Information processing apparatus

Abstract

The present application relates to an information processing apparatus. The information processing apparatus includes a first generation unit that generates a new graph structure based on graph information related to a plurality of graph structures, the plurality of graph structures being respectively generated based on a plurality of image data registered in a database (DB), the new image structure having a deviation amount from the plurality of graph structures of a first predetermined value or more; and a second generation unit that generates new image data based on the new graph structure.

Description

Technical Field

[0001] This invention relates to the technical field of information processing devices. Background Technology

[0002] As such a device, for example, a system has been proposed that enables Large Language Models (LLMs) to generate document-based query data, and uses the pairing of documents and query data for learning a retrieval model for a chatbot (see Japanese Patent Application Publication No. 2023-076413). Summary of the Invention

[0003] Large language models are language models built using massive datasets and deep learning techniques. Sometimes, it's difficult to collect large amounts of the data used for learning such models (i.e., learning data). Therefore, a data expansion technique has been proposed that artificially generates new data by modifying existing data. On the other hand, when modifications to existing data are manually set, the human workload is quite heavy.

[0004] The present invention was made in view of the above circumstances, and its object is to provide an information processing apparatus capable of generating new data.

[0005] An information processing apparatus according to one aspect of the present invention includes: a first generation unit for generating a new graph structure based on graphic information relating to a plurality of graph structures, the plurality of graph structures being generated based on a plurality of image data registered in a database, the deviation of the new graph structure from the plurality of graph structures being greater than or equal to a first predetermined value; and a second generation unit for generating new image data based on the new graph structure. Attached Figure Description

[0006] The features, advantages, and technical and industrial significance of exemplary embodiments of the present invention will now be described with reference to the accompanying drawings, in which like reference numerals denote like elements, and wherein:

[0007] Figure 1 This is a block diagram illustrating an example of the structure of an information processing apparatus according to an embodiment;

[0008] Figure 2 This is a block diagram illustrating an example of the structure of the computing device according to an embodiment;

[0009] Figure 3 This is a conceptual diagram illustrating the operation of an information processing device according to an embodiment. Detailed Implementation

[0010] Reference Figures 1 to 3 This describes the implementation method of the information processing device. Figure 1In this device, the information processing apparatus 10 includes an arithmetic unit 11, a storage unit 12, a communication unit 13, an input unit 14, and an output unit 15. The arithmetic unit 11, the storage unit 12, the communication unit 13, the input unit 14, and the output unit 15 are connected via a data bus 16.

[0011] The arithmetic unit 11 may have a processor. Furthermore, the arithmetic unit 11 may have a single processor or multiple processors. That is, the arithmetic unit 11 may have more than one processor. Additionally, the processor may be a multi-core processor. In the case where the arithmetic unit 11 has a single processor that functions as a multi-core processor, it can be said that the arithmetic unit 11 logically has multiple processors.

[0012] The processor may be at least one of a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a FPGA (Field Programmable Gate Array), and a TPU (Tensor Processing Unit).

[0013] The storage device 12 may be at least one of RAM (Random Access Memory), ROM (Read Only Memory), hard disk drive, optical disk drive, SSD (Solid State Drive), and optical disk array. That is, the storage device 12 may be implemented by a single device or by multiple devices.

[0014] The communication device 13 can communicate with external devices of the information processing device 10. Furthermore, the communication device 13 can perform both wired and wireless communication.

[0015] Input device 14 is a device capable of receiving information input to information processing device 10 from an external source. Input device 14 may include user-operable devices (e.g., keyboard, mouse, touch panel, etc.) of information processing device 10. Input device 14 may include a recording medium reading device capable of reading information recorded on a recording medium such as a USB (Universal Serial Bus) memory, which can be attached to and detached from information processing device 10. Additionally, when information is input to information processing device 10 via communication device 13 (in other words, when information processing device 10 obtains information via communication device 13), communication device 13 can function as an input device.

[0016] Output device 15 is a device capable of outputting information to the outside of information processing device 10. Output device 15 includes a display device 151 capable of outputting visual information such as text and images. Alternatively, output device 15 may also include a speaker capable of outputting auditory information such as sound. Output device 15 may include a vibration motor capable of outputting tactile information such as vibration as the aforementioned information. Output device 15 may also include a printer. Output device 15 may also be a device capable of outputting information to a recording medium removable from information processing device 10, such as a USB memory. Furthermore, when information processing device 10 outputs information via communication device 13, communication device 13 can function as an output device.

[0017] Storage device 12 is capable of storing desired data. The computer program CP executed by the arithmetic unit 11 can be stored in storage device 12. Storage device 12 can temporarily store data temporarily used by the arithmetic unit 11 while the arithmetic unit 11 is executing the computer program CP.

[0018] Furthermore, the computer program CP can also be recorded on a computer-readable and non-transitory recording medium. In this case, the recording medium can be read using a recording medium reading device (not shown) included in the information processing apparatus 10, thereby storing the computer program CP in the storage device 12. Additionally, at least one of optical discs, magnetic media, optical disks, semiconductor memory, and other media capable of storing programs can be used as the recording medium. Furthermore, the computer program CP can also be obtained from an external (not shown) device of the information processing apparatus 10 via the communication device 13. In other words, the computer program CP can be downloaded from an external device to the storage device 12 of the information processing apparatus 10.

[0019] The arithmetic unit 11 (e.g., a processor) can perform the processing to be performed by the information processing unit 10 together with the storage device 12 storing the computer program CP (in other words, together with the storage device 12 and the computer program CP stored in the storage device 12). For example, the arithmetic unit 11 may implement logical function blocks for performing the processing to be performed by the information processing unit 10 within the arithmetic unit 11 (e.g., within the processor) by executing the computer program CP.

[0020] like Figure 2As shown, the arithmetic unit 11 includes a first generation unit 111 and a second generation unit 112. The first generation unit 111 and the second generation unit 112 can also be implemented as the aforementioned logic function blocks. Alternatively, at least one of the first generation unit 111 and the second generation unit 112 can also be implemented as a physical processing circuit. At least one of the first generation unit 111 and the second generation unit 112 can be implemented in a form where logic function blocks and physical processing circuits coexist.

[0021] Reference Figure 3 The operation of the information processing device 10 configured as described above will be explained. Figure 3 In the database DB, there are multiple image data sets Imgs. Suppose that for each of the multiple image data sets Imgs, there exists graph information GSI related to multiple graph constructions generated from each of the multiple image data sets Imgs. Graph information GSI can represent information about the multiple graph constructions themselves, or it can represent information about the feature vectors after the multiple graph constructions have been vectorized. Graph information GSI can also be a distribution graph representing the distribution of multiple feature quantities involved in the multiple graph constructions in the feature space.

[0022] Furthermore, graph information (GSI) can be registered in a database (DB) or stored in a device different from the database (DB). Additionally, graph constructions can represent data consisting of node groups and edge groups, where node groups represent the relationships between each part of an object within an image related to image data, and edge groups represent the relationships between nodes. Furthermore, various existing methods can be applied to methods for generating graph constructions from image data. Therefore, a detailed description of methods for generating graph constructions from image data will be omitted. Furthermore, learning models (e.g., Graph Neural Networks: GNNs) can be used to compute the features involved in the graph construction. "Features involved in the graph construction" can refer to the features involved in the entire graph construction or the features involved in each constituent element (i.e., node) included in the graph construction.

[0023] In order to generate new image data (e.g., image data Img) using multiple image data Imgs registered in the database DB, the information processing device 10 may perform the following processing.

[0024] The first generation unit 111 of the information processing apparatus 10 can generate a new graph structure (e.g., graph structure GS) based on the graph information GSI, wherein the deviation from the multiple graph structures involved in the multiple image data Imgs is greater than or equal to a first predetermined value. Furthermore, the first generation unit 111 can also generate a new graph structure based on the graph information GSI, wherein the deviation from the aforementioned multiple graph structures is greater than or equal to a first predetermined value and less than a second predetermined value.

[0025] For example, the first generation unit 111 can determine (or estimate) the aforementioned deviation amount by calculating the distance between each of the plurality of graph structures and the candidate new graph structure. For example, the first generation unit 111 can determine (or estimate) the aforementioned deviation amount based on the feature vectors involved in each of the plurality of graph structures and the feature vectors involved in the candidate new graph structure. For example, the first generation unit 111 can determine (or estimate) the aforementioned deviation amount based on at least one of the category, positional relationship, shape, and color of objects in each of the plurality of graph structures and at least one of the category, positional relationship, shape, and color of objects in the candidate new graph structure.

[0026] For example, if the graphic information GSI is the distribution map described above, the first generation unit 111 can generate a new graph structure by including constituent elements corresponding to the blank areas of the distribution map (i.e., areas where no data points representing feature quantities exist). In this case, the first generation unit 111 can determine (or estimate) the deviation amount using any of the methods described above. Furthermore, since the first generation unit 111 generates a new graph structure, it can be referred to as a graph structure generation unit or a structure generation unit.

[0027] For example, if the dataset used for model learning includes a large number of similar data points, resulting in a bias in the data distribution, the quality of the dataset will be degraded. In this embodiment, to suppress such a quality degradation, the first generation unit 111 can generate a new graph structure whose deviation from the multiple graph structures involved in the multiple image data Imgs is greater than or equal to a first predetermined value. That is, the first predetermined value can be considered as a value used to suppress the bias in the data distribution of the dataset. The first predetermined value can be a fixed value or a variable value corresponding to any parameter.

[0028] For example, if the dataset used for model learning contains a large number of outliers or anomalies, the accuracy of the model learned using that dataset may be low. In this embodiment, in order to suppress the generation of outliers and anomalies, the first generation unit 111 can generate a new graph structure whose deviation from the multiple graph structures involved in the multiple image data Imgs is less than a second predetermined value. That is, the second predetermined value can be said to be a value used to suppress the generation of outliers and anomalies. The second predetermined value can be a fixed value or a variable value corresponding to any parameter.

[0029] The second generation unit 112 of the information processing apparatus 10 can generate new image data (e.g., image data Img) based on a new graph structure (e.g., graph structure GS) generated by the first generation unit 111. That is, the second generation unit 112 can generate new image data such that the graph structure generated based on the new image data is the same as the new graph structure generated by the first generation unit 111. Furthermore, if the second generation unit 112 is input with a new graph structure generated by the first generation unit 111, it can use a learning model (e.g., image generation AI (Artificial Intelligence)) to generate new image data. Since the second generation unit 112 generates new image data, it can also be referred to as an image generation unit.

[0030] The arithmetic unit 11 of the information processing apparatus 10 can register new image data (e.g., image data Img) generated by the second generation unit 112 into the database DB. Furthermore, the arithmetic unit 11 of the information processing apparatus 10 can control the display device 151 to display the image related to the new image data (e.g., image data Img) generated by the second generation unit 112. A user of the information processing apparatus 10 can instruct via the input device 14 whether the image data related to the image to be displayed on the display device 151 should be registered in the database 40.

[0031] Technical effect

[0032] In this embodiment, the first generation unit 111 generates a new image structure based on the graphic information GSI. The second generation unit 112 generates new image data based on the generated new image structure. That is, the information processing device 10 according to this embodiment can generate new image data. Here, in this embodiment, the information processing device 10 automatically generates new image data based on the graphic information GSI. Therefore, according to the information processing device 10, the workload of humans can be reduced.

[0033] The present invention, derived from the embodiments described above, is described below.

[0034] An information processing apparatus according to one aspect of the invention includes: a first generation unit that generates a new graph structure based on graphic information related to a plurality of graph structures, the plurality of graph structures being generated separately based on a plurality of image data registered in a database, the deviation of the new graph structure from the plurality of graph structures being a first predetermined value or more; and a second generation unit that generates new image data based on the new graph structure. In the above embodiment, "first generation unit 111" is an example of "first generation unit," and "second generation unit 112" is an example of "second generation unit."

[0035] In the information processing apparatus described above, the graphic information may be a distribution map, in which the distribution of the plurality of graphic structures is mapped, and the first generation unit may generate a graphic structure including constituent elements corresponding to the blank parts of the distribution map as the new graphic structure. If configured in this way, it is relatively easy to generate image data that differs from multiple image data already registered in the database.

[0036] In the information processing apparatus described above, the first generation unit may generate a new graph structure based on the graphic information, wherein the deviation is greater than or equal to a first predetermined value and less than a second predetermined value. If configured in this way, it is possible to suppress the generation of image data that corresponds to outliers or anomalies.

[0037] In the information processing apparatus described above, the graphic information may include information representing at least one of the categories, positional relationships, shapes, and colors of objects in the graphic structure, and the first generation unit determines the deviation amount based on at least one of the categories, positional relationships, shapes, and colors of objects in the graphic structure. If configured in this way, the deviation amount can be determined relatively easily.

[0038] This invention is not limited to the embodiments described above. Appropriate modifications may be made without departing from the spirit or spirit of the invention as read in its entirety from the claims and description, and are included within the scope of the information processing apparatus or the technology of this invention accompanying such modifications.

Claims

1. An information processing device, comprising: The first generation unit generates a new graph construction based on graph information related to multiple graph constructions, wherein... The multiple graph structures are generated separately based on multiple image data registered in the database, and the deviation of the new graph structure from the multiple graph structures is greater than or equal to a first predetermined value. as well as The second generation unit generates new image data based on the new graph structure.

2. The information processing apparatus according to claim 1, wherein, The graphical information is a distribution map, which maps the distributions constructed from the multiple graphs. The first generation unit generates a graph structure that includes constituent elements corresponding to the blank parts of the distribution map as the new graph structure.

3. The information processing apparatus according to claim 1, wherein, The first generation unit generates a new graph structure based on the graph information, wherein the deviation is greater than or equal to a first predetermined value and less than a second predetermined value.

4. The information processing apparatus according to claim 1, wherein, The graphical information includes information representing at least one of the following: the category, positional relationship, shape, and color of objects in the graphical construction. The first generation unit determines the deviation amount based on at least one of the object's category, positional relationship, shape, and color in the graph construction.

Images