Information processing system and method for processing information

1 . An information processing system configured to output an answer generated using a large language model in response to a question input by any one of users, the information processing system comprising:

at least one memory configured to store a program; and

at least one processor configured to run the program to execute a process, wherein:

the at least one processor is configured to obtain content included in electronic documents, the content including category content based on types of users and common content that is for all of the types of the users;

the at least one processor is configured to convert the content into document vectors each assigned with the any of content classification items of the content prior to the conversion, the content classification items including content category items that respectively correspond to different types of the category content, and a common content item that corresponds to the common content;

the at least one processor is configured to obtain a user classification item related to the question, the user classification item including user category items that respectively correspond to different types of the users, and an all user item that corresponds to all of the types of the users;

the at least one processor is configured to convert the question into a question vector;

the at least one processor is configured to search the document vectors under specified search conditions, and obtain for a specified number of the document vectors based on a relevance to the question vector, the specified search conditions including

(i) when the question vector is assigned with the all user item, all of the document vectors assigned with the content category items and the document vectors assigned with the common content item are determined as the search scope; and

(ii) when the question vector is assigned with one of the user category items, only the document vectors assigned with the common content item and the document vectors assigned with one of the content category items that corresponds to the one of the user category items are determined as the search scope, while document vectors assigned with other content category items are excluded from the search scope;

the at least one processor is configured to generate a prompt for input to the large language model, the prompt including the content of the specified number of the document vectors prior to the conversion, serving as context, and the question;

the at least one processor is configured to input the prompt to the large language model; and

the at least one processor is configured to output the answer generated based on text output by the large language model.

2 . The information processing system according to claim 1 , wherein:

the users are employees;

the user classification classifies the employees in accordance with types of the employees;

the at least one memory stores a user database including employee information of each of the employees and the user classification items; and

the at least one processor is configured to obtain the employee information and acquire the user classification item that corresponds to the obtained employee information from the user database.

3 . The information processing system according to claim 2 , wherein the types of the employees are specified in accordance with at least one of assigned department, work type, work location, and job title.

4 . The information processing system according to claim 1 , wherein

the at least one processor is configured to translate the question input in a first language into a second language, and

the at least one processor is configured to translate the text output by the large language model in the second language into the first language.

5 . The information processing system according to claim 1 , wherein:

the user classification classifies the users in accordance with countries or regions to which the users belong;

the user classification items include user category items that respectively correspond to the countries or regions, and an all user item;

the content classification items include content category items that respectively correspond to the countries or regions, and a common content item; and

the content assigned with the content category items is specific to a corresponding one of the countries or regions, and the content assigned with the common content item is common to all countries and regions.

6 . The information processing system according to claim 1 , wherein:

the at least one memory stores a user database including user information of each of the users and the user classification items; and

the at least one processor is configured to obtain the user information and acquire the user classification item that corresponds to the obtained user information from the user database.

7 . An information processing system configured to output an answer generated using a large language model and a vector database in response to a question input by any one of users, the vector database being configured to store document vectors converted from content based on types of users and common content that is for all of the types of the users, the information processing system comprising:

at least one memory configured to store a program; and

at least one processor configured to run the program to execute a process, wherein:

the at least one processor is configured to obtain a user classification item related to the question, the user classification item including user category items that respectively correspond to different types of the users, and an all user item that corresponds to all of the types of the users;

the at least one processor is configured to convert the question into a question vector;

the at least one processor is configured to search the document vectors under specified search conditions, and obtain a specified number of the document vectors based on a relevance to the question vector, the specified search conditions including:

(i) when the question vector is assigned with the all user item, all of the document vectors assigned with the content category items and the document vectors assigned with the common content item are determined as the search scope; and

(ii) when the question vector is assigned with one of the user category items, only the document vectors assigned with the common content item and the document vectors assigned with one of the content category items that corresponds to the one of the user category items are determined as the search scope, while document vectors assigned with other content category items are excluded from the search scope;

the at least one processor is configured to generate a prompt for input to the large language model, the prompt including the content of the specified number of the document vectors prior to the conversion, serving as context, and the question;

the at least one processor is configured to input the prompt to the large language model; and

the at least one processor is configured to output the answer generated based on text output by the large language model.

8 . A method for processing information that outputs an answer generated using a large language model in response to a question input by any one of users, the method comprising:

obtaining, with one or more computers, content included in electronic documents, the content including category content based on types of users and common content that is for all of the types of the users;

converting, with the one or more computers, the content into document vectors each assigned with the any of content classification items of the content prior to the conversion, the content classification items including content category items that respectively correspond to different types of the category content, and a common content item that corresponds to the common content;

obtaining, with the one or more computers, a user classification item related to the question, the user classification item including user category items that respectively correspond to different types of the users, and an all user item that corresponds to all of the types of the users;

converting, with the one or more computers, the question into a question vector;

searching, with the one or more computers, the document vectors under specified search conditions, and obtain a specified number of the document vectors based on relevance to the question vector, the specified search conditions including:

(i) when the question vector is assigned with the all user item, all of the document vectors assigned with the content category items and the document vectors assigned with the common content item are determined as the search scope; and

(ii) when the question vector is assigned with one of the user category items, only the document vectors assigned with the common content item and the document vectors assigned with one of the content category items that corresponds to the one of the user category items are determined as the search scope, while document vectors assigned with other content category items are excluded from the search scope;

generating, with the one or more computers, a prompt for input to the large language model, the prompt including the content of the specified number of the document vectors prior to the conversion, serving as context, and the question;

inputting, with the one or more computers, the prompt to the large language model; and

outputting, with the one or more computers, the answer generated based on text output by the large language model.