JP2019144780A - Generating device, generating method, and generating program - Google Patents

Abstract

To allow efficient search for a given object.SOLUTION: A generation device according to the present application includes an acquisition unit, a first selection unit, a second selection unit, and a generation unit. The acquisition unit acquires information on a plurality of nodes to be searched for data, a first graph and a second graph. The first graph contains information on a first edge group that is a directed edge group connecting the plurality of nodes. The second graph contains information on the plurality of nodes. The first selection unit selects a first selection node from the plurality of nodes based on the number of output edges that end at other nodes in the first graph. The second selection unit selects a second selection node based on the length of the output edge among the target nodes to which the output edge starting from the first selection node is connected in the first graph. If the number of output edges starting from the first selection node is less than a predetermined threshold in the second graph, the generation unit generates a second graph in which the output edges are added as second edges.SELECTED DRAWING: Figure 3

Description

  The present invention relates to a generation device, a generation method, and a generation program.

  Conventionally, techniques for retrieving various information have been provided. For example, a technique for performing a search using graph data generated by an undirected edge is provided. Such a technique is used for image retrieval, for example.

JP 2011-090351 A Japanese Patent No. 5208001

  However, with the above-described conventional technology, it may be difficult to enable an efficient search for a predetermined target. For example, when a search is performed using graph data in which each node corresponding to each target is connected by an undirected edge, it is difficult to suppress costs such as processing time due to a search for a detour route.

  The present application has been made in view of the above, and an object thereof is to provide a generation device, a generation method, and a generation program that enable an efficient search for a predetermined target.

  The generating apparatus according to the present application includes a first graph including information on a plurality of nodes that are targets of data search, and information on a first edge group that is a directed edge group that connects the plurality of nodes, and the plurality of nodes An acquisition unit that acquires a second graph including node information; and a first selection node that selects a first selection node based on the number of output edges that end at another node in the first graph among the plurality of nodes. A first selection unit, and a second selection unit that selects a second selection node based on a length of the output edge from among target nodes to which an output edge starting from the first selection node is connected in the first graph In the second graph, when the number of output edges starting from the first selection node is less than a predetermined threshold, the second selection node starts from the first selection node in the first edge group. The updating process for adding the output edge of the end point as the information of the second edge in the second graph, is characterized in that and a generator for generating a second graph the second edge is added.

  According to one aspect of the embodiment, there is an effect that it is possible to efficiently search for a predetermined target.

FIG. 1 is a diagram illustrating an example of a generation process according to the embodiment. FIG. 2 is a diagram illustrating a configuration example of the generation system according to the embodiment. FIG. 3 is a diagram illustrating a configuration example of the generation apparatus according to the embodiment. FIG. 4 is a diagram illustrating an example of the object information storage unit according to the embodiment. FIG. 5 is a diagram illustrating an example of the base graph data storage unit according to the embodiment. FIG. 6 is a diagram illustrating an example of the first graph data storage unit according to the embodiment. FIG. 7 is a diagram illustrating an example of the threshold information storage unit according to the embodiment. FIG. 8 is a diagram illustrating an example of the second graph data storage unit according to the embodiment. FIG. 9 is a flowchart illustrating an example of the generation process according to the embodiment. FIG. 10 is a flowchart illustrating an example of the input addition process according to the embodiment. FIG. 11 is a diagram illustrating an example of the input addition process according to the embodiment. FIG. 12 is a diagram illustrating a comparative example according to the embodiment. FIG. 13 is a flowchart illustrating an example of search processing using graph data. FIG. 14 is a hardware configuration diagram illustrating an example of a computer that implements the function of the generation device.

  Hereinafter, a generation apparatus, a generation method, and a mode for executing a generation program (hereinafter referred to as “embodiment”) according to the present application will be described in detail with reference to the drawings. Note that the generation device, the generation method, and the generation program according to the present application are not limited by this embodiment. In the following embodiments, the same portions are denoted by the same reference numerals, and redundant description is omitted.

(Embodiment)
[1. Generation process)
An example of the generation process according to the embodiment will be described with reference to FIG. FIG. 1 is a diagram illustrating an example of a generation process according to the embodiment. In FIG. 1, new graph data (hereinafter also referred to as “second graph data”) based on the graph data (hereinafter also referred to as “first graph data”) provided by the generation apparatus 100 (see FIG. 3). The case of generating is shown. In the example of FIG. 1, the generation device 100 generates first graph data from predetermined graph data (hereinafter also referred to as “base graph data”), which will be described in detail later. Hereinafter, the graph data may be simply referred to as a graph. For example, the first graph data may be described as a first graph and the second graph data may be described as a second graph. That is, FIG. 1 shows a case where the generation apparatus 100 generates first graph data from base graph data and generates second graph data from the generated first graph data. The target information (object) may be any information as long as it can be expressed as a vector. In the following, vector information for image information will be described, but the target of vector information may be other targets such as moving image information and audio information.

  Further, as illustrated in FIG. 1, the generation device 100 performs generation processing on graph data. Here, the directed edge means an edge whose data can be traced only in one direction. In the following, a source traced by an edge, that is, a node as a starting point is referred to as a reference source, and a destination traced by an edge, that is, a node as an end point is defined as a reference destination. For example, the directed edge connected from the predetermined node “A” to the predetermined node “B” indicates that the reference source is the node “A” and the reference destination is the node “B”.

  Hereinafter, the edge having the node “A” as a reference source in this way is referred to as an output edge of the node “A”. In the following, an edge with the node “B” as a reference destination is referred to as an input edge of the node “B”. In other words, the output edge and the input edge here are differences in which one of the two nodes connected to the directed edge is regarded as the center, and the one directed edge. Becomes the output edge and the input edge. In other words, the output edge and the input edge are relative concepts, and for one directional edge, when the node as the reference source is regarded as the center, the output edge is regarded as the center, and the node as the reference destination is regarded as the center. The input edge. In the present embodiment, since the edges are directed edges such as output edges and input edges, the directed edges may be simply referred to as “edges” below.

  Each node here corresponds to each object. For example, each of the plurality of local feature amounts extracted from the image may be an object. Further, for example, various data in which distances between objects are defined may be objects.

  The generation apparatus 100 performs processing on a node corresponding to, for example, millions to hundreds of millions of units of image information, but only a part thereof is illustrated in the drawing. In the example of FIG. 1, in order to simplify the description, only 12 nodes are illustrated and an outline of the process will be described. For example, the generation apparatus 100 includes a plurality of nodes as indicated by the nodes N1 to N12 in FIG. 1, starts from the second graph data GR12-1 that does not include an edge, and sequentially adds the second graph. Data GR12 is generated.

  In addition, when “node N * (* is an arbitrary numerical value)” is described in this way, this indicates that the node is identified by the node ID “N *”. For example, when “node N1” is described, the node is a node identified by the node ID “N1”.

  In addition, when “edge E * (* is an arbitrary numerical value)” is described in this way, it indicates that the edge is an edge identified by the edge ID “E *”. For example, when “edge E2” is described, the edge is an edge identified by the edge ID “E2”. For example, it is possible to trace from the node N1 to the node N2 by the edge E2 connected with the node N1 as a reference source and the node N2 as a reference destination. In this case, the edge E2, which is a directed edge, becomes an output edge when identified with the node N1 as the center, and becomes an input edge when identified with the node N2 as the center. In other words, the edge E2, which is a directed edge, when viewed from the viewpoint from the node N1 side, becomes an edge in which an arrow points from itself to another edge, that is, an outward edge, and is viewed from the viewpoint from the node N2 side. In this case, the edge is directed toward itself, that is, an inward edge. That is, the output edge here can be read as an outward edge, and the input edge can be read as an inward edge.

  Also, the spatial information VS1-1 to VS1-6 shown in FIG. 1 is a diagram schematically showing a graph data generation process, and the spaces shown in the spatial information VS1-1 to VS1-6 are the same space. May be. Hereinafter, the spatial information VS1-1 to VS1-6 will be referred to as spatial information VS1 when they are not particularly distinguished.

  Further, the spatial information VS1 in FIG. 1 may be a Euclidean space. Further, the spatial information VS1 shown in FIG. 1 is a conceptual diagram for explaining the distance between the vectors, and the spatial information VS1 is a multidimensional space. For example, the spatial information VS1 illustrated in FIG. 1 is illustrated in a two-dimensional manner to be illustrated on a plane, but is assumed to be a multidimensional space such as 100 dimensions or 1000 dimensions.

  Further, the second graph data GR12-1 to GR12-4 shown in FIG. 1 are diagrams schematically showing the generation process of the graph data, and the second graph data GR12-1 to GR12-4 are generated by the generation process. Are the same graph data. In the following description, the second graph data GR12-1 to GR12-4 will be referred to as second graph data GR12 when they are not particularly distinguished.

  In the example shown in FIG. 1, in the base graph data GR10, the first graph data GR11, and the second graph data GR12-1 to GR12-4, “node N * (* is an arbitrary numerical value)” is appropriately illustrated. Is expressed by adding the value of “*” of “node N *” in “◯”. That is, it corresponds to the node with the “*” in the “node N *” part. For example, “O” in the upper left in the spatial information VS1 and “O” with “2” inside corresponds to the node identified by the node ID “N2”. For example, in the example shown in FIG. 1, the vector data corresponding to each node may be an N-dimensional real value vector.

  In the present embodiment, the distance of each node in the spatial information VS1 is set as the similarity between corresponding objects. For example, it is assumed that the similarity of objects (image information) corresponding to each node is mapped as a distance between nodes in the spatial information VS1. For example, it is assumed that the similarity between concepts corresponding to each node is mapped to the distance between the nodes. Here, in the example shown in FIG. 1, the similarity between the objects having a short distance between the nodes in the spatial information VS1 is high, and the similarity between the objects having a long distance between the nodes in the spatial information VS1 is low. For example, in the spatial information VS1 in FIG. 1, the node identified by the node ID “N5” and the node identified by the node ID “N8” are close to each other, that is, the distance is short. Therefore, the object corresponding to the node identified by the node ID “N5” and the object corresponding to the node identified by the node ID “N8” have a high degree of similarity.

  Further, for example, in the spatial information VS1 in FIG. 1, the node identified by the node ID “N8” and the node identified by the node ID “N12” are remote, that is, have a long distance. Therefore, the degree of similarity between the object corresponding to the node identified by the node ID “N8” and the object corresponding to the node identified by the node ID “N12” is low. The distance as the index indicating the similarity may be any distance as long as it can be applied as the distance between the vectors (N-dimensional vectors), such as the Euclidean distance, the Mahalanobis distance, and the cosine distance. Various distances may be used.

  In the example of FIG. 1, the generation device 100 generates graph data GR12 by connecting nodes with directed edges by an update process described later. For example, in the search using the graph data GR12, the same processing as the graph structure type index is performed at the time of search, but the start position (start point) may be started from a start point node determined using a predetermined index. Further, for example, when the search is performed using the graph data GR12 generated by the generation device 100, the search may be performed using a predetermined starting node as a starting point. For example, at the time of searching, when the starting node is the node N1, the nodes N2 to N12 and the like may be searched by following the directed edge from the node N1. Note that the generation apparatus 100 may perform a search according to the processing procedure illustrated in FIG. 13 at the time of the search, but details will be described later.

  From here, the detail of a production | generation process is demonstrated using FIG. Each step shown in FIG. 1 is a convenient step for explaining the generation of the second graph data, and the actual processing may be performed by more detailed processing steps. The generation process performed by the generation apparatus 100 may be any processing flow as long as the second graph data GR12 as shown in the second graph data GR12-4 in FIG. 1 is generated.

  First, the generation device 100 acquires base graph data (step S10). For example, the generation device 100 acquires base graph data from the base graph data storage unit 122. In the example of FIG. 1, the generation device 100 acquires base graph data GR10 as indicated by the spatial information VS1-1. For example, the base graph data GR10 may be a neighborhood graph generated based on a predetermined standard. For example, the base graph data GR10 is k-nearest neighbor graph data. In the example of FIG. 1, the base graph data GR10 will be described as an example in which k is “10” and the output edge is connected to 410 nodes in the vicinity from each node. Note that k may be various values such as “200”.

  The base graph data is not limited to k-neighborhood graph data, but may be any graph data as long as it is a graph in which a plurality of nodes are connected to directed edges, such as approximate k-neighborhood graph data. In FIG. 1, only the nodes and edges necessary for the description of the processing are shown, and the other nodes and edges are not shown. For example, illustration of output edges from the nodes N2 to N4, N6 to N8, etc. in the base graph data GR10 is omitted.

  And the production | generation apparatus 100 produces | generates 1st graph data from base graph data (step S11). For example, the generating apparatus 100 generates first graph data in which the direction of the edge of the base graph data is reversed by inverting the direction of the edge in the base graph data. That is, the generating apparatus 100 generates first graph data in which only the direction of the edge is inverted without changing the distance between the nodes and the like. For example, the generation apparatus 100 generates first graph data in which only the direction of the edge is inverted by using the input edge of one node in the base graph data as an output edge and the output edge of one node as an input edge. To do.

  In the example of FIG. 1, the generation device 100 generates first graph data GR11 as indicated by the spatial information VS1-2 from the base graph data GR10. The generation device 100 generates first graph data GR11 that is a transposed graph of the base graph data GR1. For example, the generating apparatus 100 replaces the start point and the end point of each edge with the reference destination of each node shown in the base graph data storage unit 122 in FIG. First graph data shown in the one graph data storage unit 123 is generated. For example, the generation apparatus 100 reverses the edge E2 having the node N1 as the start point and the node N2 as the end point, thereby setting the node N2 as the start point and the node N1 as the end point as shown in the spatial information VS1-2. First graph data GR11 including the edge E2-1 is generated. For example, the generating apparatus 100 may generate a transposed graph of graph data by appropriately using various conventional techniques. Note that the generation device 100 may generate the first graph data GR11 by any process as long as the first graph data GR11 of the base graph data GR10 can be generated.

  And the production | generation apparatus 100 produces | generates 2nd graph data from 1st graph data (step S12). First, the generation device 100 generates second graph data obtained by removing an edge (first edge) from the first graph data. For example, the generating apparatus 100 generates second graph data having only nodes by removing an edge group (first edge group) from the first graph data. In the example of FIG. 1, the generation device 100 generates second graph data GR12-1 having only nodes from the first graph data GR11, as indicated by the spatial information VS1-3. The generation apparatus 100 starts from the second graph data GR12-1 having only nodes, and sequentially adds edges (hereinafter also referred to as “second edges”) that satisfy the conditions to the second graph data. Thereby, the generation device 100 generates second graph data including the second edge group.

  Then, the generation apparatus 100 selects a node (hereinafter also referred to as “first selection node”) based on the number of output edges in the first graph data among the nodes in the second graph data. Then, the generation apparatus 100 has a short output edge among the nodes (hereinafter also referred to as “target nodes”) to which the output edges from the first selection node are connected, that is, the proximity of the distance to the first selection node. Based on the above, a node (hereinafter also referred to as “second selection node”) is selected.

  In the second graph, the generation apparatus 100 has less than another threshold value (hereinafter, also referred to as “first threshold value”) that has the second selection node as the end point, and the first selection node as the start point. When the number of output edges to be performed is less than a predetermined threshold (hereinafter also referred to as “second threshold”), an output edge having the first selection node as a start point and the second selection node as an end point is added to the second graph. The generation device 100 generates the second graph data by repeating the above update process for each first selection node. In the example of FIG. 1, as illustrated in the threshold value list TL <b> 11, a case where it is less than the first threshold “4” and less than the second threshold “3” will be described as an example.

  For example, the generation apparatus 100 selects the first selection node in order from the node having the smaller output edge in the first graph data among the nodes in the second graph data. In the example of FIG. 1, the generation device 100 selects the first selection node based on the number of output edges of each of the nodes N1 to N12 as shown in the list information LT11.

  The generation apparatus 100 selects the node N1 having the smallest number of output edges (zero) as the first selection node, and performs the update process. Then, the generation apparatus 100 selects the node N2 having the smallest number of output edges (one) next to the node N1 as the first selection node, and performs the update process. In addition, the generation apparatus 100 selects the node N2 having the smallest number of output edges (one) next to the node N1 as the first selection node, and performs the update process. Further, the generation device 100 selects the node N3 having the fewest number of output edges (one) next to the node N1 as the first selection node, and performs the update process. As described above, in the example of FIG. 1, the generation apparatus 100 generates the second graph data to which the second edge is added by performing the update process using the nodes N1 to N4 as the first selection nodes sequentially (step S13). ). In other words, the generation apparatus 100 generates the second graph data with the second edge added by updating (adding) the edge (second edge) included in the second graph data by performing the update process. .

  Since the generation apparatus 100 has zero output edges starting from the node N1 in the first graph data GR11, there is no target node to be selected as the second selection node, and thus the second graph data GR12-1 has an edge. The update process using the node N1 as the first selected node is terminated without adding (second edge).

  Further, in the update process in which the node N2 is the first selection node, the generation apparatus 100 has one output edge starting from the node N2 in the first graph data GR11. Select as the second selection node. In the second graph data GR12-1, the generation device 100 has zero input edges whose end point is the node N1 that is the second selection node, which is less than the first threshold “4”, and the first selection node Since the number of output edges having the node N2 as the starting point is one less than the second threshold “3”, the edge E2-1 having the node N2 as the starting point and the node N1 as the ending point is added to the second graph data GR12. To do. As described above, the generation apparatus 100 determines whether or not the condition based on the comparison between the number of edges and the first threshold value and the second threshold value is satisfied. If the generation apparatus 100 determines that the condition is satisfied, the generation apparatus 100 adds an edge corresponding to the determination target node to the second graph data.

  In addition, in the update process in which the node N3 is the first selected node, the generation apparatus 100 has one output edge starting from the node N3 in the first graph data GR11. Select as the second selection node. In the second graph data GR12 after the addition of the edge E2-1, the generation apparatus 100 has one input edge whose end point is the node N1, which is the second selection node, less than the first threshold “4”. The number of output edges starting from the node N3 that is the first selection node is zero, which is less than the second threshold “3”. Therefore, the generation device 100 determines that the condition based on the first threshold value and the second threshold value is satisfied, and adds the edge E3-1 having the node N3 as the start point and the node N1 as the end point to the second graph data GR12.

  In addition, in the update process in which the node N4 is the first selection node, the generation apparatus 100 has one output edge starting from the node N4 in the first graph data GR11. Select as the second selection node. Then, in the second graph data GR12 after the addition of the edges E2-1 and E3-1, the generation apparatus 100 has the number of input edges whose end point is the node N1 that is the second selection node is less than the first threshold “4”. There are two, and the number of output edges starting from the node N4 that is the first selection node is zero, which is less than the second threshold “3”. Therefore, the generating apparatus 100 determines that the condition based on the first threshold value and the second threshold value is satisfied, and adds an edge E4-1 having the node N4 as a start point and the node N1 as an end point to the second graph data GR12. As a result, the generating apparatus 100, as shown in the spatial information VS1-4, the second graph data GR12- to which three edges (second edges) of the edges E2-1, E3-1, and E4-1 are added. 2 is generated.

  Then, as illustrated in the list information LT11, the generation apparatus 100 selects, as the first selection node, the node N10 having the next smallest number of output edges (three) from the nodes N2 to N4 (step S14).

  Then, the generation apparatus 100 sequentially selects the second selection node based on the length of the output edge among the target nodes to which the output edge starting from the first selection node is connected in the first graph, Second graph data is generated (step S15). In the example of FIG. 1, the generation device 100 selects the second selection node using the three nodes N1, N11, and N12 to which the output edge from the node N10 is connected as target nodes. For example, in the first graph data GR11, the generation apparatus 100 selects the second selection in order from the node to which the short edge is connected among the edges E10-1, E11-1, and E12-1 starting from the node N10. Select as a node.

  In the example of FIG. 1, as illustrated in the list information LT12, the generation apparatus 100 includes the lengths of the edges E10-1, E11-1, and E12-1, that is, the node N10 and the three nodes N1, N11, and N12. Calculate the distance to the node. For example, the generation device 100 calculates a distance between two nodes using a vector of two nodes. For example, the generating apparatus 100 calculates the length of the edge E10-1, that is, the distance between the node N10 and the node N1, using the vector of the node N10 and the vector of the node N1. Note that the edge length calculated by the generation apparatus 100, that is, the distance between nodes may be any distance, and may be, for example, a Euclidean distance, a Mahalanobis distance, a cosine distance, or the like.

  In the example of FIG. 1, as illustrated in the list information LT12, the generation apparatus 100 calculates distances between the node N10 and the three nodes N1, N11, and N12 as distances D10, D11, and D12, respectively. That is, the generation apparatus 100 calculates the length of the edge E10-1 as the distance D10, calculates the length of the edge E11-1 as the distance D11, and calculates the length of the edge E12-1 as the distance D12. Also, as shown in the list information LT12 in FIG. 1, it is assumed that the distance D11 is the smallest and the distance D10 is the largest among the distances D10, D11, and D12.

  Then, the generation device 100 selects each node indicated as a reference destination in order from the first rank shown in the list information LT12, and determines whether to add an edge. In the example of FIG. 1, the generation apparatus 100 selects the node N11 that is the reference destination of the first rank as the second selection node, and determines whether to add an edge.

  When the first selection node is the node N10 and the node N11 is the second selection node, the generation apparatus 100 has the number of input edges whose end point is the node N11 that is the second selection node being less than the first threshold “4”. The number of output edges starting from the node N10, which is the first selection node, is 0, which is less than the second threshold “3”. Therefore, the generating apparatus 100 determines that the condition based on the first threshold value and the second threshold value is satisfied, and adds the edge E11-1 having the node N10 as a start point and the node N11 as an end point to the second graph data GR12.

  Next, when the first selection node is the node N10 and the node N12 is the second selection node, the generation apparatus 100 determines that the number of input edges whose end point is the node N12 that is the second selection node is the first threshold “ The number of output edges starting from the node N10 that is the first selection node is one of the edges E11-1 and is less than the second threshold value “3”. Therefore, the generation device 100 determines that the condition based on the first threshold value and the second threshold value is satisfied, and adds the edge E12-1 having the node N10 as a start point and the node N12 as an end point to the second graph data GR12.

  Next, when the first selection node is the node N10 and the node N1 is the second selection node, the generation apparatus 100 determines that the number of input edges whose end point is the node N1 that is the second selection node is the edge E2-1. , E3-1 and E4-1, which are less than the first threshold “4”, and the number of output edges starting from the node N10 which is the first selection node is the edges E11-1 and E12-1. The number is two and is less than the second threshold “3”. Therefore, the generation device 100 determines that the condition based on the first threshold value and the second threshold value is satisfied, and adds the edge E10-1 having the node N10 as the start point and the node N1 as the end point to the second graph data GR12. As a result, the generation apparatus 100, as shown in the spatial information VS1-5, the second graph data GR12- to which three edges (second edges) of the edges E10-1, E11-1, and E12-1 are added. 3 is generated.

  Then, as illustrated in the list information LT11, the generation apparatus 100 selects, as the first selection node, the node N5 having the next smallest number of output edges (five) after the node N10 (step S16).

  Then, the generation apparatus 100 sequentially selects the second selection node based on the length of the output edge among the target nodes to which the output edge starting from the first selection node is connected in the first graph, Second graph data is generated (step S17). In the example of FIG. 1, the generation apparatus 100 selects the second selection node using five nodes N1 and N6 to N9 to which the output edge from the node N5 is connected as target nodes. For example, in the first graph data GR11, the generation apparatus 100 selects, as the second selection node, in order from the node to which the edge having the short length is connected among the edges E5-1 to E9-1 starting from the node N5. .

  In the example of FIG. 1, as illustrated in the list information LT13, the generation apparatus 100 has a length of each edge E5-1 to E9-1, that is, between the node N5 and the five nodes N1, N6 to N9. The distance is calculated. In the example of FIG. 1, as illustrated in the list information LT13, the generation apparatus 100 calculates distances between the node N5 and the five nodes N1, N6 to N9 as distances D5 to D9, respectively. That is, the generation apparatus 100 calculates the length of the edge E10-1 as the distance D10, calculates the length of the edge E11-1 as the distance D11, and calculates the length of the edge E12-1 as the distance D12. Also, as shown in the list information LT12 in FIG. 1, it is assumed that the distance D8 is the smallest and the distance D9 is the largest of the distances D5 to D9. Further, the distances D8, D6, D5, D7, and D9 are assumed to be smaller in this order.

  Then, the generation device 100 selects each node indicated as the reference destination in order from the first rank in the list information LT13, and determines whether to add an edge. In the example of FIG. 1, the generation apparatus 100 selects the node N8 that is the reference destination of the first rank as the second selection node, and determines whether to add an edge.

  First, when the first selection node is the node N5 and the node N8 is the second selection node, the generation apparatus 100 determines that the number of input edges whose end point is the node N8 that is the second selection node is the first threshold “4”. The number of output edges starting from the node N5 that is the first selection node is zero that is less than the second threshold “3”. Therefore, the generation device 100 determines that the condition based on the first threshold value and the second threshold value is satisfied, and adds the edge E8-1 having the node N5 as the start point and the node N8 as the end point to the second graph data GR12.

  Next, when the first selection node is the node N5 and the node N6 is the second selection node, the generation apparatus 100 determines that the number of input edges having the node N6 that is the second selection node as the end point is the first threshold “ The number of output edges starting from the node N5 that is the first selection node is one of the edges E8-1, and is less than the second threshold “3”. Therefore, the generating apparatus 100 determines that the condition based on the first threshold value and the second threshold value is satisfied, and adds the edge E6-1 having the node N5 as the start point and the node N6 as the end point to the second graph data GR12.

  Next, when the first selection node is the node N5 and the node N1 is the second selection node, the generation apparatus 100 determines that the number of input edges whose end point is the node N1 that is the second selection node is the edge E2-1. , E3-1, E4-1, and E10-1, and are not less than the first threshold “4”. Therefore, the generation apparatus 100 determines that the combination in which the first selection node is the node N5 and the node N1 is the second selection node does not satisfy the condition based on the first threshold, and the node N5 is the start point and the node N1 is the end point. The edge E5-1 is not added to the second graph data GR12. Thereby, the generation device 100 can suppress the concentration of input edges at a specific node.

  Next, when the first selection node is the node N5 and the node N7 is the second selection node, the generation apparatus 100 determines that the number of input edges having the node N7 that is the second selection node as the end point is the first threshold “ The number of output edges starting from the node N5 that is the first selection node is two, that is, the edges E8-1 and E6-1, and is less than the second threshold “3”. Therefore, the generation device 100 determines that the condition based on the first threshold value and the second threshold value is satisfied, and adds the edge E7-1 having the node N5 as a start point and the node N7 as an end point to the second graph data GR12.

  Next, when the first selection node is the node N5 and the node N9 is the second selection node, the generation apparatus 100 determines that the number of input edges whose end point is the node N9 that is the second selection node is the first threshold “ Since it is 0 less than 4 ”, it is determined that the first threshold condition is satisfied. On the other hand, in the generation apparatus 100, the number of output edges starting from the node N5 that is the first selection node is three edges E6-1 to E8-1, and is not less than the second threshold “3”. It is determined that the two threshold conditions are not satisfied. Therefore, the generation device 100 does not add the edge E9-1 having the node N5 as the start point and the node N9 as the end point to the second graph data GR12. Thereby, the production | generation apparatus 100 can suppress that processing time and a processing load increase by suppressing the increase in the output edge of a specific node. Accordingly, the generation device 100 generates the second graph data GR12-4 to which the three edges (second edges) of the edges E6-1 to E8-1 are added, as indicated by the spatial information VS1-6. .

  As described above, the generation apparatus 100 uses only the edge determined to satisfy the condition based on the first threshold value and the second threshold value as the second edge, out of the first edge group in the first graph data GR11. By adding to the graph data, it is possible to generate second graph data in which input edges and output edges are suppressed from being concentrated on a specific node. Thereby, the generating apparatus 100 can generate graph data that enables an efficient search for a predetermined target.

  As described above, the generation apparatus 100 determines whether to add an edge using two threshold values of the first threshold value and the second threshold value that can be appropriately set. Thereby, the generation device 100 can generate a desired second graph for the number of output edges and input edges. Therefore, the generation device 100 can generate graph data that enables an efficient search for a predetermined target.

  Conventionally, neighborhood graphs (including k-neighbor graphs and the like) in which each node is connected to neighboring nodes by directed edges in the graph structure are used for neighborhood search. In such a search using a neighborhood graph including a directed edge, a node with too few input edges is less likely to reach that node, leading to a decrease in search accuracy. A node having too many output edges increases the number of distance calculations, leading to an increase in search time. Therefore, the generation apparatus 100 uses a threshold value (first threshold value, second threshold value) corresponding to each of the input edge and the output edge to generate a node with a small number of input edges or a node with a large number of output edges when generating a graph. cut back. In other words, the generation device 100 generates the second graph so as to suppress generation of a second graph including a node having a small number of input edges and a node having a large number of output edges. Thereby, the generation device 100 generates a desired second graph, and enables the search performance to be improved by the second graph.

  Further, for example, if the graph data includes a large number (for example, 1000, 10,000, etc.) of nodes including output edges, the processing load of the search that follows the nodes increases. Further, for example, when a node including a small number (for example, 0 or 1) of input edges is included in the graph data, the possibility that the node is searched is reduced. As described above, the generation device 100 adjusts the edges so that the output edges and the input edges are not biased, and particularly the output edges of the specific nodes are not increased. In addition, it is possible to reduce the possibility of including a node that is unlikely to be searched. Therefore, the generation device 100 can generate graph data that enables an efficient search for a predetermined target. In addition, the graph data generated by the generation device 100 can enable an efficient search for a predetermined target while suppressing an increase in the number of edges.

[1-1. First graph data]
In the example of FIG. 1, the generation apparatus 100 generates the second graph data GR12 from the given first graph data GR11. However, the generation apparatus 100 uses the second graph data without using the first graph data. May be generated. For example, the generation device 100 may generate the second graph data from a plurality of nodes whose edges are not connected.

  For example, the generation device 100 may generate the second graph data based on the rank based on the distance from each node to all other nodes. For example, the generation apparatus 100 sets each node as a start point node, and in order from the shortest distance from each node, an output edge having another node of the first threshold number as an end node, and each node as an end point node, Graph data is generated by linking each node with an input edge having another node of the second threshold number as a starting point node in order from the shortest distance to the node. In this case, the generation apparatus 100 performs the same processing when the graph data in which the output edge is connected from each node to all other nodes is the target. As described above, the generation device 100 may generate the second graph data from a plurality of nodes whose edges are not connected.

[1-2. First threshold, second threshold]
In the example of FIG. 1, the first threshold is set to “4” and the second threshold is set to “3”, but various values may be selected for the first threshold and the second threshold. For example, the first threshold value and the second threshold value may be the same value. For example, the first threshold value may be smaller than the second threshold value.

[1-3. Index data)
Further, the generation device 100 may generate index data used at the time of search. For example, the generating apparatus 100 generates a search index for searching for a high-dimensional vector as index data. The high-dimensional vector here may be, for example, a vector of several hundreds to thousands of dimensions, or a vector of more dimensions.

  For example, the generating apparatus 100 may generate a search index related to a tree structure (tree structure) as index data. For example, the generating apparatus 100 may generate a search index related to a kd tree (k-dimensional tree) as index data. For example, the generation apparatus 100 may generate a search index related to a VP tree (Vantage-Point tree) as index data.

  For example, the generating apparatus 100 may generate index data having other tree structures. For example, the generation apparatus 100 may generate various index data in which leaves of tree-structured index data are connected to graph data. For example, the generating apparatus 100 may generate various index data in which the leaf of the tree-structured index data corresponds to a node in the second graph data. Further, when performing a search using such index data, the generation apparatus 100 may search for graph data from a leaf (node) that arrives by tracing the index data.

  Note that the index data as described above is an example, and the generation apparatus 100 may generate index data having any data structure as long as it can identify a query in the graph data at high speed. For example, as long as the centroid information corresponding to the query can be identified at high speed, the generation apparatus 100 may generate index data by appropriately using various conventional techniques such as a technique related to binary space division. . For example, the generation apparatus 100 may generate index data having any data structure as long as it is an index that can be used to search for a high-dimensional vector. By using the index data and the second graph data as described above, the generation device 100 can enable more efficient search regarding a predetermined target.

[2. Generation system configuration)
As illustrated in FIG. 2, the generation system 1 includes a terminal device 10, an information providing device 50, and a generation device 100. The terminal device 10, the information providing device 50, and the generating device 100 are connected via a predetermined network N so as to be communicable by wire or wirelessly. FIG. 2 is a diagram illustrating a configuration example of the generation system according to the embodiment. The generation system 1 illustrated in FIG. 2 may include a plurality of terminal devices 10, a plurality of information providing devices 50, and a plurality of generation devices 100.

  The terminal device 10 is an information processing device used by a user. The terminal device 10 receives various operations by the user. Hereinafter, the terminal device 10 may be referred to as a user. That is, hereinafter, the user can be read as the terminal device 10. The terminal device 10 described above is realized by, for example, a smartphone, a tablet terminal, a notebook PC (Personal Computer), a desktop PC, a mobile phone, a PDA (Personal Digital Assistant), or the like.

  The information providing apparatus 50 is an information processing apparatus in which information for providing various information to a user or the like is stored. For example, the information providing device 50 stores an object ID based on character information collected from various external devices such as a web server. For example, the information providing apparatus 50 is an information processing apparatus that provides an image search service to a user or the like. For example, the information providing apparatus 50 stores information for providing an image search service. For example, the information providing apparatus 50 provides the generating apparatus 100 with vector information corresponding to an image that is a target of the image search service. In addition, the information providing apparatus 50 receives the object ID indicating the image corresponding to the query from the generating apparatus 100 by transmitting the query to the generating apparatus 100.

  The generation device 100 sets each node as a start point node, an output edge having other nodes of the first threshold number in order from the shortest distance from each node as an end point node, each node as an end point node, The graph data is generated by connecting the second graph data from the first graph data with the input edge having the other node of the second threshold number as the starting node in order from the shortest distance to the respective nodes. For example, the generation device 100 generates second graph data from the first graph data.

  For example, when receiving the query information (hereinafter, also simply referred to as “query”) from the terminal device, the generation device 100 searches for an object (such as vector information) similar to the query and provides the search result to the terminal device. Further, for example, the data provided to the terminal device by the generating apparatus 100 may be data such as image information itself or information for referring to corresponding data such as a URL (Uniform Resource Locator). Good. The query and search target data may be any kind of data such as images, sounds, text data, and the like. In the present embodiment, a case where the generation apparatus 100 searches for an image will be described as an example.

[3. Configuration of the generator
Next, the configuration of the generation apparatus 100 according to the embodiment will be described with reference to FIG. FIG. 3 is a diagram illustrating a configuration example of the generation apparatus 100 according to the embodiment. As illustrated in FIG. 3, the generation device 100 includes a communication unit 110, a storage unit 120, and a control unit 130. The generation device 100 includes an input unit (for example, a keyboard and a mouse) that receives various operations from an administrator of the generation device 100 and a display unit (for example, a liquid crystal display) for displaying various types of information. May be.

(Communication unit 110)
The communication unit 110 is realized by, for example, a NIC (Network Interface Card). The communication unit 110 is connected to a network (for example, the network N in FIG. 2) by wire or wireless, and transmits and receives information to and from the terminal device 10 and the information providing device 50.

(Storage unit 120)
The storage unit 120 is realized by, for example, a semiconductor memory device such as a RAM (Random Access Memory) or a flash memory, or a storage device such as a hard disk or an optical disk. As shown in FIG. 3, the storage unit 120 according to the embodiment includes an object information storage unit 121, a base graph data storage unit 122, a first graph data storage unit 123, a threshold information storage unit 124, and a second graph. And a data storage unit 125.

(Object information storage unit 121)
The object information storage unit 121 according to the embodiment stores various types of information related to objects. For example, the object information storage unit 121 stores an object ID and vector data. FIG. 4 is a diagram illustrating an example of the object information storage unit according to the embodiment. The object information storage unit 121 illustrated in FIG. 4 includes items such as “object ID” and “vector information”.

  “Object ID” indicates identification information for identifying an object. “Vector information” indicates vector information corresponding to the object identified by the object ID. That is, in the example of FIG. 4, vector data (vector information) corresponding to an object is registered in association with an object ID for identifying the object.

  For example, in the example of FIG. 4, the object (target) identified by the ID “OB1” is associated with multidimensional vector information “10, 24, 51, 2.

  The object information storage unit 121 is not limited to the above, and may store various information according to the purpose.

(Base graph data storage unit 122)
The base graph data storage unit 122 according to the embodiment stores various types of information related to graph data. For example, the base graph data storage unit 122 stores graph data serving as a basis for generating the first graph data. In the example of FIG. 5, the base graph data storage unit 122 stores neighborhood graph data. FIG. 5 is a diagram illustrating an example of the base graph data storage unit according to the embodiment. The base graph data storage unit 122 illustrated in FIG. 5 includes items such as “node ID”, “object ID”, and “directed edge information”. The “directed edge information” includes information such as “edge ID” and “reference destination”.

  “Node ID” indicates identification information for identifying each node (target) in the graph data. The “object ID” indicates identification information for identifying the object.

  “Directed edge information” indicates information related to an edge connected to a corresponding node. In the example of FIG. 5, “directed edge information” indicates information regarding the output edge output from the corresponding node. “Edge ID” indicates identification information for identifying an edge connecting nodes. “Reference destination” indicates information indicating a reference destination (node) connected by an edge. That is, in the example of FIG. 5, with respect to a node ID for identifying a node, information for identifying an object (target) corresponding to the node and a reference destination to which a directed edge (output edge) from the node is linked ( Node) is registered in association with each other.

  For example, in the example of FIG. 5, the node (node N1) identified by the node ID “N1” corresponds to the object (target) identified by the object ID “OB1”. The node N1 indicates that the edge (edge E2) identified by the edge ID “E2” is connected to the node (node N1) identified by the node ID “N2”. That is, in the example of FIG. 5, it can be traced from the node N1 in the base graph data to the node N2 by the edge E2.

  The base graph data storage unit 122 is not limited to the above, and may store various types of information according to the purpose. For example, the base graph data storage unit 122 may store the lengths of edges connecting between nodes (vectors). That is, the base graph data storage unit 122 may store information indicating the distance between each node (vector).

(First graph data storage unit 123)
The 1st graph data storage part 123 concerning an embodiment memorizes various information about the 1st graph data. For example, the first graph data storage unit 123 stores first graph data that is a basis for generating second graph data. FIG. 6 is a diagram illustrating an example of the first graph data storage unit according to the embodiment. The first graph data storage unit 123 illustrated in FIG. 6 includes items such as “node ID”, “object ID”, and “directed edge information”. The “directed edge information” includes information such as “edge ID” and “reference destination”.

  “Node ID” indicates identification information for identifying each node (target) in the graph data. The “object ID” indicates identification information for identifying the object.

  “Directed edge information” indicates information related to an edge connected to a corresponding node. In the example of FIG. 6, “directed edge information” indicates information regarding the output edge output from the corresponding node. “Edge ID” indicates identification information for identifying an edge connecting nodes. “Reference destination” indicates information indicating a reference destination (node) connected by an edge. That is, in the example of FIG. 6, with respect to a node ID for identifying a node, information for identifying an object (target) corresponding to the node and a reference destination to which a directed edge (output edge) from the node is linked ( Node) is registered in association with each other.

  For example, in the example of FIG. 6, the node identified by the node ID “N1” corresponds to the object (target) identified by the object ID “OB1”.

  The parentheses in the node ID indicate the ID when the node ID between the graph data can be identified. In the example of FIG. 6, each node ID in the first graph data storage unit 123 stores an ID in which “−1” is added to the corresponding node ID in the base graph data storage unit 122. In this case, the node ID corresponding to the reference destination is also stored. Specifically, a node ID in which “−1” is added to the end of the node ID in FIG. 6 may be stored in the reference destination. For example, the node ID “N1-1” is stored in the reference destination corresponding to the edge E2-1.

  The example of FIG. 6 indicates that there is no output edge from the node identified by the node ID “N1”. Further, from the node (node N2) identified by the node ID “N2”, the edge (edge E2-1) identified by the edge ID “E2-1” is identified by the node ID (N1) ( Node N1) is connected. That is, in the example of FIG. 6, it can be traced from the node N2 in the first graph data to the node N1 by the edge E2-1.

  In addition, the 1st graph data memory | storage part 123 may memorize | store not only the above but various information according to the objective. For example, the first graph data storage unit 123 may store the lengths of edges connecting between nodes (vectors). That is, the first graph data storage unit 123 may store information indicating the distance between each node (vector).

(Threshold information storage unit 124)
The threshold information storage unit 124 according to the embodiment stores various types of information related to the threshold. For example, the threshold information storage unit 124 stores a first threshold and a second threshold. FIG. 7 is a diagram illustrating an example of a threshold storage unit according to the embodiment. The threshold information storage unit 124 illustrated in FIG. 7 includes items such as “threshold name”, “value”, “target node”, and “target edge”.

  “Threshold name” indicates information (name) for identifying a threshold. The “value” indicates a specific value of the corresponding threshold value. “Target node” indicates a target node for which a corresponding threshold is used. The “target edge” indicates an edge that is a target for which a corresponding threshold value is used.

  In the example of FIG. 7, the value of the threshold name “first threshold” is “4”. The target node having the first threshold is the second selection node, and the target edge is an input edge. Further, the value of the threshold name “second threshold” is “3”. The target node with the second threshold is the first selection node, and the target edge is the output edge.

  The threshold information storage unit 124 is not limited to the above, and may store various information according to the purpose.

(Second graph data storage unit 125)
The 2nd graph data storage part 125 concerning an embodiment memorizes various information about the 2nd graph data. FIG. 8 is a diagram illustrating an example of the second graph data storage unit according to the embodiment. In the example of FIG. 8, the second graph data storage unit 125 includes items such as “node ID”, “object ID”, and “directed edge information”. The “directed edge information” includes information such as “edge ID” and “reference destination”.

  “Node ID” indicates identification information for identifying each node (target) in the graph data. The “object ID” indicates identification information for identifying the object.

  “Directed edge information” indicates information related to an edge connected to a corresponding node. In the example of FIG. 8, “directed edge information” indicates information regarding output edges output from the corresponding nodes. “Edge ID” indicates identification information for identifying an edge connecting nodes. “Reference destination” indicates information indicating a reference destination (node) connected by an edge. That is, in the example of FIG. 8, with respect to a node ID for identifying a node, information for identifying an object (target) corresponding to the node and a reference destination to which a directed edge (output edge) from the node is linked ( Node) is registered in association with each other.

  For example, in the example of FIG. 8, the node identified by the node ID “N2” corresponds to the object (target) identified by the object ID “OB2”.

  The parentheses in the node ID indicate the ID when the node ID between the graph data can be identified. In the example of FIG. 8, each node ID of the second graph data storage unit 125 stores an ID in which “−2” is added to the corresponding node ID in the base graph data storage unit 122. In this case, the node ID corresponding to the reference destination is also stored. Specifically, a node ID in which “−2” is added to the end of the node ID in FIG. 8 may be stored in the reference destination. For example, the node ID “N6-2” is stored in the reference destination corresponding to the edge E6-1.

  The example of FIG. 8 indicates that there is no output edge from the node identified by the node ID “N1”. Further, from the node (node N2) identified by the node ID “N2”, the edge (edge E2-1) identified by the edge ID “E2-1” is identified by the node ID (N1) ( Node N1) is connected. That is, the example of FIG. 8 shows that the node N2 in the first graph data can be traced to the node N1 by the edge E2-1. That is, in the example of FIG. 8, since the edge E2-1 in the first graph data satisfies the condition, it is added to the second graph data.

  In addition, the 2nd graph data storage part 125 may memorize | store not only the above but various information according to the objective. For example, the second graph data storage unit 125 may store the length of the edge connecting between the nodes (vectors). That is, the second graph data storage unit 125 may store information indicating the distance between each node (vector). For example, the second graph data storage unit 125 may store information indicating the number of input edges to each node. The second graph data may include a program module that receives a query, searches for a node by tracing an edge in the graph data, extracts a node similar to the query, and outputs the extracted node. That is, the second graph data may be assumed to be used as a program module that performs a search process using the second graph. For example, the second graph data GR12 may be a program that, when vector data is input as a query, extracts and outputs nodes corresponding to vector data similar to the vector data from the second graph. For example, the second graph data GR12 may be data used as a program module for searching for a similar image corresponding to the query image. For example, the second graph data GR12 causes the computer to extract and output a node similar to the query in the second graph based on the input query.

(Control unit 130)
Returning to the description of FIG. 3, the control unit 130 is a controller and is stored in a storage device inside the generation apparatus 100 by, for example, a CPU (Central Processing Unit), an MPU (Micro Processing Unit), or the like. Various programs (corresponding to an example of a generation program) are implemented by using the RAM as a work area. The control unit 130 is a controller, and is realized by an integrated circuit such as an application specific integrated circuit (ASIC) or a field programmable gate array (FPGA).

  As illustrated in FIG. 3, the control unit 130 includes an acquisition unit 131, a selection unit 132 as a first selection unit and a second selection unit, a generation unit 133, a generation unit 134, and a provision unit 135. The functions and operations of information processing described below are realized or executed. Note that the internal configuration of the control unit 130 is not limited to the configuration illustrated in FIG. 3, and may be another configuration as long as the information processing described below is performed.

(Acquisition part 131)
The acquisition unit 131 acquires various types of information. For example, the acquisition unit 131 acquires various types of information from the storage unit 120. For example, the acquisition unit 131 receives various information from the object information storage unit 121, the base graph data storage unit 122, the first graph data storage unit 123, the threshold information storage unit 124, the second graph data storage unit 125, and the like. get. The acquisition unit 131 acquires various types of information from an external information processing apparatus.

  The acquisition unit 131 includes a first graph (first graph data GR11) including information on a plurality of nodes to be subjected to data search and information on a first edge group that is a directed edge group connecting the plurality of nodes. ) And second graph data GR12-1 including information on a plurality of nodes.

  For example, the acquisition unit 131 acquires the first graph data GR11 from the first graph data storage unit 123. For example, the acquisition unit 131 acquires the second graph data GR12-1 from the second graph data storage unit 125. The acquisition unit 131 is a neighborhood graph (base graph data GR10) generated based on a predetermined criterion, and is a neighborhood graph connected by a predetermined edge group that is a directed edge group connecting a plurality of nodes. Among these, a first graph that is a transposed graph including information of the first edge group obtained by inverting the direction of each edge of the predetermined edge group is acquired.

  For example, the acquisition unit 131 acquires base graph data. In the example of FIG. 1, the generation device 100 acquires base graph data GR10. For example, the acquisition unit 131 may acquire the first graph data. For example, the generation device 100 may acquire the first graph data GR11 from an external device such as the information providing device 50.

  For example, the acquisition unit 131 acquires information related to the search query. For example, the acquisition unit 131 acquires a search query related to image search. For example, the acquisition unit 131 acquires a query from the information providing device 50 that has received the query from the terminal device 10 to be used.

(Selection unit 132)
The selection unit 132 selects various information. FIG. 3 shows a configuration in which the first selection unit and the second selection unit are combined into the selection unit 132, but the first selection unit and the second selection unit are different from each other (the first selection unit and the first selection unit). 2 selection unit). That is, each process described as the process of the selection unit 132 may be executed by a first selection unit or a second selection unit configured individually.

  The selection unit 132 selects a node in the update process. The selection unit 132 selects a first selection node among a plurality of nodes based on the number of output edges that end at another node in the first graph. The selection unit 132 selects the second selection node based on the length of the output edge among the target nodes to which the output edge starting from the first selection node is connected in one graph.

  The selection unit 132 selects the first selection node based on the number of output edges. The selection unit 132 selects a plurality of nodes as the first selection node in order from the node with the smallest number of output edges having the other node as an end point in the first graph. The selection unit 132 selects the second selection node based on the length of the output edge. The selection unit 132 selects, from among the target nodes, the second selection node in order from the node with the short output edge starting from the first selection node.

  In the example of FIG. 1, the selection unit 132 selects the first selection node in order from the node with the fewest output edges in the first graph data among the nodes in the second graph data. The selection unit 132 selects the first selection node based on the number of output edges of each of the nodes N1 to N12 as shown in the list information LT11.

  In the example of FIG. 1, in the update process in which the node N2 is the first selection node, the selection unit 132 has one output edge starting from the node N2 in the first graph data GR11. A node N1 is selected as the second selection node. In the update process in which the node N3 is the first selection node, the selection unit 132 has one output edge starting from the node N3 in the first graph data GR11. Select as a selection node. In the update process in which the node N4 is the first selection node, the selection unit 132 has one output edge starting from the node N4 in the first graph data GR11. Select as a selection node.

  In the example of FIG. 1, as illustrated in the list information LT11, the selection unit 132 selects a node N10 having the next smallest number of output edges (three) as the first selection node after the nodes N2 to N4. The selection unit 132 selects the second selection node using the three nodes N1, N11, and N12 to which the output edge from the node N10 is connected as target nodes. In the first graph data GR11, the selection unit 132 selects the edges E10-1, E11-1, and E12-1 starting from the node N10 as the second selection node in order from the node to which the short edge is connected. select.

  In the example of FIG. 1, as illustrated in the list information LT11, the selection unit 132 selects a node N5 having the next smallest number of output edges (five) as the first selection node after the node N10. The selection unit 132 selects the second selection node using the five nodes N1 and N6 to N9 to which the output edge from the node N5 is connected as target nodes. The selection unit 132 selects, as the second selection node, in order from the node to which the edge having the short length is connected among the edges E5-1 to E9-1 starting from the node N5 in the first graph data GR11.

(Generator 133)
The generation unit 133 generates various information. For example, the generation unit 133 generates various types of information (data) from information (data) stored in the storage unit 120. For example, the generation unit 133 generates various types of information from the object information storage unit 121, the base graph data storage unit 122, the first graph data storage unit 123, and the threshold information storage unit 124. For example, the generation unit 133 generates first graph data from the base graph data. For example, the generation unit 133 generates second graph data from the first graph data.

  For example, the generation unit 133 generates the first graph data GR11. The generation unit 133 is a neighborhood graph (base graph data GR10) generated based on a predetermined criterion, and is based on a neighborhood graph connected by a predetermined edge group that is a directed edge group connecting a plurality of nodes. First graph data GR11 is generated. The generating unit 133 generates first graph data GR11 that is a transposed graph of the base graph data GR10 by inverting the direction of each edge in the base graph data GR10.

  In addition, the generation unit 133 generates a second graph including only nodes by deleting the edge of the first graph data GR11. For example, the generation unit 133 generates the second graph data GR12-1 by deleting the edge of the first graph data GR11. The generation unit 133 generates second graph data using first graph data that is a transposed graph obtained by inverting the direction of each edge of the neighborhood graph.

  When the number of output edges starting from the first selection node in the second graph is less than a predetermined threshold (second threshold) in the second graph, the generation unit 133 sets the second selection starting from the first selection node in the first edge group. A second graph is generated by an update process in which an output edge having the selected node as an end point is added to the second graph as second edge information. In the second graph, the generation unit 133 has a number of input edges having the second selected node as an end point less than another threshold (first threshold), and the number of output edges having the first selected node as a starting point is a predetermined number. If it is less than the threshold value, a second graph with the second edge added is generated by the update process. In the second graph, the generation unit 133 has a number of input edges whose end point is the second selection node that is less than the first threshold value, and a number of output edges whose start point is the first selection node is less than the second threshold value. In the first edge group, the second graph (second graph data GR12− is updated by adding an output edge having the first selected node as the start point and the second selected node as the end point to the second graph as the second edge information. 4) is generated. For example, in the second graph data GR12-1 to GR12-3, the generation unit 133 has the number of input edges whose end point is the second selection node is less than the first threshold “4”, and the first selection node is the start point. When the number of output edges to be performed is less than the second threshold value “3”, output edges having the first selection node as the start point and the second selection node as the end point in the first graph data GR11 are set to the second graph data GR12-1 Second graph data GR12-4 is generated by update processing added to GR12-3 and the like.

  When the number of input edges having the second selection node as an end point is less than the first threshold and the number of output edges having the first selection node as a start point is less than the second threshold, the generation unit 133 generates the first selection node The second graph data GR12 is generated by the update process for adding the output edge from the node to the second selection node to the second graph data. The generation unit 133 repeats the update process for each combination of the selected first selection node and second selection node.

  The generation unit 133 repeats the update process for each combination of the first selection node selected by the selection unit 132 as the first selection unit and the second selection node selected by the selection unit 132 as the second selection unit. Thus, the second graph to which the second edge is added is generated. The generation unit 133 generates the second graph to which the second edge is added by repeating the update processing on the second edge information included in the second graph at the time of the update processing.

  When there is an input no-node that is a node to which the input edge is not connected among the plurality of nodes in the second graph, the generation unit 133 has a length of the input edges connected to the node in the first graph. A second graph to which the second edge is added is generated by an input addition process for adding an edge satisfying a predetermined criterion to the second edge as an input edge of the node. When there are no input nodes that are nodes whose input edges are not connected among a plurality of nodes in the second graph, the generation unit 133 has zero input edges due to the input addition process, that is, a node that has no input edges In order to eliminate this, an input edge having no input edge is added to the second graph. The generation unit 133 generates a second graph in which at least one input edge is connected to each node. The generation unit 133 eliminates the nodes that cannot be reached by following the edges during the search, and generates a second graph in which the second edges that can reach each node are added by following the edges.

  The generation unit 133 adds an edge having the shortest length among the input edges connected to the node corresponding to the no-input node in the first graph by an input addition process for adding the second edge as the input edge of the node, A second graph with the second edge added is generated. In the input addition process, the generation unit 133 can add an edge having the shortest length among input edges connected to the node in the first graph, that is, an edge between nodes having high similarity. Thereby, the generating apparatus 100 can generate graph data that enables an efficient search for a predetermined target.

  The generation unit 133 generates a second graph to which the second edge is added by performing an input addition process during the update process. The generation unit 133 generates a second graph in which at least one input edge is connected to each node in the update process. The generation unit 133 generates the second graph to which the second edge is added by performing the input addition process after the repetition of the update process. For example, the generation unit 133 generates the second graph to which the second edge is added by performing the input addition process after the end of the update process in which each node is selected as the first selection node. For example, the generation unit 133 adds the second edge by performing the input addition process after the completion of the all update process in which all the nodes that are selection candidates of the first selection node are selected as the first selection node. A second graph is generated. The generation unit 133 performs an input addition process only on a node (node with no input) having zero input edges after the end of the update process. The generation unit 133 performs an input addition process only for a node having zero input edges after the update process in which all nodes are repeated as the first selection node.

  In the example of FIG. 1, the generation unit 133 generates graph data GR12 by connecting nodes with directed edges by an update process. As illustrated in FIG. 1, the generating unit 133 adds only the edge determined to satisfy the condition based on the first threshold value and the second threshold value as the second edge to the second graph data by adding the second edge to the second graph data. Is generated.

  In the example of FIG. 1, the generation unit 133 generates first graph data from the base graph data. The generation unit 133 generates first graph data in which the direction of the edge of the base graph data is reversed by inverting the direction of the edge in the base graph data. The generation unit 133 generates first graph data in which only the direction of the edge is inverted without changing the distance between the nodes and the like. The generation unit 133 generates first graph data in which only the direction of the edge is inverted by using the input edge of one node in the base graph data as an output edge and the output edge of one node as an input edge.

  In the example of FIG. 1, the generation unit 133 generates first graph data GR11 as shown in the spatial information VS1-2 from the base graph data GR10. The generation unit 133 generates first graph data GR11 that is a transposed graph of the base graph data GR1. For example, the generation unit 133 uses the reference destination of each node shown in the base graph data storage unit 122 in FIG. First graph data shown in the one graph data storage unit 123 is generated. For example, the generation unit 133 reverses the edge E2 having the node N1 as the start point and the node N2 as the end point, thereby setting the node N2 as the start point and the node N1 as the end point as shown in the spatial information VS1-2. First graph data GR11 including the edge E2-1 is generated. For example, the generation unit 133 may generate a transposed graph of graph data by appropriately using various conventional techniques. The generation unit 133 may generate the first graph data GR11 by any process as long as the first graph data GR11 of the base graph data GR10 can be generated.

  In the example of FIG. 1, the generation unit 133 generates second graph data from the first graph data. First, the generation unit 133 generates second graph data obtained by removing an edge (first edge) from the first graph data. For example, the generation unit 133 generates second graph data having only nodes by removing an edge group (first edge group) from the first graph data. The generation unit 133 generates second graph data GR12-1 having only nodes from the first graph data GR11 as indicated by the spatial information VS1-3. The generation unit 133 starts from the second graph data GR12-1 having only nodes, and sequentially adds edges (second edges) that satisfy the condition to the second graph data. The generation unit 133 generates second graph data including the second edge group.

  In addition, in the second graph, the generation unit 133 has the number of input edges having the second selection node as an end point less than the first threshold value, and the number of output edges having the first selection node as the start point is less than the second threshold value. In some cases, an output edge starting from the first selection node and ending at the second selection node is added to the second graph. The generation unit 133 generates second graph data by repeating the update process for each first selection node.

  In the example of FIG. 1, the generation unit 133 has zero input edges with the node N1 that is the second selection node as the end point in the second graph data GR12-1, which is less than the first threshold “4”. It is determined that the number of output edges starting from the node N2 that is the first selection node is one less than the second threshold “3”. The generation unit 133 determines whether a condition based on a comparison between the number of edges and the first threshold value and the second threshold value is satisfied. If the generation unit 133 determines that the condition is satisfied, the generation unit 133 adds an edge corresponding to the determination target node to the second graph data. In the example of FIG. 1, the generation unit 133 determines that the condition based on the comparison between the first threshold value and the second threshold value is satisfied for the combination in which the first selection node is the node N2 and the second selection node is the node N1. Then, an edge E2-1 having the node N2 as the start point and the node N1 as the end point is added to the second graph data GR12.

  In the example of FIG. 1, the generation unit 133 is a second graph in which three edges (second edges) of edges E2-1, E3-1, and E4-1 are added, as indicated by the spatial information VS1-4. Data GR12-2 is generated. In addition, as illustrated in the spatial information VS1-5, the generation unit 133 adds the second graph data GR12-3 to which three edges (second edges) of the edges E10-1, E11-1, and E12-1 are added. Is generated.

  In the example of FIG. 1, the generation unit 133 has four input edges E2-1, E3-1, E4-1, and E10-1 of the node N1, which is the second selection node, and the first threshold value. It is determined that it is not less than “4”. The generation unit 133 determines that the combination in which the first selection node is the node N5 and the node N1 is the second selection node does not satisfy the condition based on the first threshold value. In this case, the generation unit 133 does not add the edge E5-1 having the node N5 as the start point and the node N1 as the end point to the second graph data GR12.

  In the example of FIG. 1, since the number of input edges whose end point is the node N9, which is the second selection node, is 0, which is less than the first threshold “4”, the generation unit 133 satisfies the first threshold condition. judge. On the other hand, the generation unit 133 has three output edges starting from the node N5 that is the first selection node, ie, the edges E6-1 to E8-1, and is not less than the second threshold “3”. It is determined that the two threshold conditions are not satisfied. In this case, the generation unit 133 does not add the edge E9-1 having the node N5 as the start point and the node N9 as the end point to the second graph data GR12.

(Search unit 134)
The search unit 134 searches various information. For example, the search unit 134 searches for an object by searching for graph data. For example, when the query acquired by the acquisition unit 131 is acquired, the search unit 134 searches for graph data and searches for an object similar to the query. For example, the search unit 134 extracts objects similar to the query by searching for graph data. For example, the search unit 134 extracts graph-like objects by searching for graph data based on the processing procedure as shown in FIG. The search unit 134 may not include the search unit 134 when the search service is not provided.

(Providing unit 135)
The providing unit 135 provides various types of information. For example, the providing unit 135 provides various types of information to the terminal device 10 and the information providing device 50. For example, the providing unit 135 provides an object ID corresponding to the query as a search result. For example, the providing unit 135 provides the object ID retrieved by the retrieving unit 134 to the information providing apparatus 50. For example, the providing unit 135 provides the information providing apparatus 50 with the object ID extracted by the searching unit 134 through the search. The providing unit 135 provides the object ID extracted by the search unit 134 to the information providing apparatus 50 as information indicating a vector corresponding to the query.

  The providing unit 135 may provide the second graph data generated by the generating unit 133 to an external information processing apparatus. For example, the providing unit 135 may transmit the second graph data GR12 generated by the generating unit 133 to the information providing apparatus 50.

[4. Generation process flow)
Next, the procedure of the generation process by the generation system 1 according to the embodiment will be described with reference to FIG. FIG. 9 is a flowchart illustrating an example of the generation process according to the embodiment.

  As illustrated in FIG. 9, the generation apparatus 100 acquires base graph data (step S101). For example, the generation device 100 acquires base graph data from the base graph data storage unit 122. In the example of FIG. 1, the generation device 100 acquires base graph data GR10.

  And the production | generation apparatus 100 produces | generates 1st graph data from base graph data (step S102). For example, the generating apparatus 100 generates first graph data in which the direction of the edge of the base graph data is reversed by inverting the direction of the edge in the base graph data. In the example of FIG. 1, the generation device 100 generates first graph data GR11 that is a transposed graph of the base graph data GR1 from the base graph data GR10.

  Then, the generation device 100 generates second graph data obtained by removing the edge from the first graph data (step S103). For example, the generating apparatus 100 generates second graph data having only nodes by removing edges from the first graph data. In the example of FIG. 1, the generation device 100 generates second graph data GR12-1 having only nodes from the first graph data GR11.

  Then, the generation apparatus 100 selects the first selection node based on the number of output edges in the first graph data among the nodes before selection (first unselected nodes) as the first selection node in the second graph data. (Step S104). For example, the generation apparatus 100 selects the node N10 having the fewest output edges (three) in the first graph data GR11 among the first unselected nodes (nodes N5 to N11) in the second graph data GR12-1. Select as one selection node.

  Then, the generation device 100 selects the second selection node based on the length of the output edge among the target nodes to which the output edge starting from the first selection node is connected in the first graph (step S105). . For example, when the first selection node is the node N10, the generation apparatus 100 selects the second selection node using the three nodes N1, N11, and N12 to which the output edge from the node N10 is connected as target nodes. . For example, in the first graph data GR11, the generation device 100 connects the edge having the shortest length (for example, the edge E10-1) among the edges E10-1, E11-1, and E12-1 starting from the node N10. The selected edge (for example, the node N1) is selected as the second selection node.

  Then, in the second graph, the generation device 100 has the number of input edges having the second selection node as the end point less than the first threshold value, and the number of output edges having the first selection node as the start point is less than the second threshold value. If there is, an output edge starting from the first selection node and ending at the second selection node is added to the second graph (step S106). For example, in the second graph data GR12-2, the generation apparatus 100 has the number of input edges having the second selected node as the end point being less than the first threshold “4”, and the output edge having the first selected node as the start point. When the number is less than the second threshold “3”, an output edge having the first selection node as a start point and the second selection node as an end point is added to the second graph. In the example of FIG. 1, the generation apparatus 100 includes three input edges whose end point is the node N1 that is the second selection node in the second graph data GR12-2 that is less than the first threshold “4”. The number of output edges starting from the node N10 that is the first selection node is zero, which is less than the second threshold “3”. Therefore, the generation device 100 determines that the condition based on the first threshold value and the second threshold value is satisfied, and adds the edge E10-1 having the node N10 as the start point and the node N1 as the end point to the second graph data GR12.

  Then, the generation device 100 determines whether there is an unselected target node as the second selection node (step S107). Specifically, the generation device 100 determines whether there is an unselected target node as the second selection node for the currently selected first selection node. That is, each node can be selected multiple times as a second selection node for a different first selection node. For example, the generation device 100 selects one node as a second selection node for different first selection nodes a plurality of times. For example, in the case of the node N1 in which the number of input edges in the first graph data GR11 is 5, the generation apparatus 100 sets the node N1 as the second selection node for five different first selection nodes (nodes N2 to N5 and N10). Select 5 times.

  If the generation apparatus 100 determines that there is an unselected target node as the second selection node (step S107: Yes), the generation apparatus 100 returns to step S105, and in step S105, the second selection node starts from the unselected target node. Select the selected node and repeat the subsequent processing. When determining that there is no unselected target node as the second selection node (step S107: No), the generation apparatus 100 determines whether there is an unselected node as the first selection node (step S108). .

  If it is determined that there is an unselected target node as the first selection node (step S108: Yes), the generation device 100 returns to step S104 and repeats the process. If the generation apparatus 100 determines that there is no unselected target node as the first selection node (step S108: No), the process ends.

[5. (When there is no input edge)
Next, the procedure of the input addition process by the generation system 1 according to the embodiment will be described with reference to FIG. FIG. 10 is a flowchart illustrating an example of the input addition process according to the embodiment.

  As illustrated in FIG. 10, the generation apparatus 100 selects a node having the number of input edges of 0 as a processing node (Step S201). For example, the generation device 100 selects a node as a processing node when the number of input edges of the node is zero after the end of the update process performed first as the first selection node. Also. For example, the generation apparatus 100 selects a node having 0 input edges as a processing node among nodes in the second graph data after the end of the update process. In the example of FIG. 11, the generation apparatus 100 selects, as a processing node, a node N5 having zero input edges among the nodes in the second graph data GR12-21 after the end of the updating process. For example, the generation apparatus 100 may include the nodes in the second graph data GR12-21 after the end of the update process in which all the nodes (nodes N1 to N12, etc.) in the second graph are selected as the first selection nodes. Among these, the node N5 having zero input edges is selected as a processing node.

  Then, the generation apparatus 100 adds, to the second graph data, an input edge whose length satisfies a predetermined condition among the input edges connected to the processing node in the first graph data (step S202). For example, the generation apparatus 100 adds the input edge having the shortest length among the input edges connected to the processing node in the first graph data to the second graph data. In the example of FIG. 11, the generation device 100 adds the shortest edge E51-1 among the edges E51-1 and E52-1 that end at the node N5 in the first graph data GR11. 22 is generated. In the example of FIG. 1, when performing the input addition process during the update process, the generation apparatus 100 has zero input edges of the node N1 after the update process update process with the node N1 as the first selection node. Therefore, the shortest edge E2-1 is added among the edges E2-1 to E5-1 whose end point is the node N1 in the graph data GR11.

[6. Example of input addition processing)
Here, the case where an input addition process is performed is demonstrated using FIG. FIG. 11 is a diagram illustrating an example of the input addition process according to the embodiment. Specifically, FIG. 11 shows a process of adding an input edge for a node having no input edge (zero lines). The description of the same points as in FIG. 1 will be omitted as appropriate. For example, the spatial information VS1 that is the vector space shown in FIG. 11 is the same as the spatial information VS1 in FIG. Further, for example, the first graph data GR11 shown in FIG. 11 is the same as the first graph data GR11 shown in FIG.

  In the example of FIG. 11, in order to simplify the description, in the second graph data GR12-21 in the spatial information VS1-22, at least one input edge is connected to all the nodes except the node N5. Indicates the case.

  First, the generation device 100 selects a node with the number of input edges being 0 as a processing node (step S21). In the example of FIG. 11, the generation apparatus 100 selects, as a processing node, a node N5 having zero input edges among the nodes in the second graph data GR12-21 after the end of the updating process.

  Then, the generation apparatus 100 adds, to the second graph data, an input edge whose length satisfies a predetermined condition among the input edges connected to the processing node in the first graph data (step S22). For example, the generation apparatus 100 adds the input edge having the shortest length among the input edges connected to the processing node in the first graph data to the second graph data. In the example of FIG. 11, it is assumed that the distance D51 is the smallest and the distance D52 is the largest of the distances D51 and D52, as shown in the list information LT21.

  Therefore, as illustrated in the list information LT21, the generation apparatus 100 adds the shortest edge E51-1 among the edges E51-1, E52-1 having the node N5 as the end point in the first graph data GR11. As a result, the generation apparatus 100 generates the second graph data GR12-22 without the node to which the edge E51-1 is added and the input edge becomes zero, as indicated by the spatial information VS1-22. In this way, the generation apparatus 100 eliminates the nodes that cannot be reached from other nodes, thereby suppressing the occurrence of a node (target) that is not searched during the search process using the second graph. It is possible to enable an efficient search for the target of the subject.

[7. Comparative example)
Here, a comparison with the result when the selection order of the first selection node is different is shown. FIG. 12 is a diagram illustrating a comparative example according to the embodiment. The description of the same points as in FIG. 1 will be omitted as appropriate. For example, the spatial information VS1 that is the vector space shown in FIG. 12 is the same as the spatial information VS1 shown in FIG. Further, for example, the second graph data GR12-4 shown in FIG. 12 is the same as the second graph data GR12-4 shown in FIG. 2, and the result when the first selection nodes are selected in order from the one with the fewest output edges. It is.

  On the other hand, the second graph data GR31 shown in FIG. 12 is a result when the first selection nodes are selected in order from the node having the smallest node ID. The second graph data GR31 illustrated in FIG. 12 indicates the second graph data generated when the first selection node is selected in order from the smallest “*” of “node N * (* is an arbitrary numerical value)”. .

  As shown in the second graph data GR31, when the node IDs are selected in ascending order, the node N5 is selected as the first selection node before the node N10. Therefore, an edge E5-1 is added to the second graph data. As a result, the edge E10-1 and the edge E7-1 added in the second graph data GR12-4 are not added. As described above, in the second graph data GR31, since the edge E7-1 is not added, it may be difficult to increase the input edges (inward edges) of each node.

  On the other hand, in the second graph data GR12-4 shown in FIG. 12, the edge E7-1 which is an inward edge is also connected to the node N7, and the inward edge can be connected to as many nodes as possible.

[8. Search example)
Here, an example of the search using the graph data described above is shown. The search using the generated graph data is not limited to the following, and may be performed by various procedures. This point will be described with reference to FIG. FIG. 13 is a flowchart illustrating an example of search processing using graph data. The search process described below is performed by the search unit 134 of the generation device 100. Further, the object described below may be read as a node. In the following description, it is assumed that the generation apparatus 100 (search unit 134) performs a search process, but the search process may be performed by another apparatus. In this case, the generation apparatus 100 may not include the search unit 134.

  Here, the neighboring object set N (G, y) is a set of neighboring objects that are related by the edge given to the node y. “G” may be predetermined graph data (for example, the second graph data GR12 or the like). For example, the generation device 100 performs k neighborhood search processing.

  For example, the generating apparatus 100 sets the radius r of the hypersphere to ∞ (infinity) (step S300), and extracts the subset S from the existing object set (step S301). For example, the generation apparatus 100 may extract the object (node) selected as the root node as the subset S. Further, for example, a hypersphere is a virtual sphere that indicates a search range. It should be noted that the objects included in the object set S extracted in step S301 are also included in the initial set of the search result object set R at the same time.

  Next, the generation apparatus 100 extracts an object having the shortest distance from the object y from the objects included in the object set S, where y is a search query object, and sets the object as an object s (step S302). For example, when only the object (node) selected as the root node is an element of S, the generation apparatus 100 extracts the root node as the object s as a result. Next, the generating apparatus 100 excludes the object s from the object set S (Step S303).

  Next, the generating apparatus 100 determines whether or not the distance d (s, y) between the object s and the object y exceeds r (1 + ε) (step S304). Here, ε is an expansion element, and r (1 + ε) is a value indicating the radius of the search range (search only for nodes within this range. The accuracy can be increased by making it larger than the search range). is there. When the distance d (s, y) between the object s and the object y exceeds r (1 + ε) (step S304: Yes), the generation device 100 outputs the object set R as a neighborhood object set of the object y (step S305). ), The process is terminated.

  When the distance d (s, y) between the object s and the search query object y does not exceed r (1 + ε) (step S304: No), the generation device 100 includes the neighborhood object set N (G, s) of the object s. One object that is not included in the object set C is selected from the objects that are elements, and the selected object u is stored in the object set C (step S306). The object set C is provided for convenience in order to avoid duplicate search, and is set to an empty set at the start of processing.

  Next, the generating apparatus 100 determines whether or not the distance d (u, y) between the object u and the object y is equal to or less than r (1 + ε) (step S307). When the distance d (u, y) between the object u and the object y is equal to or less than r (1 + ε) (step S307: Yes), the generation apparatus 100 adds the object u to the object set S (step S308). In addition, when the distance d (u, y) between the object u and the object y is not equal to or less than r (1 + ε) (step S307: No), the generation apparatus 100 performs the determination (processing) in step S309.

  Next, the generating apparatus 100 determines whether or not the distance d (u, y) between the object u and the object y is equal to or less than r (step S309). When the distance d (u, y) between the object u and the object y exceeds r, the generation device 100 performs determination (processing) in step S315. Further, when the distance d (u, y) between the object u and the object y is not less than or equal to r (step S309: No), the generation device 100 performs the determination (processing) in step S315.

  When the distance d (u, y) between the object u and the object y is equal to or less than r (step S309: Yes), the generation apparatus 100 adds the object u to the object set R (step S310). Then, the generation device 100 determines whether or not the number of objects included in the object set R exceeds ks (step S311). The predetermined number ks is a natural number that is arbitrarily determined. For example, ks = 2 may be sufficient. When the number of objects included in the object set R does not exceed ks (step S311: No), the generation device 100 performs the determination (processing) in step S313.

  When the number of objects included in the object set R exceeds ks (step S <b> 311: Yes), the generation apparatus 100 selects an object having the longest (far) object from the object y as the object included in the object set R. Exclude from set R (step S312).

  Next, the generating apparatus 100 determines whether or not the number of objects included in the object set R matches ks (step S313). When the number of objects included in the object set R does not match ks (step S313: No), the generation device 100 performs determination (processing) in step S315. When the number of objects included in the object set R matches ks (step S313: Yes), the generation apparatus 100 has the longest (distant) object from the object y among the objects included in the object set R. And the distance from the object y are set to a new r (step S314).

  Then, the generation apparatus 100 determines whether all objects have been selected from the objects that are elements of the neighboring object set N (G, s) of the object s and stored in the object set C (step S315). If all the objects have not been selected and stored in the object set C from the objects that are elements of the neighboring object set N (G, s) of the object s (step S315: No), the generating apparatus 100 proceeds to step S306. Return and repeat the process.

  When all the objects are selected from the objects that are elements of the neighboring object set N (G, s) of the object s and stored in the object set C (step S315: Yes), the generation apparatus 100 determines that the object set S is It is determined whether it is an empty set (step S316). When the object set S is not an empty set (step S316: No), the generation device 100 returns to step S302 and repeats the process. If the object set S is an empty set (step S316: Yes), the generating apparatus 100 outputs the object set R and ends the process (step S317). For example, the generation device 100 may provide an object (node) included in the object set R as a search result corresponding to the search query (input object y) to the terminal device that has performed the search.

[9. effect〕
As described above, the generation device 100 according to the embodiment includes the acquisition unit 131, the first selection unit (selection unit 132 in the embodiment), the second selection unit (selection unit 132 in the embodiment), and the generation unit. 133. The acquisition unit 131 includes a first graph including information on a plurality of nodes that are targets of data search and information on a first edge group that is a directed edge group that connects the plurality of nodes (in the embodiment, “ First graph data GR11 ". The same applies hereinafter) and a second graph including information on a plurality of nodes (" second graph data GR12-1 "in the embodiment). The selection unit 132 serving as a first selection unit selects a first selection node based on the number of output edges having other nodes as end points in the first graph among the plurality of nodes. The selection unit 132 as the second selection unit selects a second selection node based on the length of the output edge among the target nodes to which the output edge starting from the first selection node is connected in one graph. When the number of output edges starting from the first selected node in the second graph is less than a predetermined threshold value (“second threshold value” in the embodiment, the same applies hereinafter) in the second graph, the generating unit 133 By the update process for adding the output edge having the first selection node as the start point and the second selection node as the end point to the second graph as the second edge information, the second graph (in the embodiment, “second graph data GR12-4”) is added. ) Is generated.

  As described above, the generation device 100 according to the embodiment selects the first selection node based on the number of output edges, selects the second selection node based on the length of the output edge, and starts the first selection node. Is less than the second threshold value (“3” in the embodiment), the update processing for adding the output edge from the first selection node to the second selection node to the second graph data is performed by the second update process. Graph data GR12 can be generated. Therefore, the generation device 100 can generate graph data that enables an efficient search for a predetermined target.

  Further, in the generation device 100 according to the embodiment, the generation unit 133 includes other threshold values (in the embodiment, “first threshold value”; the same applies hereinafter) with the number of input edges having the second selection node as the end point in the second graph. When the number of output edges starting from the first selected node is less than a predetermined threshold, a second graph with the second edge added is generated by the update process.

  As described above, the generation apparatus 100 according to the embodiment has an output whose number of input edges having the second selection node as the end point is less than the first threshold (in the embodiment, “4”) and whose first selection node is the start point. When the number of edges is less than the second threshold value, the second graph data GR12 can be generated by the update process for adding the output edge from the first selection node to the second selection node to the second graph data. Therefore, the generation device 100 can generate graph data that enables an efficient search for a predetermined target.

  Further, in the generation device 100 according to the embodiment, the acquisition unit 131 is a neighborhood graph (in the embodiment, “base graph data GR10”) generated based on a predetermined criterion, and connects a plurality of nodes. A first graph which is a transposed graph including information on a first edge group obtained by inverting the direction of each edge of a predetermined edge group among neighboring graphs connected by the predetermined edge group which is a directed edge group is acquired. .

  As described above, the generation apparatus 100 according to the embodiment generates the second graph data using the first graph data that is the transposed graph obtained by inverting the direction of each edge of the neighborhood graph, thereby improving the efficiency related to the predetermined target. Graph data can be generated to enable efficient search.

  In addition, in the generation device 100 according to the embodiment, the selection unit 132 as the first selection unit is the first in the order from the node with the fewest number of output edges having the other node as an end point in the first graph among the plurality of nodes. Select as a selection node.

  As described above, the generation apparatus 100 according to the embodiment selects a predetermined number of nodes as a first selection node in order from a node having a small number of output edges whose end points are other nodes in the first graph. It is possible to generate graph data that enables an efficient search for the target of the subject.

  Further, in the generation device 100 according to the embodiment, the selection unit 132 as the second selection unit is set as the second selection node in order from the node with the short output edge starting from the first selection node to the target node among the target nodes. select.

  As described above, the generation apparatus 100 according to the embodiment relates to a predetermined target by selecting, from among the target nodes, the second selection node in order from the node with the shortest output edge starting from the first selection node. Graph data that enables efficient search can be generated.

  In the generation device 100 according to the embodiment, the generation unit 133 includes a first selection node selected by the selection unit 132 as the first selection unit, and a second selected by the selection unit 132 as the second selection unit. By repeating the update process for each combination with the selected node, a second graph with the second edge added is generated.

  As described above, the generation apparatus 100 according to the embodiment relates to a predetermined target by generating the second graph by repeating the update process for each combination of the selected first selection node and second selection node. Graph data that enables efficient search can be generated.

  Further, in the generation device 100 according to the embodiment, the generation unit 133 repeats the update process for the second edge information included in the second graph at the time of the update process, thereby adding the second edge. Two graphs are generated.

  As described above, the generation device 100 according to the embodiment repeats the generation process for the second edge information included in the second graph at the time of the update process, thereby generating the second graph with the second edge added. By generating, it is possible to generate graph data that enables an efficient search for a predetermined object.

  Further, in the generation device 100 according to the embodiment, when there is an input no-node that is a node to which the input edge is not connected among the plurality of nodes in the second graph, the generation unit 133 determines the node in the first graph Among the connected input edges, the second graph with the second edge added is generated by an input addition process for adding an edge whose length satisfies a predetermined criterion to the second edge as the input edge of the node.

  As described above, the generation apparatus 100 according to the embodiment, when there is no input node among the plurality of nodes in the second graph, eliminates a node having zero input edges, that is, no input edge by the input addition process. . That is, the generation apparatus 100 can generate a second graph in which at least one input edge is connected to each node, eliminates nodes that cannot be reached by following the edges at the time of search, and traces each node by following the edges. Can be reached. Thereby, the generating apparatus 100 can generate graph data that enables an efficient search for a predetermined target.

  Further, in the generation device 100 according to the embodiment, the generation unit 133 adds the edge having the shortest length among the input edges connected to the node in the first graph to the second edge as the input edge of the node. The second graph to which the second edge is added is generated by the input adding process.

  As described above, in the input addition process, the generation device 100 according to the embodiment uses the edge having the shortest length among the input edges connected to the node in the first graph as the second edge as the input edge of the node. By adding to, an edge between nodes having high similarity can be added. Thereby, the generating apparatus 100 can generate graph data that enables an efficient search for a predetermined target.

  In the generation device 100 according to the embodiment, the generation unit 133 generates a second graph to which the second edge is added by performing an input addition process during the update process.

  As described above, the generation apparatus 100 according to the embodiment performs the input addition process at the time of the update process, so that there is a possibility that a node having zero input edges may remain when each node is subjected to the update process. By eliminating the second graph, the second graph with the second edge added is generated. That is, the generation apparatus 100 can generate a second graph in which at least one or more input edges are connected to each node in the update process, eliminate the nodes that cannot be reached by following the edges during the search, Each node can be reached by tracing. As a result, the generation device 100 can generate graph data that enables an efficient search for a predetermined target while suppressing an increase in processing time for generating the second graph.

  In the generation device 100 according to the embodiment, the generation unit 133 generates the second graph by performing the input addition process after the repetition of the update process.

  As described above, the generation apparatus 100 according to the embodiment performs the input addition process after the repetition of the update process, so that the number of input edges becomes 0 after the update process for each node is performed. Input addition processing is performed only for the existing node. Accordingly, since the generation apparatus 100 can target only the node having zero input edges after the update process, the generation apparatus 100 can reduce the number of edges added by the input addition process. Can do. That is, the generation apparatus 100 can generate the second graph in which at least one input edge is connected to each node while reducing the number of edges added by the input addition process, and the edge is searched at the time of the search. Nodes that cannot be reached by tracing are eliminated, and each node can be reached by following an edge. Accordingly, the generation device 100 can generate graph data that enables efficient search related to a predetermined target while suppressing an increase in the number of edges.

[10. Hardware configuration)
The generation apparatus 100 according to the above-described embodiment is realized by a computer 1000 configured as shown in FIG. 14, for example. FIG. 14 is a hardware configuration diagram illustrating an example of a computer that implements the function of the generation device. The computer 1000 includes a CPU 1100, a RAM 1200, a ROM (Read Only Memory) 1300, an HDD (Hard Disk Drive) 1400, a communication interface (I / F) 1500, an input / output interface (I / F) 1600, and a media interface (I / F). ) 1700.

  The CPU 1100 operates based on a program stored in the ROM 1300 or the HDD 1400 and controls each unit. The ROM 1300 stores a boot program executed by the CPU 1100 when the computer 1000 is started up, a program depending on the hardware of the computer 1000, and the like.

  The HDD 1400 stores programs executed by the CPU 1100, data used by the programs, and the like. The communication interface 1500 receives data from other devices via the network N and sends the data to the CPU 1100, and transmits data generated by the CPU 1100 to other devices via the network N.

  The CPU 1100 controls an output device such as a display and a printer and an input device such as a keyboard and a mouse via the input / output interface 1600. The CPU 1100 acquires data from the input device via the input / output interface 1600. In addition, the CPU 1100 outputs the generated data to the output device via the input / output interface 1600.

  The media interface 1700 reads a program or data stored in the recording medium 1800 and provides it to the CPU 1100 via the RAM 1200. The CPU 1100 loads such a program from the recording medium 1800 onto the RAM 1200 via the media interface 1700, and executes the loaded program. The recording medium 1800 is, for example, an optical recording medium such as a DVD (Digital Versatile Disc) or PD (Phase change rewritable disk), a magneto-optical recording medium such as an MO (Magneto-Optical disk), a tape medium, a magnetic recording medium, or a semiconductor memory. Etc.

  For example, when the computer 1000 functions as the generation apparatus 100 according to the embodiment, the CPU 1100 of the computer 1000 implements the function of the control unit 130 by executing a program loaded on the RAM 1200. The CPU 1100 of the computer 1000 reads these programs from the recording medium 1800 and executes them. However, as another example, these programs may be acquired from other devices via the network N.

  As described above, some of the embodiments of the present application have been described in detail with reference to the drawings. However, these are merely examples, and various modifications based on the knowledge of those skilled in the art, including the aspects described in the disclosure line of the invention. It is possible to implement the present invention in other forms with improvements.

[11. Others]
In addition, among the processes described in the above embodiment, all or part of the processes described as being automatically performed can be performed manually, or the processes described as being performed manually can be performed. All or a part can be automatically performed by a known method. In addition, the processing procedures, specific names, and information including various data and parameters shown in the document and drawings can be arbitrarily changed unless otherwise specified. For example, the various types of information illustrated in each drawing is not limited to the illustrated information.

  Further, each component of each illustrated apparatus is functionally conceptual, and does not necessarily need to be physically configured as illustrated. In other words, the specific form of distribution / integration of each device is not limited to that shown in the figure, and all or a part thereof may be functionally or physically distributed or arbitrarily distributed in arbitrary units according to various loads or usage conditions. Can be integrated and configured.

  In addition, the processes described in the embodiments described above can be appropriately combined within a range in which the process contents are not contradictory.

  In addition, the “section (module, unit)” described above can be read as “means” or “circuit”. For example, the acquisition unit can be read as acquisition means or an acquisition circuit.

DESCRIPTION OF SYMBOLS 1 Generation system 100 Generation apparatus 121 Object information storage part 122 Base graph data storage part 123 1st graph data storage part 124 Threshold information storage part 125 2nd graph data storage part 130 Control part 131 Acquisition part 132 Selection part (1st selection part) , Second selection unit)
133 Generating Unit 134 Searching Unit 135 Providing Unit 10 Terminal Device 50 Information Providing Device N Network

Claims (13)

A first graph including information on a plurality of nodes to be searched for data and information on a first edge group which is a directed edge group connecting the plurality of nodes; and a first graph including information on the plurality of nodes. An acquisition unit for acquiring two graphs;
A first selection unit that selects a first selection node among the plurality of nodes based on the number of output edges whose end points are other nodes in the first graph;
A second selection unit that selects a second selection node based on a length of the output edge among target nodes to which an output edge starting from the first selection node is connected in the first graph;
In the second graph, when the number of output edges starting from the first selection node is less than a predetermined threshold, the first selection node is set as the start point and the second selection node is set as the end point in the first edge group. Generating a second graph to which the second edge is added by an update process for adding the output edge to the second graph as second edge information;
A generating apparatus comprising: The generator is
In the second graph, when the number of input edges having the second selection node as an end point is less than another threshold value, and the number of output edges having the first selection node as a start point is less than the predetermined threshold value, The generation apparatus according to claim 1, wherein the update process generates a second graph to which the second edge is added. The acquisition unit
Each of the edges of the predetermined edge group among the vicinity graphs generated based on a predetermined criterion and connected by a predetermined edge group that is a directed edge group connecting the plurality of nodes. The generation apparatus according to claim 1, wherein the first graph, which is a transposed graph including information on the first edge group in which the orientation of the first edge group is inverted, is acquired. The first selection unit includes:
The node is selected as a first selection node in order from a node having a small number of output edges whose end points are other nodes in the first graph among the plurality of nodes. The generating device described in 1. The second selection unit includes:
5. The selection node according to claim 1, wherein, from among the target nodes, nodes are selected as a second selection node in order from a node with a short output edge starting from the first selection node. 6. Generator. The generator is
The second edge is added by repeating the updating process for each combination of the first selection node selected by the first selection unit and the second selection node selected by the second selection unit. The 2nd graph is produced | generated. The production | generation apparatus of any one of Claims 1-5 characterized by the above-mentioned. The generator is
The second graph to which the second edge is added is generated by repeating the update process for the information on the second edge included in the second graph at the time of the update process. Item 7. The generating device according to Item 6. The generator is
When there is an input non-node that is a node to which an input edge is not connected among the plurality of nodes in the second graph, a length of the input edge connected to the node in the first graph is a predetermined length. 8. The second graph to which the second edge is added is generated by an input addition process for adding an edge satisfying a criterion to the second edge as an input edge of the node. The generating device according to claim 1. The generator is
Of the input edges connected to the node corresponding to the no-input node in the first graph, the input adding process of adding the shortest edge as the input edge of the node to the second edge The generation apparatus according to claim 8, wherein the second graph to which two edges are added is generated. The generator is
10. The generation apparatus according to claim 8, wherein the second graph to which the second edge is added is generated by performing the input addition process at the time of the update process. The generator is
The generation apparatus according to claim 8 or 9, wherein the second graph to which the second edge is added is generated by performing the input addition processing after the end of the update processing. A generation method executed by a computer,
A first graph including information on a plurality of nodes to be searched for data and information on a first edge group which is a directed edge group connecting the plurality of nodes; and a first graph including information on the plurality of nodes. An acquisition process for acquiring two graphs;
A first selection step of selecting a first selection node among the plurality of nodes based on the number of output edges whose end points are other nodes in the first graph;
A second selection step of selecting a second selection node based on the length of the output edge among the target nodes to which the output edge starting from the first selection node is connected in the first graph;
In the second graph, when the number of output edges starting from the first selection node is less than a predetermined threshold, the first selection node is set as the start point and the second selection node is set as the end point in the first edge group. Generating the second graph by an update process for adding the output edge to the second graph as the second edge information;
A generation method comprising: A first graph including information on a plurality of nodes to be searched for data and information on a first edge group which is a directed edge group connecting the plurality of nodes; and a first graph including information on the plurality of nodes. An acquisition procedure for acquiring two graphs;
A first selection procedure for selecting a first selection node based on the number of output edges whose end points are other nodes in the first graph among the plurality of nodes;
A second selection procedure for selecting a second selection node based on the length of the output edge among the target nodes to which the output edge starting from the first selection node is connected in the first graph;
In the second graph, when the number of output edges starting from the first selection node is less than a predetermined threshold, the first selection node is set as the start point and the second selection node is set as the end point in the first edge group. A generation procedure for generating the second graph by an update process for adding the output edge to the second graph as second edge information;
A program for causing a computer to execute.

Images