WO2022145100A1 - Information processing system, information processing method, and program - Google Patents

Information processing system, information processing method, and program Download PDF

Info

Publication number
WO2022145100A1
WO2022145100A1 PCT/JP2021/037310 JP2021037310W WO2022145100A1 WO 2022145100 A1 WO2022145100 A1 WO 2022145100A1 JP 2021037310 W JP2021037310 W JP 2021037310W WO 2022145100 A1 WO2022145100 A1 WO 2022145100A1
Authority
WO
WIPO (PCT)
Prior art keywords
node
information processing
processing system
information
input
Prior art date
Application number
PCT/JP2021/037310
Other languages
French (fr)
Japanese (ja)
Inventor
潤 松▲崎▼
Original Assignee
潤 松▲崎▼
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 潤 松▲崎▼ filed Critical 潤 松▲崎▼
Publication of WO2022145100A1 publication Critical patent/WO2022145100A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
    • G06F3/048—Interaction techniques based on graphical user interfaces [GUI]
    • G06F3/0481—Interaction techniques based on graphical user interfaces [GUI] based on specific properties of the displayed interaction object or a metaphor-based environment, e.g. interaction with desktop elements like windows or icons, or assisted by a cursor's changing behaviour or appearance
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
    • G06F3/048—Interaction techniques based on graphical user interfaces [GUI]
    • G06F3/0484—Interaction techniques based on graphical user interfaces [GUI] for the control of specific functions or operations, e.g. selecting or manipulating an object, an image or a displayed text element, setting a parameter value or selecting a range
    • G06F3/04842—Selection of displayed objects or displayed text elements
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F8/00—Arrangements for software engineering
    • G06F8/30—Creation or generation of source code
    • G06F8/34—Graphical or visual programming
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/04—Inference or reasoning models
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"

Definitions

  • the present invention relates to an information processing system, an information processing method and a program.
  • Inference model Statistical inference from data and models by incorporating domain knowledge about data generation into models in predicting various events such as transmission of infectious diseases, shipments of goods, consumer trends, prices of financial products, etc. Inference model is used.
  • Non-Patent Document 1 discloses a programming language suitable for constructing such an inference model.
  • the present invention has decided to provide an information processing system or the like capable of constructing a statistical inference model more easily and efficiently.
  • an information processing system configured to perform each of the following steps:
  • the reception step the user's input to the GUI for constructing the inference model is received.
  • the input includes attribute information for each of the plurality of nodes constituting the inference model and connection relations between the plurality of nodes.
  • the first generation step based on the input, visual information including a plurality of nodes on the GUI is generated in such a manner that at least a part of the attribute information and the connection relationship can be grasped.
  • the program for realizing the software appearing in the present embodiment may be provided as a non-transitory recording medium (Non-Transity Computer-Readable Medium) that can be read by a computer, or may be downloaded from an external server. It may be provided as possible, or it may be provided so that the program is started by an external computer and the function is realized by the client terminal (so-called cloud computing).
  • Non-Transity Computer-Readable Medium Non-Transity Computer-Readable Medium
  • the "part" may include, for example, a combination of hardware resources implemented by a circuit in a broad sense and information processing of software specifically realized by these hardware resources. ..
  • various information is handled in this embodiment, and these information are, for example, physical values of signal values representing voltage and current, and signal values as a bit aggregate of a binary number composed of 0 or 1. It is represented by high and low or quantum superposition (so-called qubit), and communication / operation can be executed on a circuit in a broad sense.
  • a circuit in a broad sense is a circuit realized by at least appropriately combining a circuit, a circuit, a processor, a memory, and the like. That is, an integrated circuit for a specific application (Application Specific Integrated Circuit: ASIC), a programmable logic device (for example, a simple programmable logic device (Simple Programmable Logic Device: SPLD), a composite programmable logic device (Complex Program)) It includes a programmable gate array (Field Programmable Gate Array: FPGA) and the like.
  • FIG. 1 is a configuration diagram showing an information processing system 1 according to the present embodiment.
  • the information processing system 1 includes a user terminal 2 and a server 3, which are connected to each other through a network. These components will be further described.
  • the system exemplified by the information processing system 1 is composed of one or more devices or components. Therefore, even a single server 3 is an example of a system.
  • FIG. 2 is a block diagram showing a hardware configuration of the user terminal 2.
  • the user terminal 2 has a communication unit 21, a storage unit 22, a control unit 23, a display unit 24, and an input unit 25, and these components are inside the user terminal 2 via a communication bus 20. It is electrically connected.
  • the description of the communication unit 21, the storage unit 22, and the control unit 23 will be omitted because they are substantially the same as the communication unit 31, the storage unit 32, and the control unit 33 in the server 3.
  • the display unit 24 may be, for example, included in the housing of the user terminal 2 or may be externally attached.
  • the display unit 24 displays a screen of a graphical user interface (GUI) that can be operated by the user.
  • GUI graphical user interface
  • display devices such as a CRT display, a liquid crystal display, an organic EL display, and a plasma display properly according to the type of the user terminal 2.
  • the display unit 24 will be described as being included in the housing of the user terminal 2.
  • the input unit 25 may be included in the housing of the user terminal 2 or may be externally attached.
  • the input unit 25 may be implemented as a touch panel integrally with the display unit 24. If it is a touch panel, the user can input a tap operation, a swipe operation, and the like.
  • a switch button, a mouse, a QWERTY keyboard, or the like may be adopted. That is, the input unit 25 receives the operation input made by the user.
  • the input is transferred to the control unit 23 as a command signal via the communication bus 20, and the control unit 23 can execute predetermined control or calculation as necessary.
  • FIG. 3 is a block diagram showing a hardware configuration of the server 3.
  • the server 3 has a communication unit 31, a storage unit 32, and a control unit 33, and these components are electrically connected to each other inside the server 3 via a communication bus 30. Each component will be further described.
  • the communication unit 31 is preferably a wired communication means such as USB, IEEE1394, Thunderbolt (registered trademark), wired LAN network communication, etc., wireless LAN network communication, mobile communication such as 3G to 5G, Bluetooth (registered trademark) communication, etc. May be included as needed. That is, it is more preferable to carry out as a set of these plurality of communication means. That is, the server 3 communicates various information with the user terminal 2 via the communication unit 31 and the network.
  • wired communication means such as USB, IEEE1394, Thunderbolt (registered trademark), wired LAN network communication, etc.
  • wireless LAN network communication wireless LAN network communication
  • mobile communication such as 3G to 5G
  • Bluetooth registered trademark
  • the storage unit 32 stores various information defined by the above description. This is necessary, for example, as a storage device such as a solid state drive (Sold State Drive: SSD) that stores various programs related to the server 3 executed by the control unit 33, or temporarily related to program operations. It can be implemented as a memory such as a random access memory (Random Access Memory: RAM) for storing various information (arguments, sequences, etc.). Further, these combinations may be used. In particular, the storage unit 32 stores various programs and the like related to the server 3 executed by the control unit 33.
  • SSD solid state drive
  • RAM random access memory
  • the control unit 33 processes and controls the overall operation related to the server 3.
  • the control unit 33 is, for example, a central processing unit (CPU) (not shown).
  • the control unit 33 realizes various functions related to the server 3 by reading a predetermined program stored in the storage unit 32. That is, the information processing by the software stored in the storage unit 32 can be executed as each functional unit included in the control unit 33 by specifically realizing the information processing by the control unit 33 which is an example of the hardware. These will be described in more detail in the next section.
  • the control unit 33 is not limited to a single unit, and may be implemented so as to have a plurality of control units 33 for each function. Further, it may be a combination thereof.
  • FIG. 4 is a block diagram showing a function realized by the control unit 33 and the like in the server 3.
  • the server 3, which is an example of the information processing system 1, includes the following parts.
  • the reception unit 331 is configured to receive various information from other devices connected to the server 3 via the network.
  • the reception unit 331 may accept an input to the operation screen 4 (see FIGS. 9 to 11), which is an example of the GUI, transmitted from the user terminal 2.
  • the reading unit 332 is configured to read various information received by the receiving unit 331 or stored in advance in the storage unit 32.
  • the reading unit 332 may read the connection information of the node 5 stored in the storage unit 32 in advance.
  • the generation unit 333 is configured to generate various visual information.
  • the visual information may be information itself generated in a manner that can be visually recognized by the user, such as a screen, an image, an icon, a message, or the like.
  • a screen, an image, an icon, a message, or the like is displayed on the display unit 24 of the user terminal 2. It may be rendering information for making the user.
  • the output unit 334 is configured to output various information and transmit various information to other devices connected to the server 3 via the network.
  • the output unit 334 may output a second code which is a general-purpose programming code for the inference model generated by the user via the operation screen 4, or may output the visual information generated by the generation unit 333. It may be transmitted to the user terminal 2.
  • Node type This section describes the node type NK, which is the type of the node 5 (see FIGS. 10 and 11) that constitutes the inference model.
  • the inference model constructed by the information processing system 1 is preferably a Bayesian statistical model, and more preferably a dynamic model.
  • FIG. 5 shows a list of node types NK. As shown in FIG. 5, the node type NK includes variables, functions and probability models.
  • a plurality of nodes 5 and a link 6 indicating the connection relationship thereof are displayed and can be visually recognized by the user like a graphical model (FIGS. 10 and 10). 11).
  • a graphical model FIGS. 10 and 10). 11
  • the node represents a variable to the last, and the link connecting the nodes may have the meaning of a function or a probability model.
  • such a form lacks user intuition.
  • by giving the node 5 itself the meaning of the variable, the function, and the probability model it is possible to provide the user with a more intuitive grasp.
  • the node type NK includes at least one of Data, Transfermed Data, Parameter, Transferformed Parameter, Likelihood & Prior, and Generated Quantity defined in the programming language Santa. More preferably, the node type NK includes State, which is a type representing a state, and Time Evolution, which is a type representing time evolution. In particular, State and Time Evolution are node types NK used when the inference model is a dynamic model. Hereinafter, each node type NK will be described one by one.
  • Data which is an example of node type NK, has a data type, and node 5 of all node types NK can be connected as a child node.
  • Transformed Data which is an example of node type NK, has a data type and a conversion function of a parent node variable, and can connect a node 5 of a node type NK other than Data as a child node.
  • Parameter which is an example of node type NK, has a data type and a range, and can connect node 5 of node type NK other than Data or Dataformed Data as a child node.
  • the Transformed Parameter which is an example of the node type NK, has a data type and a conversion function of the parent node variable, and can connect a node type NK other than Data, Transformed Data, or Parameter as a child node.
  • node type NK has a data type and an initial value, and can connect a node 5 of a node type NK other than Data, Dataformed Data, Parameter or Transaction Parameter as a child node.
  • Time Evolution which is an example of the node type NK, has a description of time evolution regarding the State connected as a parent node or a child node, and only the node 5 whose node type NK is a State can be connected as a child node.
  • Likelihood & Prior which is an example of node type NK, has a description of a probability model such as likelihood and prior probability, and there is no node 5 that can be connected as a child node.
  • Generated Quantity which is an example of node type NK, has a data type and a conversion function of a parent node variable, and only node 5 whose node type NK is Generated Quantity can be connected as a child node.
  • this information processing method includes the following steps.
  • the reception step the user's input to the GUI for constructing the inference model is received.
  • the input includes an attribute information IF for each of the plurality of nodes 5 constituting the inference model, and a connection relationship CR between the plurality of nodes 5.
  • the first generation step based on the input, visual information including a plurality of nodes 5 on the GUI is generated in such a manner that at least a part of the attribute information IF and the connection relationship CR can be grasped.
  • FIG. 6 is an activity diagram showing a flow of service provision by the SaaS type model.
  • the user is described as being a data scientist who desires to build an inference model in order to make a prediction about a predetermined event. It should be noted that this is just an example, and the user is not limited to this.
  • the inference model construction service provided by the information processing system 1 is assumed to be so-called SaaS (Software as a Service) provided by accessing the server 3 from the user terminal 2 via the Internet. This will be described with reference to the activity diagram shown in 6.
  • SaaS Software as a Service
  • FIG. 7 is a schematic diagram showing an example of the login screen 9.
  • the generation unit 333 When the user accesses the URL of the service proposed by the information processing system 1, the generation unit 333 generates the login screen 9 (activity A101) on the server 3, and the output unit 334 transfers this to the user terminal 2 via the network.
  • the login screen 9 is displayed on the display unit 24 of the user terminal 2. That is, the communication unit 31 in the server 3 receives the request for the URL transmitted from the user terminal 2, and the control unit 33 in the server 3 generates a login screen 9 based on the request, which is via the communication unit 31 and the network. Is transmitted to the user terminal 2.
  • the login screen 9 includes, for example, an input field 91 for inputting a user name, an input field 92 for inputting a password, a login button 93 to be pressed at login, and a registration link 94 to be selected at registration.
  • an account registration screen (not shown) is displayed on the display unit 24 of the user terminal 2, and the user may input the initial information of the user as necessary (activity A102).
  • the information is transmitted to the server 3 via the network, and the information about the user's account, such as a user name and a password, is sent to the storage unit 32 in the server 3. Is stored as a database (activity A103).
  • the user can log in to the inference model construction service provided by the information processing system 1.
  • the user-specific user name is entered in the input field 91 and linked to the user name in the input field 92. If you enter the password and press the login button 93 by clicking or tapping it, you can log in to the inference model construction service (Activity A104).
  • the generation unit 333 in the server 3 generates the operation screen 4 which is an example of the GUI (activity A105). That is, after the information about the user's account is stored in the storage unit 32 of the server 3, the control unit 33 of the server 3 executes a predetermined program stored in the storage unit 32 to generate the operation screen 4. ing.
  • the output unit 334 transmits the operation screen 4 to the user terminal 2 via the communication unit 31 and the network, so that the operation screen 4 is displayed on the display unit 24 of the user terminal 2.
  • the user constructs an inference model by operating the operation screen 4 displayed on the display unit 24 on the input unit 25 (activity A106).
  • the specific operation flow of the operation screen 4 will be described later with reference to FIGS. 8 to 12.
  • FIG. 8 is an activity diagram showing an operation flow of the operation screen 4.
  • 9 to 11 are schematic views showing an example of the operation screen 4.
  • the operation flow of the operation screen 4 will be described with reference to the activity diagram shown in FIG.
  • the generation unit 333 generates the operation screen 4, and the output unit 334 transmits this, so that the operation screen 4 is displayed on the display unit 24 of the user terminal 2 (activity A201).
  • the reception unit 331 accepts the user's input to the operation screen 4 for constructing the inference model.
  • the input received by the server 3 via the communication unit 31 is stored in the temporary storage area, which is an example of the storage unit 32, and various processes are executed by the control unit 33.
  • the input includes an attribute information IF for each of the plurality of nodes 5 constituting the inference model, and a link 6 which is an example of the connection relationship CR between the plurality of nodes 5.
  • the user sets the node 5 by inputting the node type NK, which is an example of the attribute information IF, and the node name on the operation screen 4 (activity A202).
  • the generation unit 333 updates the operation screen 4, and the output unit 334 transmits this, so that the updated operation screen 4 is displayed on the display unit 24 of the user terminal 2. That is, the set node 5 is displayed on the operation screen 4 (activity A203).
  • the user inputs the first code, which is an example of the attribute information IF corresponding to the set node 5, to the operation screen 4 (activity A204). After repeating these a plurality of times, the user inputs the link 6 indicating the connection relationship CR between the nodes 5 to the operation screen 4 (activity A205).
  • the attribute information IF is the node type NK
  • the connection relation CR between the nodes 5 is restricted by the node type NK.
  • node type NK and its connection relation CR refer to the previous section. According to such an aspect, it is possible to prevent the user from connecting the erroneous nodes 5 to each other and thereby generating a first code or a second code having a bug.
  • the second code will be described later. In particular, not only is it restricted, but a warning that the user cannot be connected may be output. This will be described later.
  • the operation screen 4 which is an example of the GUI is an example of an object in which the node type NK can be selected from predetermined items.
  • the selection field 41 is implemented, for example, by a pull-down list.
  • the reception unit 331 can accept the selection of the node type NK by the user as an input. In this way, by letting the user select the node type NK from the predetermined items, even a person who does not remember the exact name or spelling of the node type NK can efficiently determine the node type NK. It is user-friendly because it can be used.
  • the operation screen 4 further includes an input field 42, a button 43, an area 44, and a button 46.
  • the input field 42 is a field for allowing the user to input the name (node name) of the node 5.
  • the button 43 is an object for creating the node 5.
  • the node 5 is generated in the area 44.
  • Data which is an example of node type NK
  • Patient which is a node name
  • the node 51 is generated in the inside.
  • the button 46 will be described later.
  • the input field 45 is displayed when the user selects the node 5 displayed in the area 44.
  • the user has selected node 53.
  • the node 53 has a node type NK of Time Evolution and a node name of ODE.
  • the node 51 described above and the node 52 whose node type NK is State and whose node name is Possible are nodes that have not been selected by the user.
  • the attribute information IF other than the node type NK is input to the input field 45.
  • the attribute information IF that can be input to the input field 45 includes a first code representing information processing in the corresponding node 5.
  • the declaration of a local variable used in the corresponding node 5 can be included in the first code.
  • the numerical assignment to the declared local variable, the information processing using the declared local variable, and the like can be included in the first code.
  • the first code constitutes at least a part of the second code, which is the entire source code 7 (see FIG. 12) finally output by the output unit 334.
  • the first code can be a good educational material, especially for a user who wants to implement the code without using the information processing system 1 in the future.
  • the user can set the connection relationship CR between the nodes 5.
  • the nodes 5 are connected by an arrow-shaped link 6, and the upstream-downstream relationship between the nodes 5 can be grasped. Since the link 6 is linear visual information, it is preferable to provide the operation icon 6a in order to make it easier to handle and finely adjust the position and orientation.
  • the reception unit 331 accepts an input in which two of the plurality of nodes 5 are selected in order. Then, the generation unit 333 grasps the upstream side and the downstream side by setting the first node selected first in the order, for example, the node 51 as the upstream side, and the second node selected later, for example, the node 52 as the downstream side. Generates visual information that is shown in a possible manner. For example, it is preferable that the visual information is displayed on the display unit 24 of the user terminal 2 so that it can be grasped by the direction of the arrow. In the example shown in FIG. 11, the link 61 shows the connection relationship CR between the node 51 on the upstream side and the node 52 on the downstream side.
  • the link 62 shows the connection relationship CR between the node 52 on the upstream side and the node 53 on the downstream side.
  • the user can determine the upstream / downstream of the node 5 with a simple operation, and can grasp this on the operation screen 4 which is an example of the GUI, so that it is user-friendly. Is.
  • the output unit 334 may be configured to output a warning when the connection-related CR of the unconnectable node 5 is input.
  • the reading unit 332 reads out the connection pros and cons information previously stored in the storage unit 32.
  • the connectable node type NK of each node type NK described in Section 3 is summarized in the form of a look-up table or the like.
  • the generation unit 333 When the connection is possible, the generation unit 333 generates visual information including the link 6 as shown in FIG. 11 in the operation screen 4, and the output unit 334 transmits this, so that this visual information is transmitted to the user terminal 2.
  • a warning (activity A207) not shown may be presented to the user.
  • the warning may be visual information such as an alert icon or voice information such as a warning sound.
  • the generation unit 333 generates visual information that can distinguish the node type NK in the form of a frame indicating the node 5 in relation to the node type NK.
  • the form here is a concept including a shape, a pattern or a color, or a combination thereof.
  • the node 51, the node 52, and the node 53 have different pixel information in the frame. That is, it is configured so that the user can distinguish it by having at least one of hue, lightness, and saturation different.
  • the shape of the frame itself may be distinguishable.
  • the shape of the frame for example, a rectangle, a square, a circle, an ellipse, or the like can be adopted. In this way, the form of the frame showing the node 5 allows the user to intuitively grasp the node type NK, which is user-friendly.
  • the node 5 setting operation (activity A202 to A204) and the connection-related CR setting operation (activity A205 to A207) may be freely repeated.
  • the node 51 and the node 52 are set to set the link 61 which is the connection relation CR connecting them, and then the node 53 is set and the link 62 which is the connection relation CR connecting to the node 52 is set.
  • the generation unit 333 generates a plurality of nodes 5 in a manner in which at least a part of the attribute information IF and the connection relationship CR can be grasped based on the input. Generate the visual information included on the operation screen 4.
  • the output unit 334 in the server 3 outputs a second code indicating an inference model corresponding to the mode and attribute information IF displayed in the area 44 of the operation screen 4. (Activity A209).
  • FIG. 12 shows a source code 7 which is an example of the second code output by the output unit 334.
  • the output source code 7 may be transmitted to the user terminal 2 by the output unit 334 and displayed on the display unit 24, or may be downloaded as data.
  • the code for operating the constructed inference model can be output as a source code 7 (second code) that can be used for general purposes.
  • the user can appropriately compile the second code generated in this way in the local environment using the user terminal 2 to generate an inference model, whereby inference about a desired event can be performed.
  • the source code can be transferred to a person who does not use the information processing system 1 according to the present embodiment.
  • the second code may be generated for the first time when the button 46 is pressed, but more preferably, the second code is generated based on the node 5 or the like displayed in the region 44 at the stage of constructing the inference model. Code should be generated or updated sequentially.
  • the computer may be provided with a program for causing the computer to execute each step of the information processing system 1.
  • an information processing method including each step of the information processing system 1 may be provided.
  • the information processing system 1 may further include an execution unit (not shown). That is, compilation on the user terminal 2 is not required, and the execution unit on the server 3 may be configured to be able to execute the second code. More preferably, as the second generation step, the generation unit 333 may generate visual information including the estimation result 8 by the inference model regarding the set event based on the execution result of the second code.
  • FIG. 13 is an example of the estimation result 8 obtained by the inference model. According to such an embodiment, the user can perform one-stop operations from the construction of the node 5 to the execution of the inference model, so that various events can be predicted using the inference model more quickly and efficiently. be able to.
  • a warning is issued when the nodes 5 cannot be connected to each other with the node type NK, but instead of issuing the warning, the connection relationship CR of the unreachable node 5 is issued.
  • the generation unit 333 may be configured not to generate visual information including this unconnectable connection relationship CR.
  • a warning may be issued. That is, as a third output step, when the output unit 334 does not quote the variable specified in the node 51 which is the parent node on the upstream side of the node 52 in the predetermined node 5, for example, the node 52. It may be configured to output a warning. Further, preferably, the output unit 334 may be configured to output a warning when the variable specified in the child node on the downstream side of the predetermined node 5 is not quoted in the predetermined node 5. .. According to such an aspect, it is possible to prevent omission of citation of variables.
  • an object that can refer to the variables specified in the parent node or the child node may be displayed with one click. That is, the operation screen 4 which is an example of the GUI may include an object that can refer to a variable defined in a predetermined node 5, for example, a node 51 which is a parent node on the upstream side of the node 52. Further, preferably, the operation screen 4 which is an example of the GUI may include an object which can refer to the variable specified in the child node on the downstream side of the predetermined node 5. According to such an embodiment, after noticing a variable citation omission, it can be quickly corrected.
  • the present embodiment has been described as a service provided by the SaaS model, the present invention is not limited to this, and even if the user installs a predetermined program on the user terminal 2, the service is implemented as a stand-alone service. good. Such programs may be distributed.
  • the output unit 334 may issue a warning. According to such an aspect, it is possible to prevent unnecessary creation of the node 5 and prevent link leakage.
  • FIG. 14 shows another aspect of the operation screen 4. That is, the operation screen 4 may be implemented according to such an aspect.
  • the information processing system In the information processing system, at least a part of the attribute information is a node type, and the connection relationship between the nodes is limited by the node type.
  • the information processing system is configured to further execute a first output step, and the first output step is configured to output a warning when a connection relationship of the unconnectable node is input. Things to be done.
  • the first generation step when the connection relationship of the unconnectable node is input, the visual information including the unconnectable connection relationship is not generated. ..
  • the visual information that can distinguish the node type is generated in the form of a frame indicating the node, wherein the form is a shape, a pattern, a color, or these. Things that include the binding of.
  • the GUI includes an object in which the node type can be selected from predetermined items, and in the reception step, the selection of the node type by the user is received as the input.
  • the node type includes a variable, a function, and a probability model.
  • the node type includes at least one of Data, Translated Data, Parameter, Transformed Parameter, Likelihood & Prior and Generated Quantity defined in the programming language Stan.
  • the attribute information includes a first code representing information processing in the corresponding node.
  • a second code indicating the inference model which is configured to further execute a second output step, which corresponds to the displayed embodiment and the attribute information. To output, stuff.
  • the execution step is configured to be further executed, and the execution step executes the second code.
  • the second generation step is further executed, and the second generation step is based on the inference model regarding the set event based on the execution result of the second code.
  • the attribute information includes a first code representing information processing in the corresponding node, and the first code constitutes at least a part of the second code.
  • the reception step the input in which two of the plurality of nodes are selected in order is received, and in the first generation step, the first node selected first in the order is received. Is the upstream side, and the second node selected later is the downstream side, and the visual information showing the upstream side and the downstream side in a graspable manner is generated.
  • the GUI includes an object that can refer to a variable defined by a parent node on the upstream side or a child node on the downstream side of a predetermined node.
  • An information processing method comprising each step of the information processing system.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • Business, Economics & Management (AREA)
  • Strategic Management (AREA)
  • Human Resources & Organizations (AREA)
  • Economics (AREA)
  • Human Computer Interaction (AREA)
  • Development Economics (AREA)
  • Marketing (AREA)
  • Computing Systems (AREA)
  • Evolutionary Computation (AREA)
  • Data Mining & Analysis (AREA)
  • Game Theory and Decision Science (AREA)
  • Computational Linguistics (AREA)
  • Artificial Intelligence (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Mathematical Physics (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • User Interface Of Digital Computer (AREA)
  • Stored Programmes (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

[Problem] To provide an information processing system, etc., with which it is possible to more easily and efficiently construct a statistical inference model. [Solution] According to one embodiment of the present invention, an information processing system is provided. This information processing system is constituted so as to execute each of the steps below. In an acceptance step, a user's input to a GUI for constructing an inference model is accepted. The input includes attribute information for each of a plurality of nodes constituting the inference model, and a connection relationship between the plurality of nodes. In a first generation step, visual information that includes the plurality of nodes on the GUI is generated on the basis of the input, in such a manner as for at least some of the attribute information, and the connection relationship, to be graspable.

Description

情報処理システム、情報処理方法及びプログラムInformation processing system, information processing method and program
　本発明は、情報処理システム、情報処理方法及びプログラムに関する。 The present invention relates to an information processing system, an information processing method and a program.
　感染症の伝播、商品の出荷数、消費者の動向、金融商品の価格など、様々な事象を予測する上で、データ生成に関するドメイン知識をモデルに組み込み、データとモデルから推論を行う統計学的な推論モデルが用いられている。 Statistical inference from data and models by incorporating domain knowledge about data generation into models in predicting various events such as transmission of infectious diseases, shipments of goods, consumer trends, prices of financial products, etc. Inference model is used.
　非特許文献１には、このような推論モデルを構築するのに適したプログラミング言語が開示されている。 Non-Patent Document 1 discloses a programming language suitable for constructing such an inference model.
“Ｓｔａｎ　（ｓｏｆｔｗａｒｅ）　Ｆｒｏｍ　Ｗｉｋｉｐｅｄｉａ，　ｔｈｅ　ｆｒｅｅ　ｅｎｃｙｃｌｏｐｅｄｉａ”，［令和２年１２月２５日検索，インターネット＜ＵＲＬ：ｈｔｔｐｓ：／／ｅｎ．ｗｉｋｉｐｅｄｉａ．ｏｒｇ／ｗｉｋｉ／Ｓｔａｎ＿（ｓｏｆｔｗａｒｅ）＞］"Stan (software) From Wikipedia, the free encyclopedia", [Search on December 25, 2nd year of Reiwa, Internet <URL: https: // en. wikipedia. org / wiki / Tan_ (software)>]
　しかしながら、統計学的な推論モデルにドメイン知識を組み込むには、人手での試行錯誤が必要となり、ドメイン知識を利用しない、従来の機械学習モデルや統計モデルの実装に比べて、煩雑だった。また、ドメイン知識を組み込んだ統計学的な推論モデルには、変数や関数間に複雑な依存関係が存在し、これを把握しながら誤りなくモデル構造を構築するのは、困難な作業だった。 However, incorporating domain knowledge into a statistical inference model requires manual trial and error, which is more complicated than the implementation of conventional machine learning models and statistical models that do not use domain knowledge. In addition, statistical inference models that incorporate domain knowledge have complex dependencies between variables and functions, and it was a difficult task to construct a model structure without error while grasping these.
　本発明では上記事情を鑑み、より容易かつ効率的に統計学的な推論モデルを構築可能な情報処理システム等を提供することとした。 In view of the above circumstances, the present invention has decided to provide an information processing system or the like capable of constructing a statistical inference model more easily and efficiently.
　本発明の一態様によれば、情報処理システムが提供される。この情報処理システムは、次の各ステップを実行するように構成される。受付ステップでは、推論モデルを構築するためのＧＵＩに対するユーザの入力を受け付ける。入力は、推論モデルを構成する複数のノードそれぞれに対する属性情報と、複数のノード間の接続関係とを含む。第１の生成ステップでは、入力に基づいて、属性情報の少なくとも一部と、接続関係とが把握可能な態様で、複数のノードをＧＵＩ上に含む視覚情報を生成する。 According to one aspect of the present invention, an information processing system is provided. The information processing system is configured to perform each of the following steps: In the reception step, the user's input to the GUI for constructing the inference model is received. The input includes attribute information for each of the plurality of nodes constituting the inference model and connection relations between the plurality of nodes. In the first generation step, based on the input, visual information including a plurality of nodes on the GUI is generated in such a manner that at least a part of the attribute information and the connection relationship can be grasped.
　このような一態様によれば、変数や関数間の依存関係を把握しながら、誤りを削減してより容易かつ効率的に推論モデルを構築することができる。 According to such an aspect, it is possible to reduce errors and construct an inference model more easily and efficiently while grasping the dependency relationships between variables and functions.
本実施形態に係る情報処理システム１を表す構成図である。It is a block diagram which shows the information processing system 1 which concerns on this embodiment. ユーザ端末２のハードウェア構成を示すブロック図である。It is a block diagram which shows the hardware composition of a user terminal 2. サーバ３のハードウェア構成を示すブロック図である。It is a block diagram which shows the hardware configuration of a server 3. サーバ３における制御部３３等によって実現される機能を示すブロック図である。It is a block diagram which shows the function realized by the control part 33 and the like in a server 3. ノード種別ＮＫの一覧を示している。A list of node type NK is shown. ＳａａＳ型モデルによるサービス提供の流れを示すアクティビティ図である。It is an activity diagram which shows the flow of service provision by a SaaS type model. ログイン画面９の一例を示す概要図である。It is a schematic diagram which shows an example of a login screen 9. 操作画面４の操作の流れを示すアクティビティ図である。It is an activity diagram which shows the operation flow of operation screen 4. 操作画面４の一例を示す概要図である。It is a schematic diagram which shows an example of the operation screen 4. 操作画面４の一例を示す概要図である。It is a schematic diagram which shows an example of the operation screen 4. 操作画面４の一例を示す概要図である。It is a schematic diagram which shows an example of the operation screen 4. 出力部３３４が出力した第２のコードの一例であるソースコード７を示している。The source code 7 which is an example of the 2nd code output by the output part 334 is shown. 推論モデルによって得られた推定結果８の一例である。This is an example of the estimation result 8 obtained by the inference model. 操作画面４の別の態様を示している。Another aspect of the operation screen 4 is shown.
　以下、図面を用いて本発明の実施形態について説明する。以下に示す実施形態中で示した各種特徴事項は、互いに組み合わせ可能である。 Hereinafter, embodiments of the present invention will be described with reference to the drawings. The various features shown in the embodiments shown below can be combined with each other.
　ところで、本実施形態に登場するソフトウェアを実現するためのプログラムは、コンピュータが読み取り可能な非一時的な記録媒体（Ｎｏｎ－Ｔｒａｎｓｉｔｏｒｙ　Ｃｏｍｐｕｔｅｒ－Ｒｅａｄａｂｌｅ　Ｍｅｄｉｕｍ）として提供されてもよいし、外部のサーバからダウンロード可能に提供されてもよいし、外部のコンピュータで当該プログラムを起動させてクライアント端末でその機能を実現（いわゆるクラウドコンピューティング）するように提供されてもよい。 By the way, the program for realizing the software appearing in the present embodiment may be provided as a non-transitory recording medium (Non-Transity Computer-Readable Medium) that can be read by a computer, or may be downloaded from an external server. It may be provided as possible, or it may be provided so that the program is started by an external computer and the function is realized by the client terminal (so-called cloud computing).
　また、本実施形態において「部」とは、例えば、広義の回路によって実施されるハードウェア資源と、これらのハードウェア資源によって具体的に実現されうるソフトウェアの情報処理とを合わせたものも含みうる。また、本実施形態においては様々な情報を取り扱うが、これら情報は、例えば電圧・電流を表す信号値の物理的な値、０又は１で構成される２進数のビット集合体としての信号値の高低、又は量子的な重ね合わせ（いわゆる量子ビット）によって表され、広義の回路上で通信・演算が実行されうる。 Further, in the present embodiment, the "part" may include, for example, a combination of hardware resources implemented by a circuit in a broad sense and information processing of software specifically realized by these hardware resources. .. In addition, various information is handled in this embodiment, and these information are, for example, physical values of signal values representing voltage and current, and signal values as a bit aggregate of a binary number composed of 0 or 1. It is represented by high and low or quantum superposition (so-called qubit), and communication / operation can be executed on a circuit in a broad sense.
　また、広義の回路とは、回路（Ｃｉｒｃｕｉｔ）、回路類（Ｃｉｒｃｕｉｔｒｙ）、プロセッサ（Ｐｒｏｃｅｓｓｏｒ）、及びメモリ（Ｍｅｍｏｒｙ）等を少なくとも適当に組み合わせることによって実現される回路である。すなわち、特定用途向け集積回路（Ａｐｐｌｉｃａｔｉｏｎ　Ｓｐｅｃｉｆｉｃ　Ｉｎｔｅｇｒａｔｅｄ　Ｃｉｒｃｕｉｔ：ＡＳＩＣ）、プログラマブル論理デバイス（例えば、単純プログラマブル論理デバイス（Ｓｉｍｐｌｅ　Ｐｒｏｇｒａｍｍａｂｌｅ　Ｌｏｇｉｃ　Ｄｅｖｉｃｅ：ＳＰＬＤ）、複合プログラマブル論理デバイス（Ｃｏｍｐｌｅｘ　Ｐｒｏｇｒａｍｍａｂｌｅ　Ｌｏｇｉｃ　Ｄｅｖｉｃｅ：ＣＰＬＤ）、及びフィールドプログラマブルゲートアレイ（Ｆｉｅｌｄ　Ｐｒｏｇｒａｍｍａｂｌｅ　Ｇａｔｅ　Ａｒｒａｙ：ＦＰＧＡ））等を含むものである。 Further, a circuit in a broad sense is a circuit realized by at least appropriately combining a circuit, a circuit, a processor, a memory, and the like. That is, an integrated circuit for a specific application (Application Specific Integrated Circuit: ASIC), a programmable logic device (for example, a simple programmable logic device (Simple Programmable Logic Device: SPLD), a composite programmable logic device (Complex Program)) It includes a programmable gate array (Field Programmable Gate Array: FPGA) and the like.
１．ハードウェア構成
　本節では、本実施形態のハードウェア構成について説明する。 1. 1. Hardware Configuration This section describes the hardware configuration of this embodiment.
１．１　情報処理システム１
　図１は、本実施形態に係る情報処理システム１を表す構成図である。情報処理システム１はユーザ端末２と、サーバ３とを備え、これらがネットワークを通じて接続されている。これらの構成要素についてさらに説明する。ここで、情報処理システム１に例示されるシステムとは、１つ又はそれ以上の装置又は構成要素からなるものである。したがって、サーバ３単体であってもシステムの一例となる。 1.1 Information processing system 1
FIG. 1 is a configuration diagram showing an information processing system 1 according to the present embodiment. The information processing system 1 includes a user terminal 2 and a server 3, which are connected to each other through a network. These components will be further described. Here, the system exemplified by the information processing system 1 is composed of one or more devices or components. Therefore, even a single server 3 is an example of a system.
１．２　ユーザ端末２
　図２は、ユーザ端末２のハードウェア構成を示すブロック図である。ユーザ端末２は、通信部２１と、記憶部２２と、制御部２３と、表示部２４と、入力部２５とを有し、これらの構成要素がユーザ端末２の内部において通信バス２０を介して電気的に接続されている。通信部２１、記憶部２２及び制御部２３の説明は、サーバ３における通信部３１、記憶部３２及び制御部３３と略同様のため省略する。 1.2 User terminal 2
FIG. 2 is a block diagram showing a hardware configuration of the user terminal 2. The user terminal 2 has a communication unit 21, a storage unit 22, a control unit 23, a display unit 24, and an input unit 25, and these components are inside the user terminal 2 via a communication bus 20. It is electrically connected. The description of the communication unit 21, the storage unit 22, and the control unit 23 will be omitted because they are substantially the same as the communication unit 31, the storage unit 32, and the control unit 33 in the server 3.
　表示部２４は、例えば、ユーザ端末２の筐体に含まれるものであってもよいし、外付けされるものであってもよい。表示部２４は、ユーザが操作可能なグラフィカルユーザインターフェース（Ｇｒａｐｈｉｃａｌ　Ｕｓｅｒ　Ｉｎｔｅｒｆａｃｅ：ＧＵＩ）の画面を表示する。これは例えば、ＣＲＴディスプレイ、液晶ディスプレイ、有機ＥＬディスプレイ及びプラズマディスプレイ等の表示デバイスを、ユーザ端末２の種類に応じて使い分けて実施することが好ましい。ここでは、表示部２４は、ユーザ端末２の筐体に含まれるものとして説明する。 The display unit 24 may be, for example, included in the housing of the user terminal 2 or may be externally attached. The display unit 24 displays a screen of a graphical user interface (GUI) that can be operated by the user. For example, it is preferable to use display devices such as a CRT display, a liquid crystal display, an organic EL display, and a plasma display properly according to the type of the user terminal 2. Here, the display unit 24 will be described as being included in the housing of the user terminal 2.
　入力部２５は、ユーザ端末２の筐体に含まれるものであってもよいし、外付けされるものであってもよい。例えば、入力部２５は、表示部２４と一体となってタッチパネルとして実施されてもよい。タッチパネルであれば、ユーザは、タップ操作、スワイプ操作等を入力することができる。もちろん、タッチパネルに代えて、スイッチボタン、マウス、ＱＷＥＲＴＹキーボード等を採用してもよい。すなわち、入力部２５がユーザによってなされた操作入力を受け付ける。当該入力が命令信号として、通信バス２０を介して制御部２３に転送され、制御部２３が必要に応じて所定の制御や演算を実行しうる。 The input unit 25 may be included in the housing of the user terminal 2 or may be externally attached. For example, the input unit 25 may be implemented as a touch panel integrally with the display unit 24. If it is a touch panel, the user can input a tap operation, a swipe operation, and the like. Of course, instead of the touch panel, a switch button, a mouse, a QWERTY keyboard, or the like may be adopted. That is, the input unit 25 receives the operation input made by the user. The input is transferred to the control unit 23 as a command signal via the communication bus 20, and the control unit 23 can execute predetermined control or calculation as necessary.
１．３　サーバ３
　図３は、サーバ３のハードウェア構成を示すブロック図である。サーバ３は、通信部３１と、記憶部３２と、制御部３３とを有し、これらの構成要素がサーバ３の内部において通信バス３０を介して電気的に接続されている。各構成要素についてさらに説明する。 1.3 Server 3
FIG. 3 is a block diagram showing a hardware configuration of the server 3. The server 3 has a communication unit 31, a storage unit 32, and a control unit 33, and these components are electrically connected to each other inside the server 3 via a communication bus 30. Each component will be further described.
　通信部３１は、ＵＳＢ、ＩＥＥＥ１３９４、Ｔｈｕｎｄｅｒｂｏｌｔ（登録商標）、有線ＬＡＮネットワーク通信等といった有線型の通信手段が好ましいものの、無線ＬＡＮネットワーク通信、３Ｇ～５Ｇ等のモバイル通信、Ｂｌｕｅｔｏｏｔｈ（登録商標）通信等を必要に応じて含めてもよい。すなわち、これら複数の通信手段の集合として実施することがより好ましい。すなわち、サーバ３は、通信部３１を介して、ユーザ端末２とネットワークを介して種々の情報を通信する。 Although the communication unit 31 is preferably a wired communication means such as USB, IEEE1394, Thunderbolt (registered trademark), wired LAN network communication, etc., wireless LAN network communication, mobile communication such as 3G to 5G, Bluetooth (registered trademark) communication, etc. May be included as needed. That is, it is more preferable to carry out as a set of these plurality of communication means. That is, the server 3 communicates various information with the user terminal 2 via the communication unit 31 and the network.
　記憶部３２は、前述の記載により定義される様々な情報を記憶する。これは、例えば、制御部３３によって実行されるサーバ３に係る種々のプログラム等を記憶するソリッドステートドライブ（Ｓｏｌｉｄ　Ｓｔａｔｅ　Ｄｒｉｖｅ：ＳＳＤ）等のストレージデバイスとして、あるいは、プログラムの演算に係る一時的に必要な情報（引数、配列等）を記憶するランダムアクセスメモリ（Ｒａｎｄｏｍ　Ａｃｃｅｓｓ　Ｍｅｍｏｒｙ：ＲＡＭ）等のメモリとして実施されうる。また、これらの組合せであってもよい。特に、記憶部３２は、制御部３３によって実行されるサーバ３に係る種々のプログラム等を記憶している。 The storage unit 32 stores various information defined by the above description. This is necessary, for example, as a storage device such as a solid state drive (Sold State Drive: SSD) that stores various programs related to the server 3 executed by the control unit 33, or temporarily related to program operations. It can be implemented as a memory such as a random access memory (Random Access Memory: RAM) for storing various information (arguments, sequences, etc.). Further, these combinations may be used. In particular, the storage unit 32 stores various programs and the like related to the server 3 executed by the control unit 33.
　制御部３３は、サーバ３に関連する全体動作の処理・制御を行う。制御部３３は、例えば不図示の中央処理装置（Ｃｅｎｔｒａｌ　Ｐｒｏｃｅｓｓｉｎｇ　Ｕｎｉｔ：ＣＰＵ）である。制御部３３は、記憶部３２に記憶された所定のプログラムを読み出すことによって、サーバ３に係る種々の機能を実現する。すなわち、記憶部３２に記憶されているソフトウェアによる情報処理が、ハードウェアの一例である制御部３３によって具体的に実現されることで、制御部３３に含まれる各機能部として実行されうる。これらについては、次節においてさらに詳述する。なお、制御部３３は単一であることに限定されず、機能ごとに複数の制御部３３を有するように実施してもよい。またそれらの組合せであってもよい。 The control unit 33 processes and controls the overall operation related to the server 3. The control unit 33 is, for example, a central processing unit (CPU) (not shown). The control unit 33 realizes various functions related to the server 3 by reading a predetermined program stored in the storage unit 32. That is, the information processing by the software stored in the storage unit 32 can be executed as each functional unit included in the control unit 33 by specifically realizing the information processing by the control unit 33 which is an example of the hardware. These will be described in more detail in the next section. The control unit 33 is not limited to a single unit, and may be implemented so as to have a plurality of control units 33 for each function. Further, it may be a combination thereof.
２．機能構成
　本節では、本実施形態の機能構成について説明する。前述の通り、記憶部３２に記憶されているソフトウェアによる情報処理がハードウェアの一例である制御部３３によって具体的に実現されることで、制御部３３に含まれる各機能部として実行されうる。 2. 2. Functional configuration This section describes the functional configuration of this embodiment. As described above, information processing by software stored in the storage unit 32 is specifically realized by the control unit 33, which is an example of hardware, and can be executed as each functional unit included in the control unit 33.
　図４は、サーバ３における制御部３３等によって実現される機能を示すブロック図である。具体的には、情報処理システム１の一例であるサーバ３は、次の各部を備える。 FIG. 4 is a block diagram showing a function realized by the control unit 33 and the like in the server 3. Specifically, the server 3, which is an example of the information processing system 1, includes the following parts.
　受付部３３１は、サーバ３とネットワークを介して接続された他の機器から種々の情報を受け付けるように構成される。例えば、受付部３３１は、ユーザ端末２から送信された、ＧＵＩの一例である操作画面４（図９～図１１参照）に対する入力を受け付けてもよい。 The reception unit 331 is configured to receive various information from other devices connected to the server 3 via the network. For example, the reception unit 331 may accept an input to the operation screen 4 (see FIGS. 9 to 11), which is an example of the GUI, transmitted from the user terminal 2.
　読出部３３２は、受付部３３１が受け付けた又は予め記憶部３２に記憶された種々の情報を読み出すように構成される。例えば、読出部３３２は、予め記憶部３２に記憶されたノード５の接続是非情報を読み出してもよい。 The reading unit 332 is configured to read various information received by the receiving unit 331 or stored in advance in the storage unit 32. For example, the reading unit 332 may read the connection information of the node 5 stored in the storage unit 32 in advance.
　生成部３３３は、種々の視覚情報を生成するように構成される。視覚情報とは、画面、画像、アイコン、メッセージ等といった、ユーザが視認可能な態様で生成された情報そのものでもよいし、例えばユーザ端末２の表示部２４に画面、画像、アイコン、メッセージ等を表示させるためのレンダリング情報であってもよい。 The generation unit 333 is configured to generate various visual information. The visual information may be information itself generated in a manner that can be visually recognized by the user, such as a screen, an image, an icon, a message, or the like. For example, a screen, an image, an icon, a message, or the like is displayed on the display unit 24 of the user terminal 2. It may be rendering information for making the user.
　出力部３３４は、種々の情報を出力したり、サーバ３とネットワークを介して接続された他の機器に種々の情報を送信するように構成される。例えば、出力部３３４は、操作画面４を介してユーザに生成された推論モデルを汎用のプログラミングコードである第２のコードを出力してもよいし、生成部３３３で生成された視覚情報を、ユーザ端末２に送信してもよい。 The output unit 334 is configured to output various information and transmit various information to other devices connected to the server 3 via the network. For example, the output unit 334 may output a second code which is a general-purpose programming code for the inference model generated by the user via the operation screen 4, or may output the visual information generated by the generation unit 333. It may be transmitted to the user terminal 2.
３．ノード種別
　本節では、推論モデルを構成するノード５（図１０及び図１１参照）の種別であるノード種別ＮＫについて説明する。情報処理システム１によって構築される推論モデルは、好ましくは、ベイズ統計モデルであり、さらに好ましくは、動的モデルである。図５は、ノード種別ＮＫの一覧を示している。図５に示されるように、ノード種別ＮＫは、変数、関数及び確率モデルを含む。 3. 3. Node type This section describes the node type NK, which is the type of the node 5 (see FIGS. 10 and 11) that constitutes the inference model. The inference model constructed by the information processing system 1 is preferably a Bayesian statistical model, and more preferably a dynamic model. FIG. 5 shows a list of node types NK. As shown in FIG. 5, the node type NK includes variables, functions and probability models.
　次節で詳述するＧＵＩの一例である、操作画面４には、複数のノード５と、その接続関係を示すリンク６とが表示され、グラフィカルモデルのようにユーザに視認されうる（図１０及び図１１参照）。ただし、従来のグラフィカルモデルでは、ノード種別ＮＫという概念はなく、ノードはあくまでも変数を表すものであり、ノード間を接続するリンクに関数や確率モデルの意味を持たせることがあった。しかしながら、このような形態は、ユーザの直感性に欠けるものであった。本実施形態では、ノード５自体に、変数、関数及び確率モデルの意味を持たせていることで、より直感的な把握をユーザに提供することができる。 On the operation screen 4, which is an example of the GUI described in detail in the next section, a plurality of nodes 5 and a link 6 indicating the connection relationship thereof are displayed and can be visually recognized by the user like a graphical model (FIGS. 10 and 10). 11). However, in the conventional graphical model, there is no concept of node type NK, the node represents a variable to the last, and the link connecting the nodes may have the meaning of a function or a probability model. However, such a form lacks user intuition. In the present embodiment, by giving the node 5 itself the meaning of the variable, the function, and the probability model, it is possible to provide the user with a more intuitive grasp.
　より具体的には、ノード種別ＮＫは、プログラミング言語であるＳｔａｎにおいて規定された、Ｄａｔａ、Ｔｒａｎｓｆｏｒｍｅｄ　Ｄａｔａ、Ｐａｒａｍｅｔｅｒ、Ｔｒａｎｓｆｏｒｍｅｄ　Ｐａｒａｍｅｔｅｒ、Ｌｉｋｅｌｉｈｏｏｄ　＆　Ｐｒｉｏｒ及びＧｅｎｅｒａｔｅｄ　Ｑｕａｎｔｉｔｙのうちの少なくとも１つを含む。さらに好ましくは、ノード種別ＮＫは、状態を表す種別であるＳｔａｔｅと、時間発展を表す種別であるＴｉｍｅ　Ｅｖｏｌｕｔｉｏｎとを含む。特に、Ｓｔａｔｅ及びＴｉｍｅ　Ｅｖｏｌｕｔｉｏｎは、推論モデルが動的モデルである場合に使用するノード種別ＮＫである。以下、それぞれのノード種別ＮＫについて１つずつ説明する。 More specifically, the node type NK includes at least one of Data, Transfermed Data, Parameter, Transferformed Parameter, Likelihood & Prior, and Generated Quantity defined in the programming language Santa. More preferably, the node type NK includes State, which is a type representing a state, and Time Evolution, which is a type representing time evolution. In particular, State and Time Evolution are node types NK used when the inference model is a dynamic model. Hereinafter, each node type NK will be described one by one.
　ノード種別ＮＫの一例であるＤａｔａは、データ型を有し、全てのノード種別ＮＫのノード５を子ノードとして接続することができる。 Data, which is an example of node type NK, has a data type, and node 5 of all node types NK can be connected as a child node.
　ノード種別ＮＫの一例であるＴｒａｎｓｆｏｒｍｅｄ　Ｄａｔａは、データ型と親ノード変数の変換関数とを有し、Ｄａｔａ以外のノード種別ＮＫのノード５を子ノードとして接続することができる。 Transformed Data, which is an example of node type NK, has a data type and a conversion function of a parent node variable, and can connect a node 5 of a node type NK other than Data as a child node.
　ノード種別ＮＫの一例であるＰａｒａｍｅｔｅｒは、データ型と値域とを有し、Ｄａｔａ又はＴｒａｎｓｆｏｒｍｅｄ　Ｄａｔａ以外のノード種別ＮＫのノード５を子ノードとして接続することができる。 Parameter, which is an example of node type NK, has a data type and a range, and can connect node 5 of node type NK other than Data or Dataformed Data as a child node.
　ノード種別ＮＫの一例であるＴｒａｎｓｆｏｒｍｅｄ　Ｐａｒａｍｅｔｅｒは、データ型と親ノード変数の変換関数とを有し、Ｄａｔａ、Ｔｒａｎｓｆｏｒｍｅｄ　Ｄａｔａ又はＰａｒａｍｅｔｅｒ以外のノード種別ＮＫを子ノードとして接続することができる。 The Transformed Parameter, which is an example of the node type NK, has a data type and a conversion function of the parent node variable, and can connect a node type NK other than Data, Transformed Data, or Parameter as a child node.
　ノード種別ＮＫの一例であるＳｔａｔｅは、データ型と初期値とを有し、Ｄａｔａ、Ｔｒａｎｓｆｏｒｍｅｄ　Ｄａｔａ、Ｐａｒａｍｅｔｅｒ又はＴｒａｎｓｆｏｒｍｅｄ　Ｐａｒａｍｅｔｅｒ以外のノード種別ＮＫのノード５を子ノードとして接続することができる。 State, which is an example of node type NK, has a data type and an initial value, and can connect a node 5 of a node type NK other than Data, Dataformed Data, Parameter or Transaction Parameter as a child node.
　ノード種別ＮＫの一例であるＴｉｍｅ　Ｅｖｏｌｕｔｉｏｎは、親ノード又は子ノードとして接続されたＳｔａｔｅに関する時間発展の記述を有し、ノード種別ＮＫがＳｔａｔｅのノード５のみを子ノードとして接続することができる。 Time Evolution, which is an example of the node type NK, has a description of time evolution regarding the State connected as a parent node or a child node, and only the node 5 whose node type NK is a State can be connected as a child node.
　ノード種別ＮＫの一例であるＬｉｋｅｌｉｈｏｏｄ　＆　Ｐｒｉｏｒは、尤度や事前確率等の確率モデルの記述を有し、子ノードとして接続することができるノード５がない。 Likelihood & Prior, which is an example of node type NK, has a description of a probability model such as likelihood and prior probability, and there is no node 5 that can be connected as a child node.
　ノード種別ＮＫの一例であるＧｅｎｅｒａｔｅｄ　Ｑｕａｎｔｉｔｙは、データ型と親ノード変数の変換関数とを有し、ノード種別ＮＫがＧｅｎｅｒａｔｅｄ　Ｑｕａｎｔｉｔｙのノード５のみを子ノードとして接続することができる。 Generated Quantity, which is an example of node type NK, has a data type and a conversion function of a parent node variable, and only node 5 whose node type NK is Generated Quantity can be connected as a child node.
　このように、少なくとも一部に確率的プログラミング言語の一例であるＳｔａｎにおいて規定されたノード種別ＮＫを採用することで、Ｓｔａｎの知識がある者であれば、より直感的に所望する推論モデルをＧＵＩの一例である操作画面４（図９～図１１参照）上で構築することできるため、ユーザフレンドリーである。操作画面４については次節でさらに詳しく説明する。 In this way, by adopting the node type NK specified in Tan, which is an example of a stochastic programming language, at least in part, a person who has knowledge of Santa can more intuitively obtain a desired inference model in the GUI. Since it can be constructed on the operation screen 4 (see FIGS. 9 to 11), which is an example, it is user-friendly. The operation screen 4 will be described in more detail in the next section.
４．情報処理方法
　本節では、前述した情報処理システム１の情報処理方法について説明する。 4. Information Processing Method This section describes the information processing method of the information processing system 1 described above.
　好ましくは、この情報処理方法は、次の各ステップを備える。受付ステップでは、推論モデルを構築するためのＧＵＩに対するユーザの入力を受け付ける。入力は、推論モデルを構成する複数のノード５それぞれに対する属性情報ＩＦと、複数のノード５間の接続関係ＣＲとを含む。第１の生成ステップでは、入力に基づいて、属性情報ＩＦの少なくとも一部と、接続関係ＣＲとが把握可能な態様で、複数のノード５をＧＵＩ上に含む視覚情報を生成する。 Preferably, this information processing method includes the following steps. In the reception step, the user's input to the GUI for constructing the inference model is received. The input includes an attribute information IF for each of the plurality of nodes 5 constituting the inference model, and a connection relationship CR between the plurality of nodes 5. In the first generation step, based on the input, visual information including a plurality of nodes 5 on the GUI is generated in such a manner that at least a part of the attribute information IF and the connection relationship CR can be grasped.
４．１　ＳａａＳ型モデル
　図６は、ＳａａＳ型モデルによるサービス提供の流れを示すアクティビティ図である。本実施形態では、ユーザは、所定の事象に関する予測を行うために推論モデルの構築を所望するデータサイエンティストであるものとして説明する。なお、あくまでも一例でありユーザはこの限りではない。以下、情報処理システム１によって提供される推論モデルの構築サービスは、ユーザ端末２からインターネットを介してサーバ３にアクセスすることによって提供される、いわゆるＳａａＳ（Ｓｏｆｔｗａｒｅ　ａｓ　ａ　Ｓｅｒｖｉｃｅ）であるものとして、図６に示されるアクティビティ図に沿って説明する。 4.1 SaaS type model FIG. 6 is an activity diagram showing a flow of service provision by the SaaS type model. In the present embodiment, the user is described as being a data scientist who desires to build an inference model in order to make a prediction about a predetermined event. It should be noted that this is just an example, and the user is not limited to this. Hereinafter, the inference model construction service provided by the information processing system 1 is assumed to be so-called SaaS (Software as a Service) provided by accessing the server 3 from the user terminal 2 via the Internet. This will be described with reference to the activity diagram shown in 6.
　図７は、ログイン画面９の一例を示す概要図である。ユーザが情報処理システム１によって提案されるサービスのＵＲＬにアクセスすると、サーバ３において、生成部３３３がログイン画面９を生成し（アクティビティＡ１０１）、出力部３３４がこれをネットワークを介してユーザ端末２に送信することで、ユーザ端末２の表示部２４にログイン画面９が表示される。つまり、サーバ３における通信部３１がユーザ端末２から送信されたＵＲＬのリクエストを受信し、サーバ３における制御部３３が、リクエストに基づいてログイン画面９を生成し、これが通信部３１及びネットワークを介して、ユーザ端末２に送信される。 FIG. 7 is a schematic diagram showing an example of the login screen 9. When the user accesses the URL of the service proposed by the information processing system 1, the generation unit 333 generates the login screen 9 (activity A101) on the server 3, and the output unit 334 transfers this to the user terminal 2 via the network. By transmitting, the login screen 9 is displayed on the display unit 24 of the user terminal 2. That is, the communication unit 31 in the server 3 receives the request for the URL transmitted from the user terminal 2, and the control unit 33 in the server 3 generates a login screen 9 based on the request, which is via the communication unit 31 and the network. Is transmitted to the user terminal 2.
　ログイン画面９は、例えば、ユーザ名を入力する入力欄９１と、パスワードを入力する入力欄９２と、ログイン時に押下するログインボタン９３と、登録時に選択する登録リンク９４とを備えている。 The login screen 9 includes, for example, an input field 91 for inputting a user name, an input field 92 for inputting a password, a login button 93 to be pressed at login, and a registration link 94 to be selected at registration.
　ユーザがアカウントを有していない場合は、登録リンク９４をクリックやタップする等して選択すればよい。登録リンク９４を選択した場合、ユーザ端末２の表示部２４にアカウント登録用の画面（不図示）が表示され、ユーザが必要に応じて、ユーザの初期情報を入力すればよい（アクティビティＡ１０２）。 If the user does not have an account, select it by clicking or tapping the registration link 94. When the registration link 94 is selected, an account registration screen (not shown) is displayed on the display unit 24 of the user terminal 2, and the user may input the initial information of the user as necessary (activity A102).
　アカウント登録に際して、ユーザのアカウントに関する初期情報の入力が完了されると、ネットワークを介してサーバ３にその情報が送信され、サーバ３における記憶部３２にユーザのアカウントに関する情報、例えばユーザ名やパスワード等がデータベースとして記憶される（アクティビティＡ１０３）。 When the input of the initial information about the user's account is completed at the time of account registration, the information is transmitted to the server 3 via the network, and the information about the user's account, such as a user name and a password, is sent to the storage unit 32 in the server 3. Is stored as a database (activity A103).
　アカウント登録が完了すると、ユーザは、情報処理システム１が提供する推論モデル構築サービスにログイン可能となる。換言すると、ユーザが、情報処理システム１によって提案されるサービスに対して、アカウントを有している場合は、入力欄９１にユーザ固有のユーザ名を入力し、入力欄９２にユーザ名に紐づくパスワードを入力し、ログインボタン９３をクリックやタップする等して押下すれば、推論モデル構築サービスにログインすることができる（アクティビティＡ１０４）。 When the account registration is completed, the user can log in to the inference model construction service provided by the information processing system 1. In other words, if the user has an account for the service proposed by the information processing system 1, the user-specific user name is entered in the input field 91 and linked to the user name in the input field 92. If you enter the password and press the login button 93 by clicking or tapping it, you can log in to the inference model construction service (Activity A104).
　続いて、サーバ３における生成部３３３は、ＧＵＩの一例である操作画面４を生成する（アクティビティＡ１０５）。すなわち、サーバ３における記憶部３２にユーザのアカウントに関する情報が記憶された後に、サーバ３における制御部３３が、記憶部３２に記憶された所定のプログラムを実行することで、操作画面４を生成している。出力部３３４は、操作画面４を通信部３１及びネットワークを介して、ユーザ端末２に送信することで、操作画面４がユーザ端末２の表示部２４に表示される。 Subsequently, the generation unit 333 in the server 3 generates the operation screen 4 which is an example of the GUI (activity A105). That is, after the information about the user's account is stored in the storage unit 32 of the server 3, the control unit 33 of the server 3 executes a predetermined program stored in the storage unit 32 to generate the operation screen 4. ing. The output unit 334 transmits the operation screen 4 to the user terminal 2 via the communication unit 31 and the network, so that the operation screen 4 is displayed on the display unit 24 of the user terminal 2.
　ユーザは、表示部２４に表示された操作画面４を入力部２５で操作することで、推論モデルを構築することとなる（アクティビティＡ１０６）。操作画面４の具体的な操作の流れについては、図８～図１２を参照して後述する。このようにＳａａＳ型モデルを採用することで、特別な環境構築を必要とせず広くユーザに使用する環境を提供することができる。 The user constructs an inference model by operating the operation screen 4 displayed on the display unit 24 on the input unit 25 (activity A106). The specific operation flow of the operation screen 4 will be described later with reference to FIGS. 8 to 12. By adopting the SaaS model in this way, it is possible to provide an environment that can be widely used by users without requiring a special environment construction.
４．２　操作画面４の操作
　図８は、操作画面４の操作の流れを示すアクティビティ図である。図９～図１１は、操作画面４の一例を示す概要図である。以下、図８に示されるアクティビティ図に沿って、操作画面４の操作の流れを説明する。 4.2 Operation of the operation screen 4 FIG. 8 is an activity diagram showing an operation flow of the operation screen 4. 9 to 11 are schematic views showing an example of the operation screen 4. Hereinafter, the operation flow of the operation screen 4 will be described with reference to the activity diagram shown in FIG.
　まず、生成部３３３が操作画面４を生成し、出力部３３４がこれを送信することで、ユーザ端末２の表示部２４に操作画面４が表示される（アクティビティＡ２０１）。これに対して、ユーザが操作を行うと、受付ステップとして、受付部３３１は、推論モデルを構築するための操作画面４に対するユーザの入力を受け付ける。換言すると、サーバ３が通信部３１を介して受信した入力が記憶部３２の一例である一時記憶領域に記憶され、種々の処理が制御部３３によって実行される。また、入力は、推論モデルを構成する複数のノード５それぞれに対する属性情報ＩＦと、複数のノード５間の接続関係ＣＲの一例であるリンク６とを含む。 First, the generation unit 333 generates the operation screen 4, and the output unit 334 transmits this, so that the operation screen 4 is displayed on the display unit 24 of the user terminal 2 (activity A201). On the other hand, when the user performs an operation, as a reception step, the reception unit 331 accepts the user's input to the operation screen 4 for constructing the inference model. In other words, the input received by the server 3 via the communication unit 31 is stored in the temporary storage area, which is an example of the storage unit 32, and various processes are executed by the control unit 33. Further, the input includes an attribute information IF for each of the plurality of nodes 5 constituting the inference model, and a link 6 which is an example of the connection relationship CR between the plurality of nodes 5.
具体的には、ユーザは、操作画面４に対して、属性情報ＩＦの一例であるノード種別ＮＫと、ノード名とを入力することでノード５を設定する（アクティビティＡ２０２）。続いて、生成部３３３が操作画面４を更新して、出力部３３４がこれを送信することで、更新後の操作画面４がユーザ端末２の表示部２４に表示される。すなわち、設定されたノード５が操作画面４に表示される（アクティビティＡ２０３）。続いて、ユーザは、操作画面４に対して、設定したノード５に対応する属性情報ＩＦの一例である第１のコードを入力する（アクティビティＡ２０４）。これらを複数回繰り返した後、ユーザは、操作画面４に対して、ノード５間の接続関係ＣＲを示すリンク６を入力する（アクティビティＡ２０５）。 Specifically, the user sets the node 5 by inputting the node type NK, which is an example of the attribute information IF, and the node name on the operation screen 4 (activity A202). Subsequently, the generation unit 333 updates the operation screen 4, and the output unit 334 transmits this, so that the updated operation screen 4 is displayed on the display unit 24 of the user terminal 2. That is, the set node 5 is displayed on the operation screen 4 (activity A203). Subsequently, the user inputs the first code, which is an example of the attribute information IF corresponding to the set node 5, to the operation screen 4 (activity A204). After repeating these a plurality of times, the user inputs the link 6 indicating the connection relationship CR between the nodes 5 to the operation screen 4 (activity A205).
　換言すると、属性情報ＩＦの少なくとも一部は、ノード種別ＮＫであり、ノード種別ＮＫによってノード５どうしの接続関係ＣＲが制限される。ノード種別ＮＫ及びその接続関係ＣＲの詳細については、前節を参照されたい。このような態様によれば、ユーザに誤ったノード５どうしを接続させ、これによってバグを有する第１のコード又は第２のコードが生成されるのを防止することができる。第２のコードについては後述する。特に、単に制限されるだけでなく、ユーザに接続することができない旨の警告が出力されてもよい。これについては、後に説明する。 In other words, at least a part of the attribute information IF is the node type NK, and the connection relation CR between the nodes 5 is restricted by the node type NK. For details of node type NK and its connection relation CR, refer to the previous section. According to such an aspect, it is possible to prevent the user from connecting the erroneous nodes 5 to each other and thereby generating a first code or a second code having a bug. The second code will be described later. In particular, not only is it restricted, but a warning that the user cannot be connected may be output. This will be described later.
　また、図９に示されるように、初期状態等、ノード５が非選択の状態において、ＧＵＩの一例である操作画面４は、ノード種別ＮＫを予め定められた項目から選択可能なオブジェクトの一例である選択欄４１を含む。選択欄４１は、例えばプルダウンリストによって実施される。そして、サーバ３において受付部３３１は、ユーザによるノード種別ＮＫの選択を入力として受け付けることができる。このように、予め定められた項目からユーザにノード種別ＮＫを選択させることで、ノード種別ＮＫの正確な名前やスペルを覚えていない者であっても、効率的にノード種別ＮＫを決定することができるため、ユーザフレンドリーである。 Further, as shown in FIG. 9, in a state where the node 5 is not selected, such as in the initial state, the operation screen 4 which is an example of the GUI is an example of an object in which the node type NK can be selected from predetermined items. Includes a selection field 41. The selection field 41 is implemented, for example, by a pull-down list. Then, in the server 3, the reception unit 331 can accept the selection of the node type NK by the user as an input. In this way, by letting the user select the node type NK from the predetermined items, even a person who does not remember the exact name or spelling of the node type NK can efficiently determine the node type NK. It is user-friendly because it can be used.
　操作画面４は、さらに、入力欄４２と、ボタン４３と、領域４４と、ボタン４６とを備えている。入力欄４２は、ユーザにノード５の名称（ノード名）を入力させる欄である。ボタン４３は、ノード５を生成するためのオブジェクトである。ユーザが選択欄４１の選択及び入力欄４２への入力を済ませ、ボタン４３を押下することで、領域４４内にノード５が生成される。図１０に示される例では、選択欄４１ではノード種別ＮＫの一例であるＤａｔａが選択され、入力欄４２にはノード名であるＰａｔｉｅｎｔが入力され、ユーザがボタン４３を押下することで、領域４４内にノード５１が生成されている。ボタン４６については後述する。 The operation screen 4 further includes an input field 42, a button 43, an area 44, and a button 46. The input field 42 is a field for allowing the user to input the name (node name) of the node 5. The button 43 is an object for creating the node 5. When the user completes the selection of the selection field 41 and the input to the input field 42 and presses the button 43, the node 5 is generated in the area 44. In the example shown in FIG. 10, Data, which is an example of node type NK, is selected in the selection field 41, Patient, which is a node name, is input in the input field 42, and the user presses the button 43 to display the area 44. The node 51 is generated in the inside. The button 46 will be described later.
　さらに、図１１に示されるように、ユーザが、領域４４に表示されたノード５を選択すると、入力欄４５が表示されることが好ましい。図１１では、ユーザがノード５３を選択している。ノード５３は、ノード種別ＮＫがＴｉｍｅ　Ｅｖｏｌｕｔｉｏｎであり且つノード名がＯＤＥである。また、先に説明したノード５１と、ノード種別ＮＫがＳｔａｔｅであり且つノード名がＳｕｓｃｅｐｔｉｂｌｅであるノード５２とは、ユーザに選択されていないノードである。入力欄４５には、ノード種別ＮＫ以外の属性情報ＩＦが入力されることが好ましい。特に、入力欄４５に入力されうる属性情報ＩＦは、対応するノード５における情報処理を表す第１のコードを含む。 Further, as shown in FIG. 11, it is preferable that the input field 45 is displayed when the user selects the node 5 displayed in the area 44. In FIG. 11, the user has selected node 53. The node 53 has a node type NK of Time Evolution and a node name of ODE. Further, the node 51 described above and the node 52 whose node type NK is State and whose node name is Possible are nodes that have not been selected by the user. It is preferable that the attribute information IF other than the node type NK is input to the input field 45. In particular, the attribute information IF that can be input to the input field 45 includes a first code representing information processing in the corresponding node 5.
　具体的には例えば、対応するノード５において使用するローカル変数の宣言、宣言したローカル変数への数値代入、宣言したローカル変数を用いた情報処理等が、第１のコードに含まれうる。このように、各ノードに対して、属性情報ＩＦの１つとして、コードの入力を可能とすることで、効率を重視しつつも、推論モデルのより汎用的な設計を実現することができる。 Specifically, for example, the declaration of a local variable used in the corresponding node 5, the numerical assignment to the declared local variable, the information processing using the declared local variable, and the like can be included in the first code. In this way, by making it possible to input a code as one of the attribute information IFs for each node, it is possible to realize a more general-purpose design of the inference model while emphasizing efficiency.
　また好ましくは、第１のコードは、最終的に出力部３３４が出力する全体のソースコード７（図１２参照）である第２のコードの少なくとも一部を構成する。第１のコードが、第２のコードの少なくとも一部を構成することで、特に将来的に情報処理システム１を用いずにコードを実装したいユーザにとって、良い教育材料となりうる。 Also preferably, the first code constitutes at least a part of the second code, which is the entire source code 7 (see FIG. 12) finally output by the output unit 334. By forming at least a part of the second code, the first code can be a good educational material, especially for a user who wants to implement the code without using the information processing system 1 in the future.
　また、図１１に示されるように、ユーザは、ノード５間の接続関係ＣＲを設定することができる。具体的には、ノード５間が矢印型のリンク６で接続されており、ノード５間の上流下流の関係が把握可能な構成となっている。なお、リンク６は線状の視覚情報であるため、より扱いやすく位置や向きを微調整するために、操作アイコン６ａを備えていることが好ましい。 Further, as shown in FIG. 11, the user can set the connection relationship CR between the nodes 5. Specifically, the nodes 5 are connected by an arrow-shaped link 6, and the upstream-downstream relationship between the nodes 5 can be grasped. Since the link 6 is linear visual information, it is preferable to provide the operation icon 6a in order to make it easier to handle and finely adjust the position and orientation.
　より具体的には、サーバ３において、受付部３３１が複数のノード５のうちを順番に２つ選択した入力を受け付ける。そして、生成部３３３が順番のうち先に選択された第１のノード、例えばノード５１を上流側、後に選択された第２のノード、例えばノード５２を下流側として、上流側及び下流側を把握可能な態様で示す視覚情報を生成する。例えば矢印の向きによって把握可能に、ユーザ端末２の表示部２４に視覚情報が表示されるとよい。図１１に示される例では、リンク６１は、上流側であるノード５１と下流側であるノード５２との接続関係ＣＲを示している。同様に、リンク６２は、上流側であるノード５２と下流側であるノード５３との接続関係ＣＲを示している。このような態様によれば、ユーザは、簡単な操作で、ノード５の上流下流を決定することができるとともに、これをＧＵＩの一例である操作画面４上で把握することができるため、ユーザフレンドリーである。 More specifically, in the server 3, the reception unit 331 accepts an input in which two of the plurality of nodes 5 are selected in order. Then, the generation unit 333 grasps the upstream side and the downstream side by setting the first node selected first in the order, for example, the node 51 as the upstream side, and the second node selected later, for example, the node 52 as the downstream side. Generates visual information that is shown in a possible manner. For example, it is preferable that the visual information is displayed on the display unit 24 of the user terminal 2 so that it can be grasped by the direction of the arrow. In the example shown in FIG. 11, the link 61 shows the connection relationship CR between the node 51 on the upstream side and the node 52 on the downstream side. Similarly, the link 62 shows the connection relationship CR between the node 52 on the upstream side and the node 53 on the downstream side. According to such an aspect, the user can determine the upstream / downstream of the node 5 with a simple operation, and can grasp this on the operation screen 4 which is an example of the GUI, so that it is user-friendly. Is.
　また、前述の通り、ノード種別ＮＫによって接続可能なノード５が制限されているため、ユーザにこれを教示可能に構成されることが好ましい。換言すると、出力部３３４は、接続不能なノード５の接続関係ＣＲが入力された場合に、警告を出力するように構成されるとよい。例えば、ユーザによって、２つのノード５が順番に選択されることで接続関係ＣＲが入力されると、読出部３３２は、予め記憶部３２に記憶された接続是非情報を読み出す。接続是非情報には、第３節で説明した各ノード種別ＮＫの接続可能なノード種別ＮＫがルックアップテーブル等の形式でまとめられているとよい。接続可能である場合は、生成部３３３が図１１に示されるようなリンク６を操作画面４に含む視覚情報を生成し、出力部３３４がこれを送信することで、この視覚情報がユーザ端末２における表示部２４に表示され（アクティビティＡ２０６）、接続不可である場合は、ユーザにこれを知らせる不図示の警告が提示されるとよい（アクティビティＡ２０７）。かかる警告は、アラートアイコン等の視覚情報でもよいし、警告音等の音声情報でもよい。 Further, as described above, since the node 5 that can be connected is limited by the node type NK, it is preferable that the node 5 can be taught to the user. In other words, the output unit 334 may be configured to output a warning when the connection-related CR of the unconnectable node 5 is input. For example, when the connection relationship CR is input by the user selecting the two nodes 5 in order, the reading unit 332 reads out the connection pros and cons information previously stored in the storage unit 32. In the connection pros and cons information, it is preferable that the connectable node type NK of each node type NK described in Section 3 is summarized in the form of a look-up table or the like. When the connection is possible, the generation unit 333 generates visual information including the link 6 as shown in FIG. 11 in the operation screen 4, and the output unit 334 transmits this, so that this visual information is transmitted to the user terminal 2. (Activity A206), if the connection is not possible, a warning (activity A207) not shown may be presented to the user. The warning may be visual information such as an alert icon or voice information such as a warning sound.
　さらに、ノード種別ＮＫに関連し、生成部３３３がノード５を示す枠の形態でノード種別ＮＫを区別可能な視覚情報を生成することが好ましい。ここでの形態は、形状、模様若しくは色彩又はこれらの結合を含む概念である。具体的には、ノード５１と、ノード５２と、ノード５３とは枠内のピクセル情報が異なるものである。すなわち、色相、明度及び彩度の少なくとも何れかが異なることで、ユーザが区別可能に構成されている。あるいは、ピクセル情報に代えて、枠の形状そのものを区別可能としてもよい。枠の形状としては、例えば、長方形、正方形、円及び楕円等が採用されうる。このようにノード５を示す枠の形態によって、ユーザが直感的にノード種別ＮＫを把握できるため、ユーザフレンドリーである、 Further, it is preferable that the generation unit 333 generates visual information that can distinguish the node type NK in the form of a frame indicating the node 5 in relation to the node type NK. The form here is a concept including a shape, a pattern or a color, or a combination thereof. Specifically, the node 51, the node 52, and the node 53 have different pixel information in the frame. That is, it is configured so that the user can distinguish it by having at least one of hue, lightness, and saturation different. Alternatively, instead of the pixel information, the shape of the frame itself may be distinguishable. As the shape of the frame, for example, a rectangle, a square, a circle, an ellipse, or the like can be adopted. In this way, the form of the frame showing the node 5 allows the user to intuitively grasp the node type NK, which is user-friendly.
　なお、図８に示されるように、ノード５の設定操作（アクティビティＡ２０２～Ａ２０４）と、接続関係ＣＲの設定操作（アクティビティＡ２０５～Ａ２０７）とを自由に繰り返してよい。例えば、ノード５１とノード５２とを設定してこれらを接続する接続関係ＣＲとなるリンク６１を設定し、続いてノード５３を設定してノード５２と接続する接続関係ＣＲであるリンク６２を設定してもよい。このような一連の操作により、第１の生成ステップとして、生成部３３３は、入力に基づいて、属性情報ＩＦの少なくとも一部と、接続関係ＣＲとが把握可能な態様で、複数のノード５を操作画面４上に含む視覚情報を生成する。 As shown in FIG. 8, the node 5 setting operation (activity A202 to A204) and the connection-related CR setting operation (activity A205 to A207) may be freely repeated. For example, the node 51 and the node 52 are set to set the link 61 which is the connection relation CR connecting them, and then the node 53 is set and the link 62 which is the connection relation CR connecting to the node 52 is set. You may. By such a series of operations, as a first generation step, the generation unit 333 generates a plurality of nodes 5 in a manner in which at least a part of the attribute information IF and the connection relationship CR can be grasped based on the input. Generate the visual information included on the operation screen 4.
　所望する推論モデルが構築されると、ユーザは、ボタン４６を押下する（アクティビティＡ２０８）。これを契機に、第２の出力ステップとして、サーバ３における出力部３３４は、操作画面４の領域４４において表示された態様及び属性情報ＩＦに対応する、推論モデルを示す第２のコードを出力する（アクティビティＡ２０９）。図１２は、出力部３３４が出力した第２のコードの一例であるソースコード７を示している。出力されたソースコード７は、出力部３３４によってユーザ端末２に送信され、表示部２４に表示されてもよいし、データとしてダウンロードされるようにしてもよい。 When the desired inference model is constructed, the user presses the button 46 (activity A208). Taking this as an opportunity, as a second output step, the output unit 334 in the server 3 outputs a second code indicating an inference model corresponding to the mode and attribute information IF displayed in the area 44 of the operation screen 4. (Activity A209). FIG. 12 shows a source code 7 which is an example of the second code output by the output unit 334. The output source code 7 may be transmitted to the user terminal 2 by the output unit 334 and displayed on the display unit 24, or may be downloaded as data.
　このような態様によれば、構築した推論モデルを動作させるためのコードを汎用的に使用可能なソースコード７（第２のコード）として出力することができる。このように生成された第２のコードを、ユーザがユーザ端末２を用いてローカル環境下で適宜コンパイルすることで推論モデルを生成し、これにより所望の事象に関する推論を実施することができる。あるいは、本実施形態に係る情報処理システム１を利用しない者にソースコードを譲渡する等することが可能となる。 According to such an aspect, the code for operating the constructed inference model can be output as a source code 7 (second code) that can be used for general purposes. The user can appropriately compile the second code generated in this way in the local environment using the user terminal 2 to generate an inference model, whereby inference about a desired event can be performed. Alternatively, the source code can be transferred to a person who does not use the information processing system 1 according to the present embodiment.
　なお、ボタン４６の押下を契機として初めて第２のコードが生成されてもよいが、さらに好ましくは、推論モデルの構築段階において、領域４４に表示されているノード５等に基づいて、第２のコードが逐次生成又は更新されるとよい。 The second code may be generated for the first time when the button 46 is pressed, but more preferably, the second code is generated based on the node 5 or the like displayed in the region 44 at the stage of constructing the inference model. Code should be generated or updated sequentially.
　以上に説明した実施形態によれば、変数や関数間の依存関係を把握しながら、誤りを削減してより容易かつ効率的に推論モデルを構築することができる。また、このような実施形態によれば、プログラミング技術の個人差を緩和し、より容易かつ効率的に推論モデルを構築することができる。 According to the embodiment described above, it is possible to reduce errors and construct an inference model more easily and efficiently while grasping the dependency relationships between variables and functions. Further, according to such an embodiment, it is possible to alleviate individual differences in programming techniques and construct an inference model more easily and efficiently.
５．その他
　本実施形態に係る情報処理システム１に関して、以下のような態様を採用してもよい。 5. In addition, the following aspects may be adopted for the information processing system 1 according to the present embodiment.
（１）以上の実施形態では、情報処理システム１の構成として説明したが、コンピュータに、コンピュータに情報処理システム１の各ステップを実行させるプログラムが提供されてもよい。あるいは、情報処理システム１の各ステップを備える情報処理方法が提供されてもよい。 (1) In the above embodiments, although the configuration of the information processing system 1 has been described, the computer may be provided with a program for causing the computer to execute each step of the information processing system 1. Alternatively, an information processing method including each step of the information processing system 1 may be provided.
（２）情報処理システム１は、不図示の実行部をさらに備えてもよい。すなわち、ユーザ端末２でのコンパイルを不要とし、サーバ３における実行部が、第２のコードを実行可能に構成されてもよい。さらに好ましくは、第２の生成ステップとして、生成部３３３が、第２のコードの実行結果に基づいて、設定された事象に関する、推論モデルによる推定結果８を含む視覚情報を生成してもよい。図１３は、推論モデルによって得られた推定結果８の一例である。このような態様によれば、ユーザは、ノード５の構築から推論モデルの実行までをワンストップで行うことができるため、より迅速に効率的に、推論モデルを用いた様々な事象の予測を行うことができる。 (2) The information processing system 1 may further include an execution unit (not shown). That is, compilation on the user terminal 2 is not required, and the execution unit on the server 3 may be configured to be able to execute the second code. More preferably, as the second generation step, the generation unit 333 may generate visual information including the estimation result 8 by the inference model regarding the set event based on the execution result of the second code. FIG. 13 is an example of the estimation result 8 obtained by the inference model. According to such an embodiment, the user can perform one-stop operations from the construction of the node 5 to the execution of the inference model, so that various events can be predicted using the inference model more quickly and efficiently. be able to.
（３）本実施形態においては、ノード５どうしが接続できないノード種別ＮＫである場合に警告を出すように実施しているが、警告を出すことに代えて、接続不能なノード５の接続関係ＣＲが入力された場合に、生成部３３３がこの接続不能な接続関係ＣＲを含む視覚情報を生成しないように構成されてもよい。 (3) In the present embodiment, a warning is issued when the nodes 5 cannot be connected to each other with the node type NK, but instead of issuing the warning, the connection relationship CR of the unreachable node 5 is issued. Is input, the generation unit 333 may be configured not to generate visual information including this unconnectable connection relationship CR.
（４）１つ上流のノード５である親ノードまたは１つ下流のノード５である子ノードに規定される変数を参照していない場合に、警告を出すように実施してもよい。つまり、第３の出力ステップとして、出力部３３４は、所定のノード５、例えばノード５２において、ノード５２の上流側にある親ノードであるノード５１に規定される変数を引用していない場合に、警告を出力するように構成されてもよい。また好ましくは、出力部３３４は、所定のノード５において、所定のノード５の下流側にある子ノードに規定される変数を引用していない場合に、警告を出力するように構成されてもよい。このような態様によれば、変数の引用漏れを防止することができる。 (4) When the variable specified in the parent node which is one upstream node 5 or the child node which is one downstream node 5 is not referred to, a warning may be issued. That is, as a third output step, when the output unit 334 does not quote the variable specified in the node 51 which is the parent node on the upstream side of the node 52 in the predetermined node 5, for example, the node 52. It may be configured to output a warning. Further, preferably, the output unit 334 may be configured to output a warning when the variable specified in the child node on the downstream side of the predetermined node 5 is not quoted in the predetermined node 5. .. According to such an aspect, it is possible to prevent omission of citation of variables.
（５）ユーザによってあるノード５が選択された場合に、ワンクリックで親ノードまたは子ノードに規定される変数を参照可能な不図示のオブジェクトが表示されるように実施してもよい。つまり、ＧＵＩの一例である操作画面４は、所定のノード５、例えばノード５２の上流側にある親ノードであるノード５１に規定される変数を参照可能なオブジェクトを含んでもよい。また好ましくは、ＧＵＩの一例である操作画面４は、所定のノード５の下流側にある子ノードに規定される変数を参照可能なオブジェクトを含んでもよい。このような態様によれば、変数の引用漏れに気づいた後に、素早く修正することができる。 (5) When a certain node 5 is selected by the user, an object (not shown) that can refer to the variables specified in the parent node or the child node may be displayed with one click. That is, the operation screen 4 which is an example of the GUI may include an object that can refer to a variable defined in a predetermined node 5, for example, a node 51 which is a parent node on the upstream side of the node 52. Further, preferably, the operation screen 4 which is an example of the GUI may include an object which can refer to the variable specified in the child node on the downstream side of the predetermined node 5. According to such an embodiment, after noticing a variable citation omission, it can be quickly corrected.
（６）本実施形態をＳａａＳ型モデルで提供されるサービスとして説明したが、これに限定されず、ユーザがユーザ端末２に所定のプログラムをインストールすることで、スタンドアロン型のサービスとして実施してもよい。このようなプログラムが頒布されてもよい。 (6) Although the present embodiment has been described as a service provided by the SaaS model, the present invention is not limited to this, and even if the user installs a predetermined program on the user terminal 2, the service is implemented as a stand-alone service. good. Such programs may be distributed.
（７）どのノード５とも接続されていない浮いた状態のノード５が存在する場合に、出力部３３４が警告を出すように実施してもよい。このような態様によれば、無駄なノード５の作成を防止することや、リンク漏れを防止することができる。 (7) When there is a floating node 5 that is not connected to any node 5, the output unit 334 may issue a warning. According to such an aspect, it is possible to prevent unnecessary creation of the node 5 and prevent link leakage.
（８）図１４は、操作画面４の別の態様を示している。すなわち、このような態様によって操作画面４を実施してもよい。 (8) FIG. 14 shows another aspect of the operation screen 4. That is, the operation screen 4 may be implemented according to such an aspect.
　さらに、次に記載の各態様で提供されてもよい。
　前記情報処理システムにおいて、前記属性情報の少なくとも一部は、ノード種別であり、前記ノード種別によって前記ノード間の前記接続関係が制限される、もの。
　前記情報処理システムにおいて、第１の出力ステップをさらに実行するように構成され、前記第１の出力ステップでは、接続不能な前記ノードの接続関係が入力された場合に、警告を出力するように構成される、もの。
　前記情報処理システムにおいて、前記第１の生成ステップでは、接続不能な前記ノードの接続関係が入力された場合に、この接続不能な前記接続関係を含む視覚情報を生成しないように構成される、もの。と
　前記情報処理システムにおいて、前記第１の生成ステップでは、前記ノードを示す枠の形態で前記ノード種別を区別可能な前記視覚情報を生成し、ここで前記形態は、形状、模様若しくは色彩又はこれらの結合を含む、もの。
　前記情報処理システムにおいて、前記ＧＵＩは、前記ノード種別を予め定められた項目から選択可能なオブジェクトを含み、前記受付ステップでは、前記ユーザによる前記ノード種別の選択を前記入力として受け付ける、もの。
　前記情報処理システムにおいて、前記ノード種別は、変数、関数及び確率モデルを含む、もの。
　前記情報処理システムにおいて、前記ノード種別は、プログラミング言語であるＳｔａｎにおいて規定された、Ｄａｔａ、Ｔｒａｎｓｆｏｒｍｅｄ　Ｄａｔａ、Ｐａｒａｍｅｔｅｒ、Ｔｒａｎｓｆｏｒｍｅｄ　Ｐａｒａｍｅｔｅｒ、Ｌｉｋｅｌｉｈｏｏｄ　＆　Ｐｒｉｏｒ及びＧｅｎｅｒａｔｅｄ　Ｑｕａｎｔｉｔｙのうちの少なくとも１つを含む、もの。
　前記情報処理システムにおいて、前記属性情報は、対応する前記ノードにおける情報処理を表す第１のコードを含む、もの。
　前記情報処理システムにおいて、第２の出力ステップをさらに実行するように構成され、前記第２の出力ステップでは、表示された前記態様及び前記属性情報に対応する、前記推論モデルを示す第２のコードを出力する、もの。
　前記情報処理システムにおいて、実行ステップをさらに実行するように構成され、前記実行ステップでは、前記第２のコードを実行する、もの。
　前記情報処理システムにおいて、第２の生成ステップをさらに実行するように構成され、前記第２の生成ステップでは、前記第２のコードの実行結果に基づいて、設定された事象に関する、前記推論モデルによる推定結果を含む第２の視覚情報を生成する、もの。
　前記情報処理システムにおいて、前記属性情報は、対応する前記ノードにおける情報処理を表す第１のコードを含み、前記第１のコードは、前記第２のコードの少なくとも一部を構成する、もの。
　前記情報処理システムにおいて、前記受付ステップでは、前記複数のノードのうちを順番に２つ選択した前記入力を受け付け、前記第１の生成ステップでは、前記順番のうち先に選択された第１のノードを上流側、後に選択された第２のノードを下流側として、前記上流側及び前記下流側を把握可能な態様で示す視覚情報を生成する、もの。
　前記情報処理システムにおいて、前記ＧＵＩは、所定のノードの上流側にある親ノード又は下流側にある子ノードに規定される変数を参照可能なオブジェクトを含む、もの。
　情報処理方法であって、前記情報処理システムの各ステップを備える、方法。
　プログラムであって、コンピュータに前記情報処理システムの各ステップを実行させる、もの。
　もちろん、この限りではない。 Furthermore, it may be provided in each of the following embodiments.
In the information processing system, at least a part of the attribute information is a node type, and the connection relationship between the nodes is limited by the node type.
The information processing system is configured to further execute a first output step, and the first output step is configured to output a warning when a connection relationship of the unconnectable node is input. Things to be done.
In the information processing system, in the first generation step, when the connection relationship of the unconnectable node is input, the visual information including the unconnectable connection relationship is not generated. .. In the information processing system, in the first generation step, the visual information that can distinguish the node type is generated in the form of a frame indicating the node, wherein the form is a shape, a pattern, a color, or these. Things that include the binding of.
In the information processing system, the GUI includes an object in which the node type can be selected from predetermined items, and in the reception step, the selection of the node type by the user is received as the input.
In the information processing system, the node type includes a variable, a function, and a probability model.
In the information processing system, the node type includes at least one of Data, Translated Data, Parameter, Transformed Parameter, Likelihood & Prior and Generated Quantity defined in the programming language Stan.
In the information processing system, the attribute information includes a first code representing information processing in the corresponding node.
In the information processing system, a second code indicating the inference model, which is configured to further execute a second output step, which corresponds to the displayed embodiment and the attribute information. To output, stuff.
In the information processing system, the execution step is configured to be further executed, and the execution step executes the second code.
In the information processing system, the second generation step is further executed, and the second generation step is based on the inference model regarding the set event based on the execution result of the second code. A device that produces a second visual piece of information, including estimation results.
In the information processing system, the attribute information includes a first code representing information processing in the corresponding node, and the first code constitutes at least a part of the second code.
In the information processing system, in the reception step, the input in which two of the plurality of nodes are selected in order is received, and in the first generation step, the first node selected first in the order is received. Is the upstream side, and the second node selected later is the downstream side, and the visual information showing the upstream side and the downstream side in a graspable manner is generated.
In the information processing system, the GUI includes an object that can refer to a variable defined by a parent node on the upstream side or a child node on the downstream side of a predetermined node.
An information processing method, comprising each step of the information processing system.
A program that causes a computer to execute each step of the information processing system.
Of course, this is not the case.
　最後に、本発明に係る種々の実施形態を説明したが、これらは、例として提示したものであり、発明の範囲を限定することは意図していない。当該新規な実施形態は、その他の様々な形態で実施されることが可能であり、発明の要旨を逸脱しない範囲で、種々の省略、置き換え、変更を行うことができる。当該実施形態やその変形は、発明の範囲や要旨に含まれるとともに、特許請求の範囲に記載された発明とその均等の範囲に含まれるものである。 Finally, various embodiments of the present invention have been described, but these are presented as examples and are not intended to limit the scope of the invention. The novel embodiment can be implemented in various other embodiments, and various omissions, replacements, and changes can be made without departing from the gist of the invention. The embodiment and its modifications are included in the scope and gist of the invention, and are included in the scope of the invention described in the claims and the equivalent scope thereof.
１　　　　：情報処理システム
２　　　　：ユーザ端末
２０　　　：通信バス
２１　　　：通信部
２２　　　：記憶部
２３　　　：制御部
２４　　　：表示部
２５　　　：入力部
３　　　　：サーバ
３０　　　：通信バス
３１　　　：通信部
３２　　　：記憶部
３３　　　：制御部
３３１　　：受付部
３３２　　：読出部
３３３　　：生成部
３３４　　：出力部
４　　　　：操作画面
４１　　　：選択欄
４２　　　：入力欄
４３　　　：ボタン
４４　　　：領域
４５　　　：入力欄
４６　　　：ボタン
５　　　　：ノード
５１　　　：ノード
５２　　　：ノード
５３　　　：ノード
６　　　　：リンク
６１　　　：リンク
６２　　　：リンク
６ａ　　　：操作アイコン
７　　　　：ソースコード
８　　　　：推定結果
９　　　　：ログイン画面
９１　　　：入力欄
９２　　　：入力欄
９３　　　：ログインボタン
９４　　　：登録リンク
ＣＲ　　　：接続関係
ＩＦ　　　：属性情報
ＮＫ　　　：ノード種別 1: Information processing system 2: User terminal 20: Communication bus 21: Communication unit 22: Storage unit 23: Control unit 24: Display unit 25: Input unit 3: Server 30: Communication bus 31: Communication unit 32: Storage unit 33: Control unit 331: Reception unit 332: Read unit 333: Generation unit 334: Output unit 4: Operation screen 41: Selection field 42: Input field 43: Button 44: Area 45: Input field 46: Button 5: Node 51: Node 52 : Node 53: Node 6: Link 61: Link 62: Link 6a: Operation icon 7: Source code 8: Estimated result 9: Login screen 91: Input field 92: Input field 93: Login button 94: Registration link CR: Connection relationship IF: Attribute information NK: Node type

Claims (17)

  1. 情報処理システムであって、
    　次の各ステップを実行するように構成され、
    　受付ステップでは、推論モデルを構築するためのＧＵＩに対するユーザの入力を受け付け、ここで前記入力は、前記推論モデルを構成する複数のノードそれぞれに対する属性情報と、前記複数のノード間の接続関係とを含み、
    　第１の生成ステップでは、前記入力に基づいて、前記属性情報の少なくとも一部と、前記接続関係とが把握可能な態様で、前記複数のノードを前記ＧＵＩ上に含む視覚情報を生成する、もの。     It is an information processing system
    It is configured to perform each of the following steps
    In the reception step, the user's input to the GUI for constructing the inference model is accepted, and here, the input receives the attribute information for each of the plurality of nodes constituting the inference model and the connection relationship between the plurality of nodes. Including
    In the first generation step, based on the input, visual information including the plurality of nodes on the GUI is generated in a manner in which at least a part of the attribute information and the connection relationship can be grasped. ..
  2. 請求項１に記載の情報処理システムにおいて、
    　前記属性情報の少なくとも一部は、ノード種別であり、前記ノード種別によって前記ノード間の前記接続関係が制限される、もの。     In the information processing system according to claim 1,
    At least a part of the attribute information is a node type, and the connection relationship between the nodes is limited by the node type.
  3. 請求項２に記載の情報処理システムにおいて、
    　第１の出力ステップをさらに実行するように構成され、
    　前記第１の出力ステップでは、接続不能な前記ノードの接続関係が入力された場合に、警告を出力するように構成される、もの。     In the information processing system according to claim 2,
    It is configured to perform the first output step further,
    In the first output step, a warning is output when the connection relationship of the node that cannot be connected is input.
  4. 請求項２又は請求項３に記載の情報処理システムにおいて、
    　前記第１の生成ステップでは、接続不能な前記ノードの接続関係が入力された場合に、この接続不能な前記接続関係を含む視覚情報を生成しないように構成される、もの。と     In the information processing system according to claim 2 or 3.
    The first generation step is configured so as not to generate visual information including the connectable connection relationship when the connection relationship of the connectable node is input. When
  5. 請求項２～請求項４の何れか１つに記載の情報処理システムにおいて、
    　前記第１の生成ステップでは、前記ノードを示す枠の形態で前記ノード種別を区別可能な前記視覚情報を生成し、ここで前記形態は、形状、模様若しくは色彩又はこれらの結合を含む、もの。     In the information processing system according to any one of claims 2 to 4.
    In the first generation step, the visual information that can distinguish the node type is generated in the form of a frame indicating the node, wherein the form includes a shape, a pattern, a color, or a combination thereof.
  6. 請求項２～請求項５の何れか１つに記載の情報処理システムにおいて、
    　前記ＧＵＩは、前記ノード種別を予め定められた項目から選択可能なオブジェクトを含み、
    　前記受付ステップでは、前記ユーザによる前記ノード種別の選択を前記入力として受け付ける、もの。     In the information processing system according to any one of claims 2 to 5.
    The GUI includes an object whose node type can be selected from predetermined items.
    In the reception step, the selection of the node type by the user is accepted as the input.
  7. 請求項２～請求項６の何れか１つに記載の情報処理システムにおいて、
    　前記ノード種別は、変数、関数及び確率モデルを含む、もの。     In the information processing system according to any one of claims 2 to 6.
    The node type includes variables, functions and probability models.
  8. 請求項２～請求項７の何れか１つに記載の情報処理システムにおいて、
    　前記ノード種別は、プログラミング言語であるＳｔａｎにおいて規定された、Ｄａｔａ、Ｔｒａｎｓｆｏｒｍｅｄ　Ｄａｔａ、Ｐａｒａｍｅｔｅｒ、Ｔｒａｎｓｆｏｒｍｅｄ　Ｐａｒａｍｅｔｅｒ、Ｌｉｋｅｌｉｈｏｏｄ　＆　Ｐｒｉｏｒ及びＧｅｎｅｒａｔｅｄ　Ｑｕａｎｔｉｔｙのうちの少なくとも１つを含む、もの。     In the information processing system according to any one of claims 2 to 7.
    The node type includes at least one of Data, Transformed Data, Parameter, Transformed Parameter, Likelihood & Prior, and Generated Quantity defined in the programming language Stan.
  9. 請求項１～請求項８の何れか１つに記載の情報処理システムにおいて、
    　前記属性情報は、対応する前記ノードにおける情報処理を表す第１のコードを含む、もの。     In the information processing system according to any one of claims 1 to 8.
    The attribute information includes a first code representing information processing in the corresponding node.
  10. 請求項１～請求項９の何れか１つに記載の情報処理システムにおいて、
    　第２の出力ステップをさらに実行するように構成され、
    　前記第２の出力ステップでは、表示された前記態様及び前記属性情報に対応する、前記推論モデルを示す第２のコードを出力する、もの。     In the information processing system according to any one of claims 1 to 9.
    It is configured to perform a second output step further,
    In the second output step, a second code indicating the inference model corresponding to the displayed aspect and the attribute information is output.
  11. 請求項１０に記載の情報処理システムにおいて、
    　実行ステップをさらに実行するように構成され、
    　前記実行ステップでは、前記第２のコードを実行する、もの。     In the information processing system according to claim 10,
    Configured to perform more execution steps,
    In the execution step, the second code is executed.
  12. 請求項１１に記載の情報処理システムにおいて、
    　第２の生成ステップをさらに実行するように構成され、
    　前記第２の生成ステップでは、前記第２のコードの実行結果に基づいて、設定された事象に関する、前記推論モデルによる推定結果を含む第２の視覚情報を生成する、もの。     In the information processing system according to claim 11,
    It is configured to perform a second generation step further,
    In the second generation step, based on the execution result of the second code, the second visual information including the estimation result by the inference model regarding the set event is generated.
  13. 請求項１２に記載の情報処理システムにおいて、
    　前記属性情報は、対応する前記ノードにおける情報処理を表す第１のコードを含み、
    　前記第１のコードは、前記第２のコードの少なくとも一部を構成する、もの。     In the information processing system according to claim 12,
    The attribute information includes a first code representing information processing at the corresponding node.
    The first code constitutes at least a part of the second code.
  14. 請求項１～請求項１３の何れか１つに記載の情報処理システムにおいて、
    　前記受付ステップでは、前記複数のノードのうちを順番に２つ選択した前記入力を受け付け、
    　前記第１の生成ステップでは、前記順番のうち先に選択された第１のノードを上流側、後に選択された第２のノードを下流側として、前記上流側及び前記下流側を把握可能な態様で示す視覚情報を生成する、もの。     In the information processing system according to any one of claims 1 to 13.
    In the reception step, the input in which two of the plurality of nodes are selected in order is received.
    In the first generation step, the upstream side and the downstream side can be grasped by setting the first node selected first in the order as the upstream side and the second node selected later as the downstream side. The one that produces the visual information shown in.
  15. 請求項１～請求項１４の何れか１つに記載の情報処理システムにおいて、
    　前記ＧＵＩは、所定のノードの上流側にある親ノード又は下流側にある子ノードに規定される変数を参照可能なオブジェクトを含む、もの。     In the information processing system according to any one of claims 1 to 14.
    The GUI includes an object that can refer to a variable specified in a parent node on the upstream side or a child node on the downstream side of a predetermined node.
  16. 情報処理方法であって、
    　請求項１～請求項１５の何れか１つに記載の情報処理システムの各ステップを備える、方法。     It is an information processing method
    A method comprising each step of the information processing system according to any one of claims 1 to 15.
  17. プログラムであって、
    　コンピュータに請求項１～請求項１５の何れか１つに記載の情報処理システムの各ステップを実行させる、もの。     It ’s a program,
    A computer that causes a computer to execute each step of the information processing system according to any one of claims 1 to 15.
PCT/JP2021/037310 2021-01-04 2021-10-08 Information processing system, information processing method, and program WO2022145100A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2021-000136 2021-01-04
JP2021000136A JP2022105387A (en) 2021-01-04 2021-01-04 Information processing system, information processing method and program

Publications (1)

Publication Number Publication Date
WO2022145100A1 true WO2022145100A1 (en) 2022-07-07

Family

ID=82259206

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2021/037310 WO2022145100A1 (en) 2021-01-04 2021-10-08 Information processing system, information processing method, and program

Country Status (2)

Country Link
JP (1) JP2022105387A (en)
WO (1) WO2022145100A1 (en)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2012181740A (en) * 2011-03-02 2012-09-20 Secom Co Ltd Information storage device
JP2015041121A (en) * 2013-08-20 2015-03-02 株式会社日立製作所 Sales forecast system and sales forecast method
JP2019125198A (en) * 2018-01-17 2019-07-25 株式会社日立製作所 Analysis model creation system, programming device, and analysis model creation method

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2012181740A (en) * 2011-03-02 2012-09-20 Secom Co Ltd Information storage device
JP2015041121A (en) * 2013-08-20 2015-03-02 株式会社日立製作所 Sales forecast system and sales forecast method
JP2019125198A (en) * 2018-01-17 2019-07-25 株式会社日立製作所 Analysis model creation system, programming device, and analysis model creation method

Also Published As

Publication number Publication date
JP2022105387A (en) 2022-07-14

Similar Documents

Publication Publication Date Title
US20210034990A1 (en) Rule Assignments and Templating
JP6351869B2 (en) Engineering tools, systems and modules
US9262126B2 (en) Recommendation system for agile software development
JP6978141B1 (en) How to generate a blueprint for a block object
US20220284286A1 (en) Method and apparatus for providing recommendations for completion of an engineering project
US7835894B2 (en) System and method for simulating home network
US20210248305A1 (en) System and method for interactive datasheets
WO2020220891A1 (en) Method and apparatus for generating configuration file of site in internet of things system
Chu et al. Accelerated norm-optimal iterative learning control algorithms using successive projection
WO2022145100A1 (en) Information processing system, information processing method, and program
CN109857791B (en) Data batch processing method and device
JP2020514870A (en) Image generation method, image generation device, image generation system, and program
WO2022037019A1 (en) System, method and device for implementing man-machine multi-round conversation
JP2023133285A (en) Information processing system, information processing method and program
CN115525321A (en) Distributed task generation method, device, equipment and storage medium
KR102254178B1 (en) Test device for artficial intellignece model service using user interface and method for testing thereof
US11995036B2 (en) Automated customized modeling of datasets with intuitive user interfaces
Hernández-Ocaña et al. Bacterial Foraging Based Algorithm Front-end to Solve Global Optimization Problems.
JP7406293B1 (en) Information processing system, information processing method and program
JP2015103160A (en) Component list creating system, component list creating method, and component list creating program
Nakata et al. Characterizing smart systems with interactive personalization
US11468067B2 (en) Information storage system for user inquiry-directed recommendations
JP7506867B1 (en) Document correction program, document correction device, and document correction method
JP3315890B2 (en) Data processing system
JP2023179349A (en) Information processing system, program and information processing method

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21914961

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21914961

Country of ref document: EP

Kind code of ref document: A1