graph display process in the thinking assistance device 10.

[0079] As shown in Figure 6, the processes from step S501 to step S504 are the same as those described above from step S301 to step S304, so their explanation will be omitted. In step S505, the selection unit 140 performs a process to automatically select (identify) a node representing a more abstract concept as the selected node.

[0080] Furthermore, in step S506, the generation unit 120 performs the same process as in step S306 described above to generate an extension node that is technically related to the node (selected node) selected (identified) by the selection unit 140. Furthermore, in step S507, the generation unit 120 performs the same process as in step S307 described above to generate an extended edge that indicates the relationship with the extended node. Then, in step S508, the display unit 130 performs the same process as in step S308 described above, to display a new extended graph that includes the extended nodes and extended edges generated by the generation unit 120.

[0081] Then, returning to step S504, the reception unit 110 performs a process to determine whether or not it has received an instruction to further expand the graph (extended graph) displayed by the display unit 130, similar to step S304 described above. If the reception unit 110 receives an instruction to expand the graph (YES in S504), it proceeds to step S505, where it selects (identifies) the selected node, and then repeatedly performs subsequent processes such as generating the expanded node (step S506), generating the expanded edge (step S507), displaying the new expanded graph with the generated expanded node and expanded edge added (step S508), and determining whether or not an expansion instruction has been given (step S504). On the other hand, if the reception unit 110 has not received an instruction to expand the graph (NO in S504), the operation of the thinking assistance device 10 ends (END).

[0082] Next, we will explain the processing flow in System 1 when receiving instructions at each stage of the knowledge (technical) training for new engineers and other personnel, and manually performing a series of extension processes. Figure 7 is a flowchart illustrating the operation of the node selection process in the thinking assistance device 10.

[0083] As shown in Figure 7, the processes from step S601 to step S603 are the same as those described in steps S301 to step S303, so their explanation will be omitted. In step S604, the reception unit 110 performs a process to determine whether or not it has received an instruction to select a node in order to expand the graph displayed by the display unit 130. If the reception unit 110 receives an instruction to select a node (YES in S604), the process proceeds to step S605. On the other hand, if the reception unit 110 does not receive an instruction to select a node (NO in S604), the operation of the thinking assistance device 10 ends (END). In step S605, the selection unit 140 performs a process to identify the selected node as the selected node.

[0084] Furthermore, in step S606, the generation unit 120 performs the same process as in step S306 described above to generate an extension node that is technically related to the node (selected node) selected (identified) by the selection unit 140. Furthermore, in step S607, the generation unit 120 performs the same process as in step S307 described above to generate an extended edge that indicates the relationship with the extended node. Then, in step S608, the display unit 130 performs the same process as in step S308 described above, to display a new extended graph that includes the extended nodes and extended edges generated by the generation unit 120.

[0085] After that, returning to step S604, the reception unit 110 determines whether or not it has received an instruction to select a node in order to further expand the graph (extended graph) displayed by the display unit 130, and then repeats the subsequent processing.

[0086] Next, we will explain the processing flow in System 1 when displaying multiple extension node candidates, accepting the selection of an extension node from among them, and displaying explanations of the extension node candidates. Figure 8 is a flowchart illustrating the operation of node selection processing and node explanation display processing in the thinking assistance device 10.

[0087] As shown in Figure 8, the processes from step S701 to step S705 are the same as those described in steps S601 to S605, so their explanation will be omitted. In step S706, the display unit 130 performs a process to display multiple extension node candidates based on the node identified in step S705. Next, in step S707, the reception unit 110 performs a process to determine whether or not it has received an instruction to select one of the multiple extension node candidates displayed by the display unit 130. If the reception unit 110 receives an instruction to select an extension node candidate (YES in S707), the process proceeds to step S708. On the other hand, if the reception unit 110 has not received an instruction to select an extension node candidate (NO in S707), the process in step S707 is repeated.

[0088] In step S708, the display unit 130 performs the process of displaying a technical explanation of the selected extension node candidate. Furthermore, in step S709, the reception unit 110 performs a process to determine whether or not it has received an instruction to confirm the selection of an extension node candidate. If the reception unit 110 receives an instruction to confirm the selection of an extension node candidate (YES in S709), the process proceeds to step S710. On the other hand, if the reception unit 110 has not received an instruction to confirm the selection of an extension node candidate (NO in S709), the process returns to step S707 and repeats the subsequent processing.

[0089] In step S710, the selection unit 140 performs a process to identify the selected extension node candidate as an extension node. Furthermore, in step S711, the display unit 130 performs a process to hide all extension node candidates other than the selected extension node candidate. In this embodiment, the display unit 130 hides other extension node candidates that are not selected, thereby suppressing the divergence of the user's thinking as they consider the technical content of the knowledge graph. Furthermore, in step S712, the generation unit 120 performs the same process as in step S607 described above to generate extended edges that indicate relationships with extended nodes. Then, in step S713, the display unit 130 performs a process to display a new extended graph which is the extended node identified by the selection unit 140 and the extended edge generated by the generation unit 120.

[0090] After that, returning to step S704, the reception unit 110 determines whether or not it has received an instruction to select a node in order to further expand the graph (extended graph) displayed by the display unit 130, and then repeats the subsequent processing.

[0091] Next, we will explain the processing flow when switching between and displaying multiple extension node candidates in this system 1. Figure 9 is a flowchart illustrating the operation of the extended node switching display process in the thinking assistance device 10.

[0092] As shown in Figure 9, the processes from step S801 to step S805 are the same as those described above from step S601 to step S605, so their explanation will be omitted. In step S806, the display unit 130 performs the process of displaying one candidate extension node based on the selected node identified in step S805. Next, in step S807, the reception unit 110 performs a process to determine whether or not it has received an instruction to select (identify) an extended node candidate displayed by the display unit 130.

[0093] If the reception unit 110 receives an instruction to confirm the selection of an extension node candidate (YES in S807), the process proceeds to step S808. On the other hand, if the reception unit 110 has not received an instruction to confirm the selection of an extension node candidate (for example, it has received an instruction to select the selected node identified in step S805) (NO in S807), the process returns to step S806, displays a different extension node candidate, and repeats the subsequent processing.

[0094] In step S808, the selection unit 140 performs a process to identify the selected (identified) extension node candidate as an extension node. Furthermore, in step S809, the generation unit 120 performs the same process as in step S607 described above to generate extended edges that indicate relationships with extended nodes. Then, in step S810, the display unit 130 performs a process to display a new extended graph which is the extended node identified by the selection unit 140 and the extended edge generated by the generation unit 120.

[0095] After that, returning to step S804, the reception unit 110 determines whether or not it has received an instruction to select a node in order to further expand the graph (extended graph) displayed by the display unit 130, and then repeats the subsequent processing.

[0096] Next, we will explain the processing flow when searching for unknown knowledge data in this system 1. Figure 10 is a flowchart illustrating the operation of the unknown knowledge data search process in the thinking assistance device 10.

[0097] As shown in Figure 10, the processes from step S901 to step S908 are the same as those described above from step S301 to step S308, so their explanation will be omitted. In step S909, the verification unit 150 refers to the structured information DB220 and performs a process to check whether there is other technical information that has nodes that are technically related to or match the newly generated extended node in the displayed extended graph. If the verification unit 150 determines that there is other technical information (YES in S909), the process proceeds to step S910. On the other hand, if the verification unit 150 determines that there is no other technical information (NO in S909), the process proceeds to step S911.

[0098] In step S910, the display unit 130 performs a process to display the two graphs together such that at least a portion of the nodes of other technically matching (related) technical information and the extended nodes of the extended graph displayed by the display unit 130 overlap. After that, the operation of the thinking assistance device 10 ends (END).

[0099] Furthermore, in step S911, the selection unit 140 performs a process to select (identify) the extended node generated in step S906 as a selected node. Then, the process returns to step S906 and repeats subsequent steps such as generating extension nodes, generating extension edges (step S907), displaying a new extension graph with the generated extension nodes and extension edges added (step S908), and determining whether there is other technical information (step S909).

[0100] Next, we will explain the processing flow when combining multiple pieces of knowledge in this system 1. Figure 11 is a flowchart illustrating the operation of the thinking assistance device 10 in which multiple pieces of knowledge are combined.

[0101] As shown in Figure 11, in step S1001, the reception unit 110 performs a process to determine whether or not it has received an instruction to select multiple pieces of technical information. The multiple technical information pieces to be selected can be exemplified by, for example, one technical information piece and other technical information pieces specifically designated by the user. Furthermore, if, for example, keyword search or vector search is used, there are matching keywords or concepts between one technical information piece and other technical information pieces, or if there are similar keywords or concepts, these one technical information piece and other technical information pieces can be selected as multiple technical information pieces.

[0102] If the reception unit 110 receives an instruction to select technical information (YES in S1001), the process proceeds to step S1002. On the other hand, if the reception unit 110 does not receive an instruction to select technical information (NO in S1001), the process in step S1001 is repeated.

[0103] In step S1002, the generation unit 120 performs the process of generating a graph for each piece of technical information received by the reception unit 110. Next, in step S1003, the display unit 130 performs the process of displaying together the graphs generated by the generation unit 120. Furthermore, in step S1004, the reception unit 110 performs a process to determine whether or not it has received an instruction to select one node as a selected node in any of the graphs displayed by the display unit 130. If the reception unit 110 receives an instruction to select a node (YES in S1004), the process proceeds to step S1005. On the other hand, if the reception unit 110 does not receive an instruction to select a node (NO in S1004), the process in step S1004 is repeated.

[0104] In step S1005, the selection unit 140 performs the same process as in step S605 described above to identify the selected node as the selected node. Furthermore, in step S1006, the generation unit 120 performs the same process as in step S606 described above to generate an extension node that is technically related to the node (selected node) selected (identified) by the selection unit 140. Furthermore, in step S1007, the generation unit 120 performs the same process as in step S607 described above to generate extended edges that indicate relationships with extended nodes. Then, in step S1008, the display unit 130 performs the same process as in step S608 described above, displaying a new extended graph that includes the extended nodes and extended edges generated by the generation unit 120.

[0105] Furthermore, in step S1009, the verification unit 150 performs a process to verify whether the newly generated extended node in the displayed extended graph is technically related to or matches any node in any of the other displayed graphs. If the verification unit 150 determines that the extension node matches (is related to) any node (YES in S1009), the process proceeds to step S1010. On the other hand, if the verification unit 150 determines that the extension node does not match (is related to) any node (NO in S1009), the process returns to step S1004, where the reception unit 110 determines whether or not it has received an instruction to select a node to further extend the graph (extended graph) displayed by the display unit 130, and then repeats the subsequent process.

[0106] Then, in step S1010, the display unit 130 performs a process to display the two graphs together such that at least a portion of the nodes of the extended graph overlap with the nodes of other technically matching (related) graphs. After that, the operation of the thinking assistance device 10 ends (END).

[0107] As described above, System 1 is designed to allow for the acquisition of clues without limiting the scope, even if the number of researchers and engineers decreases, by expanding nodes and edges based on existing knowledge (technical information). Therefore, it is expected to serve as a tool to support engineers' thinking, such as for knowledge (technical) education for new engineers, searching for unknown knowledge data, and connecting two pieces of knowledge, thereby providing a catalyst for breakthroughs.

[0108] Next, an example of the screen configuration displayed on the engineer's terminal 30 in this embodiment will be described. First, in order to use System 1, the engineer accesses the System 1 site using the communication function of the engineer terminal 30, installs (downloads) the dedicated application software, and sets up personal identification information (ID) and a password.

[0109] When connecting to this system 1 from the engineer's terminal 30, the first screen displayed can be, for example, the one shown in Figure 12. Figure 12 is a schematic diagram showing an example configuration of the technical information input screen 300 displayed on the engineer's terminal 30.

[0110] In Figure 12, the technical information input screen 300 includes a research objective input field 301, a data file input section 302, a technical field input field 303, and a "Next: Build Basic Graph" button 304. The research objective input field 301 is used to input technical information as a question, for example, in the education of new engineers or the exploration of unknown knowledge data. This research objective input field 301 is equipped with a pull-down menu button 3011 for selecting accumulated technical information instead of a question, such as when linking two pieces of knowledge. The data file input section 302 is used to input technical information as research results. The technology field input field 303 is used to input the technology field of the technical information entered in the research objective input field 301 and the data file input section 302. The "Next: Construct basic graph" button 304 instructs the system to generate a graph of the entered technical information.

[0111] In Figure 12, the research objective input field 301 is shown to have the question, "In the synthesis of ether by SN2 reaction, we investigated the reaction temperature and solvent amount, but were unable to achieve the expected 80% yield. We would like to know what further conditions to investigate," and the technical field input field 303 is shown to have "Investigation of SN2 reaction conditions" entered.

[0112] If the "Next: Build Basic Graph" button 304 is selected in this state, for example, the knowledge graph display screen 310 shown in Figure 13 will be displayed. Figure 13 is a schematic diagram showing an example configuration of the knowledge graph display screen 310 displayed on the engineer's terminal 30.

[0113] In Figure 13, the knowledge graph display screen 310 is equipped with a knowledge graph 311 and an "Expand" button 314. This knowledge graph 311 consists of a first node 3121 labeled "ether synthesis," a second node 3122 labeled "yield," a first edge 3131 showing the relationship between the first node 3121 and the second node 3122, a third node 3123 labeled "temperature," a second edge 3132 showing the relationship between the first node 3121 and the third node 3123, a fourth node 3124 labeled "solvent amount," and a third edge 3133 showing the relationship between the first node 3121 and the fourth node 3124.

[0114] Furthermore, in the knowledge graph 311, the first edge 3131 is shown by an arrow pointing from the first node 3121 to the second node 3122, indicating the relationship between the first node 3121 and the second node 3122, where "yield" is measured by "ether synthesis." The second edge 3132 is shown by an arrow pointing from the third node 3123 to the first node 3121, indicating the relationship between the first node 3121 and the third node 3123, where "temperature" affects "ether synthesis." The third edge 3133 is shown by an arrow pointing from the fourth node 3124 to the first node 3121, indicating the relationship between the first node 3121 and the fourth node 3124, where "solvent amount" affects "ether synthesis."

[0115] When the "Expand" button 314 is selected while this knowledge graph 311 is displayed, for example, the expanded graph display screen 320 shown in Figure 14 will be displayed. In other words, when explaining development results and past technical information to new engineers, etc., it is expected that this system 1 will be used to ask questions and display knowledge data (expanded graph) that leads to the answers, thereby assisting their thinking process, with the expectation that knowledge (technical) education will be provided. Figure 14 is a schematic diagram showing an example configuration of the extended graph display screen 320 displayed on the engineer's terminal 30.

[0116] In Figure 14, the extended graph display screen 320 is equipped with an extended graph 321, an "Export Results" button 325, and a "Start New Analysis" button 326. This extended graph 321 shows the result of an automatic series of extensions to the knowledge graph 311 shown on the knowledge graph display screen 310, in response to a question entered in the research objective input field 301 of the technical information input screen 300.

[0117] Extended graph 321 shows the first extended node 3211, labeled "SN2 reaction," which is generated based on the first node 3121, "ether synthesis," and an arrow pointing from the first node 3121 to the first extended node 3211, which is generated as an edge showing the relationship between the first node 3121 and the first extended node 3211, indicating that "ether synthesis" is classified as an "SN2 reaction." Additionally, it shows the second extended node 3212, labeled "reaction rate," which is generated in relation to the first extended node 3211 and the second node 3122, and an arrow pointing from the first extended node 3211 to the second extended node 3212, which is generated as an edge showing the relationship between the first extended node 3211 and the second extended node 3212, and an arrow pointing from the second extended node 3212 to the second node 3122, which is generated as an edge showing the relationship between the second extended node 3212 and the second node 3122.

[0118] Furthermore, a third extension node 3213, labeled "solvent effect," is generated based on the first extension node 3211, and an arrow is shown pointing from the third extension node 3213 to the first extension node 3211, which is generated as an edge indicating the relationship between the first extension node 3211 and the third extension node 3213. Additionally, a fourth extension node 3214, labeled "solvent," is generated based on the third extension node 3213, and an arrow is shown pointing from the fourth extension node 3214 to the third extension node 3213, which is generated as an edge indicating the relationship between the third extension node 3213 and the fourth extension node 3214.

[0119] Furthermore, the fifth extension node 3215, labeled "non-protic polar solvent," and the sixth extension node 3216, labeled "protic polar solvent," are generated in relation to the first extension node 3211 and the fourth extension node 3214. An arrow is generated from the fifth extension node 3215 to the first extension node 3211 as an edge indicating the relationship between the first extension node 3211 and the fifth extension node 3215, and the relationship between the fourth extension node 3214 and the fifth extension node 3215 is also shown. An arrow is shown pointing from the fifth extension node 3215, which was generated as an edge, to the fourth extension node 3214; an arrow is shown pointing from the sixth extension node 3216, which was generated as an edge indicating the relationship between the first extension node 3211 and the sixth extension node 3216, to the first extension node 3211; and an arrow is shown pointing from the sixth extension node 3216, which was generated as an edge indicating the relationship between the fourth extension node 3214 and the sixth extension node 3216, to the fourth extension node 3214.

[0120] Furthermore, the seventh extension node 3217, named "MeCN (acetonitrile)", and the eighth extension node 3218, named "DMF (dimethylformamide)", which were generated based on the fifth extension node 3215, are shown, along with an arrow pointing from the seventh extension node 3217 to the fifth extension node 3215, which was generated as an edge showing the relationship between the fifth extension node 3215 and the seventh extension node 3217, and an arrow pointing from the eighth extension node 3218 to the fifth extension node 3215, which was generated as an edge showing the relationship between the fifth extension node 3215 and the eighth extension node 3218.

[0121] This extended graph 321 shows that to achieve an 80% yield in the synthesis of ethers by SN2 reactions, the conditions should be considered using aprotic polar solvents such as "MeCN" or "DMF," and this can be used to explore unknown knowledge data. In other words, by expanding the scope of existing knowledge (technical information) by using not just one piece of technical information but multiple pieces of technical information, clues to insights can be obtained, and it is expected that thinking will be efficiently and effectively supported.

[0122] Then, the engineer, as a user of System 1, selects the "Export Results" button 325 to output the results shown on the extended graph display screen 320. Furthermore, if you wish to ask another question, selecting the "Start a new analysis" button 326 will display the technical information input screen 300 shown in Figure 12.

[0123] Furthermore, in this system 1, instead of automatically expanding the knowledge graph 311 shown on the knowledge graph display screen 310 in response to the questions entered in the research objective input field 301 of the technical information input screen 300, as described above, it is also possible to expand it sequentially based on the instructions of an engineer, for the purpose of knowledge (technology) education for new engineers, etc. That is, by selecting the "Expand" button 314 on the knowledge graph display screen 310 shown in Figure 13, the expanded graph display screens 330 shown in Figures 15-1 to 15-5 can be displayed sequentially depending on the settings. Figures 15-1 to 15-5 are schematic diagrams illustrating the configuration of an extended graph display screen 330, which sequentially explains the procedure for displaying the extended graph shown on the engineer terminal 30 in educational support for thinking.

[0124] In Figure 15-1, the extended graph display screen 330 is equipped with an extended graph 331, a "Show Next Node" button 336, and an "Exit" button 337. This extended graph 331 is the first thing that appears when the "Extend" button 314 is selected on the knowledge graph display screen 310 shown in Figure 13. The extended graph 331 shows the first extended node 3311, which is "SN2 reaction" generated based on the first node 3121, "ether synthesis," and the first edge 3321, which is an arrow pointing from the first node 3121 to the first extended node 3311, indicating the relationship between the first node 3121 and the first extended node 3311.

[0125] Therefore, if the engineer wishes to further expand this extended graph 331 (exploring ideas), they select the "Show next node" button 336, while if they wish to stop expanding this extended graph 331 (generating new extended nodes), they select the "Finish" button 337. If the "Show next node" button 336 is selected at this point, for example, the extended graph display screen 330 shown in Figure 15-2 will continue to be displayed.

[0126] In Figure 15-2, the extended graph display screen 330 is equipped with an extended graph 332, a "Show Next Node" button 336, and an "Exit" button 337. This extended graph 332 shows a second extended node 3312 labeled "reaction rate" which is generated based on the first extended node 3311, "SN2 reaction," a second edge 3322 indicating the relationship between the first extended node 3311 and the second extended node 3312 as an arrow pointing from the first extended node 3311 to the second extended node 3312, and a third edge 3323 indicating the relationship between the second node 3122 and the second extended node 3312 as an arrow pointing from the second extended node 3312 to the second node 3122 labeled "yield." Furthermore, the extended graph 332 shows a third extended node 3313 labeled "solvent effect" which is generated based on the first extended node 3311, "SN2 reaction," and a fourth edge 3324 indicating the relationship between the first extended node 3311 and the third extended node 3313 as an arrow pointing from the third extended node 3313 to the first extended node 3311.

[0127] If the "Finish" button 337 is selected, further expansion of the extended graph 332 will end. If the "Show next node" button 336 is selected, for example, the extended graph display screen 330 shown in Figure 15-3 will continue to be displayed.

[0128] In Figure 15-3, the extended graph display screen 330 is equipped with an extended graph 333, a "Show Next Node" button 336, and an "Exit" button 337. This extended graph 333 shows a fourth extended node 3314, "solvent," which is generated based on the third extended node 3313, "solvent effect," and a fifth edge 3325, which is an arrow pointing from the fourth extended node 3314 to the third extended node 3313, indicating the relationship between the third extended node 3313 and the fourth extended node 3314.

[0129] If the "Finish" button 337 is selected, further expansion of the extended graph 333 will end. If the "Show next node" button 336 is selected, for example, the extended graph display screen 330 shown in Figure 15-4 will continue to be displayed.

[0130] In Figure 15-4, the extended graph display screen 330 is equipped with an extended graph 334, a "Show Next Node" button 336, and an "Exit" button 337. This extended graph 334 shows a fifth extended node 3315, which is "aprotic polar solvent" generated based on the first extended node 3311, "SN2 reaction," and the fourth extended node 3314, "solvent," and a sixth extended node 3316, "protic polar solvent." The extended graph 334 also shows a sixth edge 3326, which is an arrow pointing from the fifth extended node 3315 to the first extended node 3311, indicating the relationship between the first extended node 3311 and the fifth extended node 3315, and a seventh edge 3327, which is an arrow pointing from the fifth extended node 3315 to the fourth extended node 3314, indicating the relationship between the fourth extended node 3314 and the fifth extended node 3315. Furthermore, the extended graph 334 shows an eighth edge 3328, which represents the relationship between the first extended node 3311 and the sixth extended node 3316, as an arrow pointing from the sixth extended node 3316 to the first extended node 3311, and a ninth edge 3329, which represents the relationship between the fourth extended node 3314 and the sixth extended node 3316, as an arrow pointing from the sixth extended node 3316 to the fourth extended node 3314.

[0131] If the "Finish" button 337 is selected, further expansion of the extended graph 334 will end. If the "Show next node" button 336 is selected, for example, the extended graph display screen 330 shown in Figure 15-5 will continue to be displayed.

[0132] In Figure 15-5, the extended graph display screen 330 is equipped with an extended graph 335, an "Export Results" button 338, and a "Start New Analysis" button 339. This extended graph 335 shows the seventh extended node 3317, "MeCN (acetonitrile)", and the eighth extended node 3318, "DMF (dimethylformamide)", which are generated based on the fifth extended node 3315, "aprotic polar solvent". The extended graph 335 also shows the tenth edge 3330, which is an arrow pointing from the seventh extended node 3317 to the fifth extended node 3315, indicating the relationship between the fifth extended node 3315 and the seventh extended node 3317, and the eleventh edge 3331, which is an arrow pointing from the eighth extended node 3318 to the fifth extended node 3315, indicating the relationship between the fifth extended node 3315 and the eighth extended node 3318.

[0133] This extended graph 335 shows that to achieve an 80% yield in the synthesis of ethers by SN2 reactions, the conditions should be considered using aprotic polar solvents such as "MeCN" or "DMF." This will assist in the knowledge (technical) education of new engineers and in the search for unknown knowledge data.

[0134] Then, the engineer selects the "Export Results" button 338 to output the results shown on the extended graph display screen 330. Furthermore, if you wish to ask another question, selecting the "Start a new analysis" button 339 will display the technical information input screen 300 shown in Figure 12.

[0135] Furthermore, in this system 1, the selection of nodes to be expanded can be left to the discretion of the engineer, rather than automatically selecting nodes representing more abstract concepts that the system 1 can explore in depth by selecting buttons as described above. In other words, instead of selecting the "Show Next Node" button 336 on the expanded graph display screen 330 shown in Figures 15-1 to 15-4, the system can be expanded sequentially by the engineer "selecting" (clicking or tapping) any node, for example, as shown in Figure 16. Figure 16 is a schematic diagram of the extended graph display screen 340, which shows an example illustrating extended nodes and extended edges generated by node selection in the knowledge graph displayed on the engineer terminal 30.

[0136] In Figure 16, the extended graph display screen 340 shows an extended graph 341 in which engineers can select nodes. This extended graph 341 can be expanded by generating extended nodes and extending the knowledge graph, not by selecting the "Extend" button 314 on the knowledge graph display screen 310 shown in Figure 13, but by selecting (clicking or tapping) nodes by the engineer, as indicated by the white arrows superimposed in the figure.

[0137] The extended graph 341 shows that when the first node 3121, labeled "ether synthesis" and shown in a different color from the other nodes in the figure, is selected as the selected node, the first extended node 3411, labeled "SN2 reaction," is generated based on the first node 3121, and the first edge 3421, which indicates the relationship between the first node 3121 and the first extended node 3411, is generated as an arrow pointing from the first node 3121 to the first extended node 3411.

[0138] Although subsequent expansions are not shown in the diagram, engineers can continue to select nodes to expand according to their own will and sequentially expand the expansion graph 341, thereby freely exploring their thoughts rather than being limited to standardized thinking.

[0139] Furthermore, in this system 1, instead of automatically generating, selecting, and displaying the extension nodes generated by selecting the extension node, it is also possible to allow the selection to be left to the discretion of the engineer. That is, in the extension graph display screen 340 shown in Figure 16, when a node is selected (clicked or tapped), for example, as shown in Figure 17-1, multiple extension node candidates can be generated and displayed, and as shown in Figure 17-2, one extension node candidate can be selected as the extension node. Figure 17-1 is a schematic diagram of the extended graph display screen 350, which shows an example illustrating multiple extended node candidates and extended edge candidates generated by node selection in the knowledge graph displayed on the engineer terminal 30. Figure 17-2 is a schematic diagram showing an example configuration of the extended graph display screen 350, illustrating the state in which one extended node candidate has been selected in the knowledge graph shown in Figure 17-1.

[0140] In Figure 17-1, the extended graph display screen 350 shows an extended graph 351 in which an engineer can select an extended node. As shown by the white arrows in the diagram, this extended graph 351 generates multiple candidate extended nodes when a node is selected (clicked or tapped) by an engineer, and one of these candidate extended nodes can be selected as the extended node.

[0141] The extended graph 351 shows that when the first node 3121, labeled "ether synthesis" and shown in a different color from the other nodes in the figure, is selected as the selected node, the following extension node candidates are displayed as shown by the dotted lines in the figure: the first extension node candidate 3511, labeled "alkylating agent," the second extension node candidate 3512, labeled "SN2 reaction," and the third extension node candidate 3513, labeled "side reaction," all of which are generated based on the first node 3121. Although Figure 17-1 shows three candidate extension nodes, the number is not limited to three; it could be two or three or more, and can be selected arbitrarily.

[0142] Furthermore, in Figure 17-2, the extended graph display screen 350 shows the extended graph 352 after the extended node has been selected by the engineer. In this extended graph 352, the second extended node candidate 3512, labeled "SN2 reaction," is shown by an overlaid white arrow and in a different color from the other nodes in the figure. When this second extended node candidate 3512 is selected by the engineer, the remaining first extended node candidate 3511 and third extended node candidate 3513 disappear, and the selected second extended node candidate 3512 is displayed as the first extended node.

[0143] Although subsequent extensions are not shown in the diagram, engineers can continue to select extension nodes of their own volition and sequentially extend the extension graph 341, thereby freely exploring their thoughts rather than being limited to standardized thinking.

[0144] Furthermore, in this system 1, along with the display of multiple extension node candidates, it is possible to display explanations of the nodes to assist in selecting an extension node. That is, in the extension graph display screen 350 shown in Figure 17-1, by selecting an extension node candidate (mouse over), an explanation of the node can be displayed, for example, as shown in Figure 18. Figure 18 is a schematic diagram showing an example configuration of the extended graph display screen 360, which explains the technical explanation displayed when one extended node candidate is selected in the knowledge graph shown in Figure 17-1.

[0145] In Figure 18, the extended graph display screen 360 shows an extended graph 361 displaying an explanation of the second extended node candidate 3512 selected in Figure 17-2. As shown by the white arrows superimposed in the figure, when a node is selected (mouse over) by an engineer, an explanatory display section 362 will appear near the selected second candidate extended node 3512, for example, "The SN2 reaction is one of the common reaction mechanisms in organic chemistry, producing an ether from an alkyl halide and an alkoxide ion..."

[0146] Therefore, before selecting an extension node (click or tap), the engineer can view descriptions of each of the multiple extension node candidates displayed by hovering the mouse pointer over them. Furthermore, moving the mouse pointer away from a candidate extension node (deselecting it) will hide the display of the technical explanation.

[0147] Furthermore, in this system 1, instead of displaying multiple extension node candidates at once, they can be displayed sequentially. That is, in the extension graph display screen 340 shown in Figure 16, when a node is selected (clicked or tapped), different extension node candidates can be generated and displayed, for example, as shown in Figures 19-1 to 19-3, and one extension node candidate can be selected as the extension node. Figures 19-1 to 19-3 are schematic diagrams illustrating the configuration of an extended graph display screen 370, which explains the switching of extended node candidates generated by node selection in the knowledge graph displayed on the engineer terminal 30.

[0148] In Figure 19-1, the extended graph display screen 370 shows an extended graph 371 in which engineers can select nodes. This extended graph 371, as shown by the white arrows superimposed in the knowledge graph display screen 310 in Figure 13, allows engineers to select (click or tap) nodes, which in turn allows them to sequentially switch and display candidates for extended nodes.

[0149] In the extended graph 371, the first node 3121, labeled "Ether Synthesis," which is shown in a different color from the other nodes in the figure, is selected (clicked or tapped) as a selected node, and the first extended node candidate 3711, labeled "Alkylating Agent," which was generated based on the first node 3121, is displayed.

[0150] Furthermore, by selecting (clicking or tapping) the first node 3121 in this state, it is possible to display, for example, the extended graph 372 shown in Figure 19-2. In Figure 19-2, the extended graph 372 is shown displaying a second extended node candidate 3712, labeled "SN2 reaction," which was generated based on the first node 3121. Furthermore, by selecting (clicking or tapping) the first node 3121 in this state, it is possible to display, for example, the extended graph 373 shown in Figure 19-3. In Figure 19-3, the extended graph 372 is shown displaying a third candidate extended node 3713, labeled "Side Reaction," which was generated based on the first node 3121.

[0151] Although the subsequent switching of extension node candidates is not illustrated, the user can continue to select (click or tap) the first node 3121 in the same manner to sequentially switch and display the extension node candidates. Therefore, it is possible to display more extension node candidates compared to displaying them all at once, allowing the user to select the appropriate extension node at their own discretion.

[0152] Furthermore, in this system 1, it is also possible to change the extension node based on the edge. That is, when a node is selected (clicked or tapped), the orientation of the edge is reversed, and the displayed extension node is changed to another extension node in relation to the orientation of the edge, for example, as shown in Figure 20-1. Figure 20-1 is a schematic diagram showing an example configuration of the extended graph display screen 380, which illustrates the switching of extended node candidates generated by edge selection in the extended graph displayed on the engineer terminal 30.

[0153] In Figure 20-1, the extended graph display screen 380 shows an extended graph 381 in which engineers can select edges. In this extended graph 381, when the first node 3121, labeled "ether synthesis," is selected as the selected node in the extended graph display screen 340 shown in Figure 16, and the first extended node 3411, labeled "SN2 reaction," is generated based on the first node 3121, selecting (clicking or tapping) the first edge 3421, which shows the relationship between the first node 3121 and the first extended node 3411, reverses the direction of the edge, and based on the relationship determined by the direction of the edge, another extended node 3811, labeled "dehydration reaction," is generated, indicating that ether is synthesized by a dehydration reaction.

[0154] By changing the extension nodes based on edges in this way, engineers can select the nodes to extend according to their own will and sequentially extend the extension graph 341, thereby exploring their thoughts freely rather than being limited to standardized thinking.

[0155] Figure 20-2 is a schematic diagram showing an example configuration of the extended graph display screen 380, which explains the switching of extended node candidates generated by operations on edges in the extended graph displayed on the engineer terminal 30.

[0156] As shown in Figure 20-2, System 1 displays a new edge connected to the first node 3121 in the knowledge graph when the first node 3121 is selected as the selected node. Then, System 1 displays an extended node in response to an operation on the new edge. Here, new edges are displayed extending in a predetermined direction from the selected node. The generated edges contain information about the predetermined relationships between nodes. Furthermore, the relationships of the generated edges are determined based on the technical elements represented by the selected node. Furthermore, the new extension node includes the technological elements represented by the selected node, the relationships indicated by the new edges, and new technological elements generated based on the arrangement of the new extension node with other nodes.

[0157] As shown in Figure 20-2(A-1), System 1 selects the first node 3121 in the graph as the selected node. Then, as shown in Figure 20-2(A-2), System 1 displays multiple edges (E101, E102, E103) connected to the first node 3121 as the selected node. The system then accepts a selection operation from the user for one of the multiple edges. In the example shown in the figure, edge E103 is selected. Edge E103 is associated with a predetermined concept indicating the relationship between nodes. In addition, the relationship between the nodes to which each edge is connected can be displayed as text. Then, in response to the selection of edge E103, system 1 generates a new extension node based on the technical elements indicated by the first node 3121 and the relationships between nodes defined by edge E103.

[0158] In this way, depending on the selected node, multiple new edges connected to the selected node can be displayed, and depending on the selection of one of these edges, an extended node can be displayed. This makes it possible to generate a new extended node using the edge selected by the user from among multiple edges that define the relationships between nodes.

[0159] As shown in Figure 20-2(B-1), the first node 3121 in the graph is selected as the selected node. Then, System 1 displays the edge E105 connected to the first node 3121, which is the selected node. Edge E105 is associated with a predetermined concept representing the relationship between nodes.

[0160] As shown in Figure 20-2(B-2), System 1 accepts user input to move the end of edge E105 in the graph. That is, while one end of edge E105 is connected to the selected node, it accepts input to move the position of the other end. Edge E105 changes in length, angle, etc., in response to the user's operation to move its end.

[0161] Subsequently, when the user's movement operation on edge E105 is completed, system 1 generates a new extended node based on the technical elements indicated by the first node 3121, the relationships between nodes defined by edge E105, and the relationships with other nodes adjacent to edge E105 (in the example shown, the node indicating "solvent amount").

[0162] In this way, the system accepts move operations on new edges and can display extended nodes according to the position of the new edges in the graph resulting from the move operation. This makes it possible for users to generate new extended nodes by moving the edges that define the relationships between nodes.

[0163] Furthermore, System 1 can assist engineers in their thinking process by searching for technically relevant unknown knowledge data based on the technical information entered as a question. In other words, if a technical development project fails to achieve the desired results and development is halted, an experienced technical developer might be able to provide effective advice based on past technical information. However, given the current shortage of researchers and engineers, such advice cannot be expected in the future. Therefore, System 1 is used to search for unknown knowledge data that is thought to be technically relevant to the current technical information, with the expectation that it will help find past technical information that can provide effective advice for current technical development.

[0164] For the sake of explanation, we will use the knowledge graphs of technical information regarding the fungicide trifmine, shown in Figures 21(A) and 21(B). Figure 21 shows (A) a schematic diagram illustrating the graph of the first technical information and (B) a schematic diagram illustrating the graph of the second technical information, which are used to explain the generation of the extended graph.

[0165] Furthermore, as shown in Figures 22-1 to 22-5, it is possible to search one knowledge graph from another, thereby assisting in thinking by connecting knowledge with different purposes. Figures 22-1 to 22-5 are schematic diagrams showing examples of the configuration of an extended graph display screen 390 that sequentially explains the procedure for displaying the extended graph shown on the engineer terminal 30 during the search for unknown knowledge data.

[0166] Figure 21(A) shows a knowledge graph of the first technical information that studied the covalent crystal of "trifmine," and is shown as being composed of the first node A1, "trifmine," the second node A2, "hydrogen bonding," the third node A3, "hydroquinone," the fourth node A4, "cocrystal," the fifth node A5, "acidic conformer," and the sixth node A6, "residual rate." This first piece of technical information is stored in the technical information DB210, while the knowledge graph shown in Figure 21(A) is stored in the structured information DB220.

[0167] Furthermore, Figure 21(B) shows a knowledge graph of the second piece of technical information that studied the interaction of "trifmine," and is shown as being composed of the first node B1, "trifmine," the second node B2, "ziram," the third node B3, "agricultural fungicide," the fourth node B4, "interaction," and the fifth node B5, "caking." This second piece of technical information is also stored in the technical information DB210, while the knowledge graph shown in Figure 21(B) is stored in the structured information DB220.

[0168] Below, based on Figures 22-1 to 22-5, we will explain how to search the knowledge graph of the second technical information as unknown knowledge data from the knowledge graph of the first technical information, with current technical information designated as the first technical information and past technical information designated as the second technical information. In Figure 22-1, the extended graph display screen 390 shows the knowledge graph 391 of the first technical information. In knowledge graph 391, as indicated by the overlapping white arrows and the different color used for other nodes, when the fourth node A4, "cocrystal," is selected as the selected node, the extended graph 392 shown in Figure 22-2 is displayed.

[0169] In the extended graph 392 shown in Figure 22-2, a first extended node 3921 called "complex cocrystal" is generated based on the fourth node A4, which is "cocrystal," and a first extended edge 3922 is generated as an arrow pointing from the fourth node A4 to the first extended node 3921. In conjunction with the generation of this extended graph 392, the question posed to System 1 is to check for the existence of other technical information knowledge graphs that have nodes that are technically related to or match the generated extended node "complex cocrystal".

[0170] If there are no other technical information knowledge graphs that can be connected to extended graph 392, then the first extended node 3921, "complex cocrystal," is selected as the selected node, for example, by overlapping it with a white arrow in the figure and indicating it with a different color from other nodes. This will result in the extended graph 393 shown in Figure 22-3.

[0171] In the extended graph 393 shown in Figure 22-3, a second extended node 3931 called "coordinate bond" is generated based on the first extended node 3921, "complex cocrystal," and a second extended edge 3932 is generated as an arrow pointing from the first extended node 3921 to the second extended node 3931. With the generation of this extended graph 393, the question posed to System 1 is to check for the existence of other knowledge graphs of technical information that have nodes that are technically related to or match the newly generated extended node called "coordinate coherence".

[0172] If there are no other technical information knowledge graphs that can be connected to extended graph 393, and the second extended node 3931, "Coordination Coupling," is selected as the selected node, as indicated by the white arrow in the figure and the different color from the other nodes, then extended graph 394 shown in Figure 22-4 will be displayed.

[0173] In the extended graph 394 shown in Figure 22-4, a third extended node 3941 called "interaction" is generated based on the second extended node 3931, which is "coordinate bond," and a third extended edge 3942 is generated as an arrow pointing from the second extended node 3931 to the third extended node 3941. With the generation of this extended graph 394, the question posed to System 1 is to check for the existence of other knowledge graphs of technical information that have nodes that are technically related to or match the newly generated extended node called "Interaction".

[0174] As a result, if there is a second piece of technical information, shown in Figure 21(B), that can be connected to the extended graph 394 based on the newly generated extended node called "interaction," then the extended graph 395 shown in Figure 22-5 will be displayed.

[0175] In the extended graph 395 shown in Figure 22-5, the third extended node 3941 and the fourth node B4 in the knowledge graph of the second technical information shown in Figure 21(B) are displayed in such a way that they overlap at least partially.

[0176] Furthermore, System 1 can also perform the reverse of the search described above. For example, if there is past technical information where desirable results were obtained but the reason (basis) was unknown and further technological development did not progress, the system can use current knowledge (technical information) to search for unknown knowledge data that is thought to be technically related to the past technical information, with the expectation that the reason why the desirable results were obtained will be revealed and contribute to further technological development.

[0177] Here, for the sake of explanation, we will use the knowledge graphs shown in Figures 21(A) and 21(B). Furthermore, as shown in Figures 23-1 to 23-5, it is possible to search one knowledge graph from the other, thereby assisting in thinking by connecting knowledge with different purposes. Figures 23-1 to 23-5 are schematic diagrams showing examples of the configuration of the extended graph display screen 400, which sequentially explains other display procedures for the extended graph shown on the engineer terminal 30 during the search for unknown knowledge data.

[0178] The following explanation, based on Figures 23-1 to 23-5, describes how to search for the knowledge graph of the first technical information (current technical information) as unknown knowledge data from the knowledge graph of the second technical information (past technical information). In Figure 23-1, the extended graph display screen 400 shows the knowledge graph 401 of the second technical information. In knowledge graph 401, as indicated by the overlapping white arrows in the figure and the different color used for other nodes, when the fourth node B4, "Interaction," is selected as the selected node, the extended graph 402 shown in Figure 23-2 is displayed.

[0179] In the extended graph 402 shown in Figure 23-2, a first extended node 4021 called "coordinate bond" is generated based on the third node B3, "agricultural fungicide," and the fourth node B4, "interaction." It is also shown that a first extended edge 4022 is generated as an arrow from the third node B3 to the first extended node 4021, and a second extended edge 4023 is generated as an arrow from the fourth node B4 to the first extended node 4021. In conjunction with the generation of this extended graph 402, the question posed to System 1 is to check for the existence of other knowledge graphs of technical information that have nodes that are technically related to or match the generated extended node called "coordinate coherence".

[0180] If there are no other technical information knowledge graphs that can be connected to extended graph 402, then the first extended node 4021, "coordinate bond," is selected as the selected node, for example, by overlaying a white arrow in the figure and indicating it with a different color from other nodes. This will result in the extended graph 403 shown in Figure 23-3.

[0181] In the extended graph 403 shown in Figure 23-3, a second extended node 4031, labeled "complex cocrystal," is generated based on the third node B3, "agricultural fungicide," and the first extended node 4021, "coordinate bond." Additionally, a third extended edge 4032 is generated as an arrow pointing from the third node B3 to the second extended node 4031, and a fourth extended edge 4033 is generated as an arrow pointing from the first extended node 4021 to the second extended node 4031. With the generation of this extended graph 403, the question posed to System 1 is to check for the existence of other technical information knowledge graphs that have nodes that are technically related to or match the newly generated extended node "complex cocrystal".

[0182] If there are no other technical information knowledge graphs that can be connected to extended graph 403, and the second extended node 4031, "complex cocrystal," is selected as the selected node, as indicated by the white arrow in the figure and the different color from the other nodes, then extended graph 404 shown in Figure 23-4 will be displayed.

[0183] In the extended graph 404 shown in Figure 23-4, a third extended node 4041, named "cocrystal," is generated based on the second extended node 4031, "complex cocrystal," and a fifth extended edge 4042 is generated as an arrow pointing from the second extended node 4031 to the third extended node 4041. With the generation of this extended graph 404, the question posed to System 1 is to check for the existence of other technical information knowledge graphs that have nodes that are technically related to or match the newly generated extended node called "cocrystal".

[0184] As a result, if there is the first piece of technical information shown in Figure 21(A) that can be connected to the extended graph 404 based on the newly generated extended node called "cocrystal," then the extended graph 405 shown in Figure 23-5 will be displayed.

[0185] In the extended graph 405 shown in Figure 23-5, the third extended node 4041 and the fourth node A4 in the knowledge graph of the first technical information shown in Figure 21(A) are displayed in such a way that they overlap at least partially.

[0186] Furthermore, System 1 can assist engineers in connecting two pieces of knowledge that seem technically related but whose basis is unclear. In other words, for the sake of explanation, using the knowledge graphs shown in Figures 21(A) and 21(B), as shown in Figures 24-1 to 24-4, one of the knowledge graphs can be extended to assist in connecting pieces of knowledge with different purposes.

[0187] Figures 24-1 to 24-4 are schematic diagrams showing examples of the configuration of the extended graph display screen 410, which sequentially explains the procedure for displaying the extended graph shown on the engineer terminal 30 when linking two pieces of knowledge.

[0188] In the technical information input screen 300 shown in Figure 12, select the first technical information shown in Figure 21(A) and the second technical information shown in Figure 21(B) from the pull-down menu button 3011 in the research objective input field 301. Then, when the "Next: Build Basic Graph" button 304 is selected, the extended graph display screen 410, as shown in Figure 24-1, will be displayed.

[0189] In Figure 24-1, the extended graph display screen 410 shows a knowledge graph 411 which includes the knowledge graph for the first technical information and the knowledge graph for the second technical information, respectively. In the knowledge graph 411, for example, if the fourth node A4 in the knowledge graph of the first technical information, "cocrystal," is selected as the selection node, as indicated by the overlapping white arrows in the figure and the different color used for other nodes, the extended graph 412 shown in Figure 24-2 will be displayed.

[0190] In the extended graph 412 shown in Figure 24-2, a first extended node 4121 called "complex cocrystal" is generated based on the fourth node A4, which is "cocrystal," and a first extended edge 4122 is generated as an arrow pointing from the fourth node A4 to the first extended node 4121. Next, the first extended node 4121, "complex cocrystal," is selected as the selected node, for example, by overlapping it with a white arrow in the figure and indicating it with a different color from the other nodes. This results in the extended graph 413 shown in Figure 24-3.

[0191] In the extended graph 413 shown in Figure 24-3, a second extended node 4131 called "coordinate bond" is generated based on the first extended node 4121, "complex cocrystal," and a second extended edge 4132 is generated as an arrow pointing from the first extended node 4121 to the second extended node 4131. Furthermore, as indicated by the overlapping white arrows in the diagram and the different color used for the other nodes, when the second extended node 4131, "coordinate bond," is selected as the selected node, the extended graph 414 shown in Figure 24-4 is displayed.

[0192] In the extended graph 414 shown in Figure 24-4, it is shown that the second extended node 4131, "coordinate coherence," is coupled with the fourth node B4, "interaction," in the knowledge graph of the second technical information.

[0193] This allows engineers to recognize that the "cocrystal" in the first technical information is the "interaction" in the second technical information, and to use the second technical information for further examination of the first technical information.

[0194] [Note] (Note 1) It is a system that assists the user's thinking process. Having one or more processors, The aforementioned processor, We accept technical information, which is information related to technology. The aforementioned technical information is displayed as a graph consisting of multiple nodes and multiple edges. In the graph, when one of the multiple nodes is selected as a selected node, an extension node, which is a new node technically related to the selected node, and an extension edge, which is a new edge indicating the relationship between the selected node and the extension node, are generated. Display the extended graph, which includes the generated extended nodes and extended edges added to the graph. An auxiliary system characterized by the following features. (Note 2) The auxiliary system according to Appendix 1, characterized in that it generates the extension node and the extension edge based on the technical elements represented by other nodes that are located close to the selected node in the graph. (Note 3) In response to the selection of the extension node in the extension graph as the selected node, a new extension node and a new extension edge are generated. The auxiliary system according to Appendix 1, characterized by displaying a new extended graph in which the newly generated extended nodes and extended edges have been added to the extended graph. (Note 4) The auxiliary system according to Appendix 1, characterized in that, among the multiple nodes in the graph, a node containing a more abstract concept is selected as the selected node. (Note 5) The auxiliary system according to Appendix 1, characterized in that it accepts a selection from the user to designate one of the multiple nodes in the graph as the selected node. (Note 6) The auxiliary system according to Appendix 5, characterized in that, in response to the selection of the selected node, it displays a plurality of new nodes that are technically related to the selected node as candidates for extended nodes. (Note 7) The auxiliary system according to Appendix 6, characterized in that, when one of the multiple extension node candidates is selected as the extension node, a technical explanation of the selected extension node candidate is displayed. (Note 8) The auxiliary system according to Appendix 6, characterized in that, when one of the multiple extension node candidates is selected as the extension node, the other extension node candidates that are not selected are hidden. (Note 9) Multiple nodes that are technically related to the selected node are generated as candidates for extended nodes. The auxiliary system according to Appendix 5, characterized in that it displays a plurality of generated extension node candidates one by one in order according to the selection of the selected node. (Note 10) Depending on the selection of the selected node in the graph, a new edge connected to the selected node is displayed. The auxiliary system according to Appendix 1, characterized in that it displays the extended node in response to an operation on the new edge. (Note 11) Depending on the selection of the aforementioned node, multiple new edges connected to the aforementioned node are displayed. The auxiliary system according to Appendix 1, characterized in that it displays the extended node in response to the selection operation of one of the plurality of edges. (Note 12) Accepting a move operation on the aforementioned new edge, The auxiliary system according to Appendix 1, characterized in that it displays the extended node according to the position of the new edge in the graph resulting from the movement operation. (Note 13) Each time the extended graph is displayed, the system searches for other graphs representing other technical information that have nodes that are technically related to or match the newly generated extended node. If there is another graph representing the other technical information, the graph of the other technical information and the extended graph are displayed such that at least a portion of the technically related or matching nodes and the extended nodes of the extended graph overlap. The auxiliary system according to Appendix 1, characterized in that, if there is no other graph representing the other technical information, a new extended graph is displayed which further generates the extended nodes and extended edges, and another graph has nodes that are technically related to or match the newly generated extended nodes. (Note 14) The system receives multiple pieces of the aforementioned technical information, generates and displays the respective graphs, Each time a new extended graph is displayed in one graph, which has further generated the extended nodes and extended edges, the presence or absence of nodes that are technically related to or match the newly generated extended nodes is identified in other graphs. If there are technically related or matching nodes in the other graph, the first graph and the other graph are displayed such that at least a portion of the nodes and the extension nodes overlap. The auxiliary system according to Appendix 1, characterized in that, if there are no technically related or matching nodes in the other graph, a new extended graph is displayed in the first graph, further generating the extended nodes and extended edges, and the presence or absence of nodes that are technically related or matching to the newly generated extended nodes is again identified in the other graph. (Note 15) In the graph, the user selects the selected node from among the multiple nodes. The system generates several new extension node candidates, which are technically related to the selected node, and extension edges, which are new edges indicating the relationship between the selected node and the extension node, depending on the selection of one of the extension node candidates. The extended graph, including the extended nodes and extended edges, is displayed. The auxiliary system according to Appendix 1, characterized in that for each selection of the extension node, a new extension graph is displayed that includes the extension node and the extension edge. (Note 16) It is a device that assists the user's thinking process. Having one or more processors, The aforementioned processor, We accept technical information, which is information related to technology. The aforementioned technical information is displayed as a graph consisting of multiple nodes and multiple edges. In the graph, when one of the multiple nodes is selected as a selected node, an extension node, which is a new node technically related to the selected node, and an extension edge, which is a new edge indicating the relationship between the selected node and the extension node, are generated. The extended graph, which includes the generated extended nodes and extended edges added to the graph, is displayed on the screen. An auxiliary device characterized by the following features. (Note 17) On the computer, A function to receive technical information, which is information related to technology, A function to display a graph representing the aforementioned technical information using multiple nodes and multiple edges, In the graph, when one of the multiple nodes is selected as a selected node, a function is provided to generate an extension node, which is a new node that is technically related to the selected node, and an extension edge, which is a new edge that shows the relationship between the selected node and the extension node. A function to display an extended graph in which the generated extended nodes and extended edges are added to the graph, A program that makes this possible.

[0195] Although the embodiments described above are illustrative examples intended to facilitate understanding of the present invention, they are not intended to limit the interpretation of the present invention. Furthermore, the present invention is not limited to the embodiments described above, and any modifications, improvements, etc., that can achieve the objectives of the present invention are included in the present invention. [Explanation of symbols]

[0196] 10...Thinking assistance device (server), 30...Technician terminal, 110...Reception unit, 120...Generation unit, 130...Display unit, 140...Selection unit, 150...Confirmation unit, 210...Technical information DB, 220...Structured information DB, 230...Unstructured information DB

Claims

1. It is a system that assists the user's thinking process. Having one or more processors, The aforementioned processor, We accept technical information, which is information related to technology. The aforementioned technical information is displayed as a graph consisting of multiple nodes and multiple edges. In the graph, when one of the multiple nodes is selected as a selected node, an extension node, which is a new node technically related to the selected node, and an extension edge, which is a new edge indicating the relationship between the selected node and the extension node, are generated. The extended graph is displayed, with the generated extended nodes and extended edges added to the graph. In the graph, the extension nodes and extension edges are generated based on the technical elements represented by other nodes that are located close to the selected node. An auxiliary system characterized by the following features.

2. It is a system that assists the user's thinking process. Having one or more processors, The aforementioned processor, We accept technical information, which is information related to technology. The aforementioned technical information is displayed as a graph consisting of multiple nodes and multiple edges. In the graph, when one of the multiple nodes is selected as a selected node, an extension node, which is a new node technically related to the selected node, and an extension edge, which is a new edge indicating the relationship between the selected node and the extension node, are generated. The extended graph is displayed, with the generated extended nodes and extended edges added to the graph. Among the multiple nodes in the graph, the node containing a more abstract concept is selected as the selected node. An auxiliary system characterized by the following features.

3. It is a system that assists the user's thinking process. Having one or more processors, The aforementioned processor, We accept technical information, which is information related to technology. The aforementioned technical information is displayed as a graph consisting of multiple nodes and multiple edges. In the graph, when one of the multiple nodes is selected as a selected node, an extension node, which is a new node technically related to the selected node, and an extension edge, which is a new edge indicating the relationship between the selected node and the extension node, are generated. The extended graph is displayed, with the generated extended nodes and extended edges added to the graph. The system accepts a selection from the user to designate one of the multiple nodes in the graph as the selected node. An auxiliary system characterized by the following features.

4. Depending on the selection of the aforementioned node, a number of new nodes technically related to the aforementioned node are displayed as extension node candidates. The auxiliary system according to claim 3, characterized in that way.

5. When one of the multiple extension node candidates is selected as the extension node, a technical explanation of the selected extension node candidate is displayed. The auxiliary system according to feature 4.

6. When one of the multiple extension node candidates is selected as the extension node, the other extension node candidates that are not selected are hidden. The auxiliary system according to feature 4.

7. Multiple nodes that are technically related to the selected node are generated as candidates for extended nodes. Depending on the selection of the aforementioned node, the generated multiple candidate extension nodes are displayed one by one in sequence. The auxiliary system according to claim 3, characterized in that way.

8. It is a system that assists the user's thinking process. Having one or more processors, The aforementioned processor, We accept technical information, which is information related to technology. The aforementioned technical information is displayed as a graph consisting of multiple nodes and multiple edges. In the graph, when one of the multiple nodes is selected as a selected node, an extension node, which is a new node technically related to the selected node, and an extension edge, which is a new edge indicating the relationship between the selected node and the extension node, are generated. The extended graph is displayed, with the generated extended nodes and extended edges added to the graph. Depending on the selection of the selected node in the graph, a new edge connected to the selected node is displayed. In response to the operation on the new edge, the extended node is displayed. An auxiliary system characterized by the following features.

9. It is a system that assists the user's thinking process. Having one or more processors, The aforementioned processor, We accept technical information, which is information related to technology. The aforementioned technical information is displayed as a graph consisting of multiple nodes and multiple edges. In the graph, when one of the multiple nodes is selected as a selected node, an extension node, which is a new node technically related to the selected node, and an extension edge, which is a new edge indicating the relationship between the selected node and the extension node, are generated. The extended graph is displayed, with the generated extended nodes and extended edges added to the graph. Depending on the selection of the aforementioned node, multiple new edges connected to the aforementioned node are displayed. An auxiliary system characterized by displaying the extended node in response to a selection operation of one of the aforementioned multiple edges.

10. It is a system that assists the user's thinking process. Having one or more processors, The aforementioned processor, We accept technical information, which is information related to technology. The aforementioned technical information is displayed as a graph consisting of multiple nodes and multiple edges. In the graph, when one of the multiple nodes is selected as a selected node, an extension node, which is a new node technically related to the selected node, and an extension edge, which is a new edge indicating the relationship between the selected node and the extension node, are generated. The extended graph is displayed, with the generated extended nodes and extended edges added to the graph. Accepting a move operation on the aforementioned new edge, The extended node is displayed according to the position of the new edge in the graph resulting from the movement operation. An auxiliary system characterized by the following features.

11. It is a system that assists the user's thinking process. Having one or more processors, The aforementioned processor, We accept technical information, which is information related to technology. The aforementioned technical information is displayed as a graph consisting of multiple nodes and multiple edges. In the graph, when one of the multiple nodes is selected as a selected node, an extension node, which is a new node technically related to the selected node, and an extension edge, which is a new edge indicating the relationship between the selected node and the extension node, are generated. The extended graph is displayed, with the generated extended nodes and extended edges added to the graph. Each time the extended graph is displayed, the system searches for other graphs representing other technical information that have nodes that are technically related to or match the newly generated extended node. If there is another graph representing the other technical information, the graph of the other technical information and the extended graph are displayed such that at least a portion of the technically related or matching nodes and the extended nodes of the extended graph overlap. If there are no other graphs representing the aforementioned other technical information, a new extended graph is displayed, further generating the extended nodes and extended edges, and another graph is searched again for nodes that are technically related to or match the newly generated extended nodes. An auxiliary system characterized by the following features.

12. It is a system that assists the user's thinking process. Having one or more processors, The aforementioned processor, We accept technical information, which is information related to technology. The aforementioned technical information is displayed as a graph consisting of multiple nodes and multiple edges. In the graph, when one of the multiple nodes is selected as a selected node, an extension node, which is a new node technically related to the selected node, and an extension edge, which is a new edge indicating the relationship between the selected node and the extension node, are generated. The extended graph is displayed, with the generated extended nodes and extended edges added to the graph. The system receives multiple pieces of the aforementioned technical information, generates and displays the respective graphs, Each time a new extended graph is displayed in one graph, which has further generated the extended nodes and extended edges, the presence or absence of nodes that are technically related to or match the newly generated extended nodes is identified in other graphs. If there are technically related or matching nodes in the other graph, the first graph and the other graph are displayed such that at least a portion of the nodes and the extension nodes overlap. If there are no technically related or matching nodes in the other graph, a new extended graph is displayed in the first graph, further generating the extended nodes and extended edges, and the presence or absence of technically related or matching nodes in the other graph is identified again. An auxiliary system characterized by the following features.

13. It is a system that assists the user's thinking process. Having one or more processors, The aforementioned processor, We accept technical information, which is information related to technology. The aforementioned technical information is displayed as a graph consisting of multiple nodes and multiple edges. In the graph, when one of the multiple nodes is selected as a selected node, an extension node, which is a new node technically related to the selected node, and an extension edge, which is a new edge indicating the relationship between the selected node and the extension node, are generated. The extended graph is displayed, with the generated extended nodes and extended edges added to the graph. In the graph, the user selects the selected node from among the multiple nodes. The system generates several new extension node candidates, which are technically related to the selected node, and extension edges, which are new edges indicating the relationship between the selected node and the extension node, depending on the selection of one of the extension node candidates. The extended graph, including the extended nodes and extended edges, is displayed. For each selection of the extension node, a new extension graph is displayed, which includes the extension node and the extension edge. An auxiliary system characterized by the following features.

14. It is a device that assists the user's thinking process. Having one or more processors, The aforementioned processor, We accept technical information, which is information related to technology. The aforementioned technical information is displayed as a graph consisting of multiple nodes and multiple edges. In the graph, when one of the multiple nodes is selected as a selected node, an extension node, which is a new node technically related to the selected node, and an extension edge, which is a new edge indicating the relationship between the selected node and the extension node, are generated. The extended graph, which includes the generated extended nodes and extended edges added to the graph, is displayed on the screen. In the graph, the extension nodes and extension edges are generated based on the technical elements represented by other nodes that are located close to the selected node. An auxiliary device characterized by the following features.

15. On the computer, A function to receive technical information, which is information related to technology, A function to display a graph representing the aforementioned technical information using multiple nodes and multiple edges, In the graph, when one of the multiple nodes is selected as a selected node, a function is provided to generate an extension node, which is a new node that is technically related to the selected node, and an extension edge, which is a new edge that shows the relationship between the selected node and the extension node. A function to display an extended graph in which the generated extended nodes and extended edges are added to the graph, The function generates the extension node and the extension edge based on the technical elements represented by other nodes that are located close to the selected node in the graph, A program that makes this possible.

Images