Information processing apparatus

Abstract

The present invention addresses the problem of increasing training data relating to a user on the basis of behavior records. This information processing device is provided with: an input means for inputting, into a first large-scale language model, a presentation including record information relating to a behavior record of a user and facility information relating to one or more facilities; an acquisition means for acquiring, from the first large-scale language model, an answer indicating whether or not the user accesses one or more facilities; and a generation means for generating, on the basis of the recorded information and the answer, training data for machine learning of a model different from the first large-scale language model.

Description

Technical Field

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

[0002] As such a device, for example, a device relating to a recommendation service that uses a user’s usage record information to recommend items based on the user’s preferences has been proposed (see Patent Document 1).

[0003] Patent Document 1: Japanese Patent Application Publication No. 2014-002492 Summary of the Invention

[0004] For example, a recommendation service can be provided using a trained model built through machine learning. However, when focusing on a single user, the number of items that user utilizes is significantly less than the number of items offered. Therefore, even if a user's usage history is used as training data, that user's characteristics may not be reflected in the trained model.

[0005] The present invention was made in view of the above-mentioned problems, and its objective is to provide an information processing apparatus capable of adding training data related to a user.

[0006] An information processing apparatus according to one aspect of the present invention comprises: an input unit that inputs a prompt including record information related to a user's behavior records and facility information related to one or more facilities into a first large-scale language model; an acquisition unit that acquires from the first large-scale language model a response indicating whether the user has visited the one or more facilities; and a generation unit that generates training data for machine learning of a model different from the first large-scale language model based on the record information and the response. Attached Figure Description

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

[0008] Figure 2 This is a flowchart illustrating the operation of the information processing apparatus according to the first embodiment.

[0009] Figure 3 This is a block diagram illustrating the structure of the information processing apparatus according to the second embodiment.

[0010] Figure 4 This is a flowchart illustrating the operation of the information processing apparatus according to the second embodiment. Detailed Implementation

[0011] <First Embodiment>

[0012] refer to Figure 1 and Figure 2 The first embodiment of the information processing apparatus will be described. Figure 1 The information processing device 10 includes a computing unit 11, a storage unit 12, a communication unit 13, an input unit 14, and an output unit 15. The computing unit 11, storage unit 12, communication unit 13, input unit 14, and output unit 15 are connected via a data bus 16. Furthermore, the information processing device 10 can be a personal computer, a tablet terminal, or a smartphone.

[0013] Information processing device 10 can access server 20, which provides services using Large Scale Language Model (LLM), via network NW. That is, information processing device 10 and server 20 are configured to communicate via network NW. Furthermore, server 20 can be a cloud server. Network NW can be a wide area network such as the Internet, or a local area network such as a local area network (LAN). Additionally, information processing device 10 and server 20 can constitute information processing system 1.

[0014] 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.

[0015] The processor may be at least one of a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), a Field Programmable Gate Array (FPGA), and a Tensor Processing Unit (TPU).

[0016] Storage device 12 may be at least one of random access memory (RAM), read-only memory (ROM), hard disk drive, magneto-optical disk drive, solid-state drive (SSD), and optical disk array. That is, storage device 12 may be implemented by a single device or by multiple devices.

[0017] The communication device 13 can communicate with external devices of the information processing device 10. In addition, the communication device 13 can perform wired communication or wireless communication.

[0018] Input device 14 is a device capable of accepting information input from an external source to information processing device 10. 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 microphone capable of collecting the user's voice. In this case, the user can perform voice input. Input device 14 may, for example, include a recording medium reading device capable of reading information recorded on a recording medium removable from information processing device 10, such as a USB (Universal Serial Bus) memory. Furthermore, 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.

[0019] 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 characters and images. Additionally, output device 15 may 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. Output device 15 may include a printer. Output device 15 can output information to recording media 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.

[0020] 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. When the arithmetic unit 11 executes the computer program CP, storage device 12 can temporarily store data temporarily used by the arithmetic unit 11.

[0021] Furthermore, the computer program CP can be recorded on a computer-readable and non-temporary recording medium. In this case, the computer program CP can be stored in the storage device 12 by reading the recording medium using a recording medium reading device (not shown) included in the information processing device 10. Additionally, at least one of optical discs, magnetic media, magneto-optical discs, semiconductor memory, and any medium capable of storing other programs can be used as the recording medium. Furthermore, the computer program CP can also be obtained from an external (not shown) device outside the information processing device 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 device 10.

[0022] 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 can 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.

[0023] In this embodiment, the arithmetic unit 11 includes an input unit 111, an acquisition unit 112, and a generation unit 113 as functional blocks. Alternatively, at least one of the input unit 111, the acquisition unit 112, and the generation unit 113 can be implemented as a physical processing circuit. Or, at least one of the input unit 111, the acquisition unit 112, and the generation unit 113 can be implemented as a mixture of logic functional blocks and physical processing circuits.

[0024] For example, there is a recommendation service that recommends restaurants based on user preferences. When attempting to provide this recommendation service using a trained model built through machine learning, the following problem arises: When focusing on a single user, the number of restaurants visited by that user is significantly less than the total number of restaurants the recommendation service targets. Therefore, even if a user's records are used as training data, the trained model may fail to recommend restaurants that match that user's preferences. The information processing apparatus 10 of this embodiment performs the following operations to increase the training data related to a single user.

[0025] refer to Figure 2 The flowchart illustrates the operation of the information processing device 10. Figure 2 In step S101, the input unit 111 of the information processing device 10 sends a prompt containing record information related to a user's behavior record and facility information related to more than one facility to the server 20 via the network NW. As a result, the prompt is input into the large-scale language model.

[0026] A behavior record is a concept that includes at least one of the following: access records, purchase records, usage records, and browsing records. A user's behavior record can be stored in storage device 12 or in a device different from information processing device 10. A facility can be an indoor facility or an outdoor facility. For example, a prompt input into the large-scale language model may include the question, "Did a user access OO (facility name)?" Similarly, a prompt input into the large-scale language model may include the question, "If a user is going to △△ (region name), which facility would they visit?" Additionally, a prompt input into the large-scale language model may include a statement specifying a day of the week or a time period.

[0027] When the large-scale language model outputs a response to a prompt (e.g., a response indicating whether a user has accessed more than one facility), the server 20 sends the response to the information processing device 10 via the network NW. As a result, the acquisition unit 112 of the information processing device 10 acquires the response from the large-scale language model (step S102).

[0028] Next, the generation unit 113 of the information processing device 10 generates training data (i.e., training data for machine learning of a model different from the large-scale language model) based on a user's recorded information and the response of the large-scale language model (step S103).

[0029] (Technical effect)

[0030] The prompts input into the large-scale language model include recorded information related to a user's behavior. Therefore, it can be expected that the large-scale language model's response is based on a user's behavior record. As a result, a user's behavior record can be manually generated based on the large-scale language model's response. By generating training data based on a user's actual behavior record and the large-scale language model's response using the generation unit 113 of the information processing device 10, the amount of training data related to a user can be significantly increased.

[0031] By performing machine learning using the training data generated above, a training model capable of making recommendations that match a user's preferences can be constructed. Furthermore, from the perspective of response time and the cost of model application, recommendation services using such training models are more efficient than those using large-scale language models. Additionally, a method is proposed to generate training data based on a user's records and the records of other users similar to that user, when training data based on a single user's records is limited. However, when a user's records are less abundant than those of other users, the accuracy of the constructed training model can sometimes be relatively low. In contrast, in this embodiment, since the records of other users are not utilized, the possibility of low accuracy in the constructed training model can be prevented.

[0032] <Second Implementation>

[0033] refer to Figure 3 and Figure 4 The second embodiment of the information processing apparatus will be described here. Furthermore, regarding the second embodiment, descriptions that overlap with those of the first embodiment will be appropriately omitted. Additionally, common parts in the accompanying drawings will be indicated by the same symbols. Figure 3In this system, information processing device 10a includes a computing device 11, a storage device 12, a communication device 13, an input device 14, and an output device 15. Information processing device 10a can access servers 20 and 30, which provide services using a large-scale language model, via a network NW. Furthermore, server 30 can be a cloud server. Additionally, information processing device 10a, server 20, and 30 can constitute information processing system 2.

[0034] In this embodiment, the arithmetic unit 11 includes an input unit 111, an acquisition unit 112, a generation unit 113, a determination unit 114, and an adjustment unit 115 as functional blocks. Alternatively, at least one of the input unit 111, acquisition unit 112, generation unit 113, determination unit 114, and adjustment unit 115 can be implemented as a physical processing circuit. Or, at least one of the input unit 111, acquisition unit 112, generation unit 113, determination unit 114, and adjustment unit 115 can be implemented as a mixture of logic functional blocks and physical processing circuits.

[0035] refer to Figure 4 The flowchart below explains the operation of the information processing device 10a. After the processing in step S102, the determination unit 114 of the information processing device 10a determines whether the response from the large-scale language model of the server 20 is appropriate (step S201). In the processing of step S201, the determination unit 114 can send the prompt input into the large-scale language model of the server 20 in the processing of step S101 and the response obtained in the processing of step S102 to the server 30 via the network NW. As a result, the prompt and response can be input into the large-scale language model of the server 30. Then, the determination unit 114 can determine whether the response from the large-scale language model of the server 20 is appropriate based on the response from the large-scale language model of the server 30. That is, the determination unit 114 can use the large-scale language model of the server 30 to determine whether the response from the large-scale language model of the server 20 is appropriate. This method is called "LLM-as-a-Judge".

[0036] In step S201, if the server 20's large-scale language model's response is deemed appropriate (step S201: Yes), the aforementioned step S103 is performed. In this case, the generation unit 113 generates training data based on a user's record information and the appropriate response determined by the determination unit 114.

[0037] In step S201, if the response from the large-scale language model of server 20 is determined to be inappropriate (step S201: No), the adjustment unit 115 of the information processing device 10a can adjust the prompt input into the large-scale language model of server 20 during the process of step S101 (step S202). Then, the process of step S101 can be performed. Alternatively, the adjustment unit 115 can, for example, use the large-scale language model of server 30 to adjust the prompt.

[0038] (Technical effect)

[0039] In this embodiment, the determination unit 114 determines whether the response from the large-scale language model of the server 20 is appropriate. Furthermore, if the response from the large-scale language model of the server 20 is determined to be appropriate, the generation unit 113 generates training data. Therefore, according to this embodiment, training data can be generated based on the highly accurate responses from the large-scale language model.

[0040] Hereinafter, various aspects of the invention derived from the embodiments described above will be described.

[0041] An information processing apparatus according to one aspect of the invention comprises: an input unit for inputting a prompt containing record information related to a user's behavior and facility information related to one or more facilities into a first large-scale language model; an acquisition unit for acquiring a response from the first large-scale language model indicating whether the user has visited the one or more facilities; and a generation unit for generating training data for machine learning of a model different from the first large-scale language model based on the record information and the response.

[0042] In the above embodiments, "input unit 111" is equivalent to an example of an "input unit", "acquisition unit 112" is equivalent to an example of an "acquisition unit", and "generation unit 113" is equivalent to an example of a "generation unit". In the information processing device of the above manner, even when there are relatively few records of a person's behavior, it is possible to increase the training data related to a user based on those behavior records.

[0043] The information processing apparatus described above may include a determination unit for determining whether the answer is appropriate. In the above embodiment, "determination unit 114" is an example of a "determination unit". In this case, the determination unit may use a second large-scale language model, different from the first large-scale language model, to determine whether the answer is appropriate. Furthermore, the generation unit may generate the training data based on the recorded information and the answers determined as appropriate by the determination unit.

[0044] The information processing device described above may include an adjustment unit that adjusts the prompt when the determination unit determines that the answer is inappropriate. In the above embodiment, "adjustment unit 115" is an example of "adjustment unit".

[0045] The present invention is not limited to the embodiments described above, and appropriate modifications can be made without departing from the spirit or concept of the invention as read in its entirety from the claims and description. Information processing apparatuses that accompany such modifications are also included within the technical scope of the present invention.

[0046] Symbol Explanation

[0047] 10, 10a - Information processing device; 111 - Input unit; 112 - Acquisition unit; 113 - Generation unit; 114 - Judgment unit; 115 - Adjustment unit.

Claims

1. An information processing device, characterized in that, have: The input unit inputs prompts, including record information related to the user's behavior records and facility information related to more than one facility, into the first large-scale language model; The acquisition unit acquires a response from the first large-scale language model indicating whether the user has accessed more than one facility; and The generation unit generates training data for machine learning of a model different from the first large-scale language model, based on the recorded information and the response.

2. The information processing device according to claim 1, characterized in that, have: The determination unit determines whether the answer is appropriate.

3. The information processing device according to claim 2, characterized in that, The determination unit uses a second large-scale language model, which is different from the first large-scale language model, to determine whether the answer is appropriate.

4. The information processing apparatus according to claim 2, characterized in that, The generation unit generates the training data based on the recorded information and the answers that the determination unit determines are appropriate.

5. The information processing apparatus according to claim 2, characterized in that, have: An adjustment unit adjusts the prompt if the determination unit determines that the answer is inappropriate.

Images