Interest program acquisition device and interest program acquisition program

Abstract

To provide an interest program acquisition device and an interest program acquisition program that can properly express individuality of a robot.SOLUTION: An interesting program acquisition device 1 comprises: a view data extraction part 11 which extracts a keyword from a program that a user views in a predetermined period; a score generation part 12 which calculates scores indicative of the appearance frequencies of keywords extracted from programs that other users view; an addition score part 13 which calculates a weighting value indicative of the appearance frequency of the keyword extracted from the program that the user views; an interest score part 14 which calculates interest scores, obtained by adjusting scores according to the weighting value, by keywords; and a program determination part 15 which selects an interest program based upon the interest scores for keywords extracted from information on respective viewable programs.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a technology for setting programs of interest to a robot that watches content such as television programs together with a person. [Background technology]

[0002] 2. Description of the Related Art Research and development of communication robots capable of conversing with humans has been progressing. For example, Patent Document 1 proposes a technology that extracts keywords from the video, audio, and subtitle data of a television program and generates speech sentences, allowing a robot that watches television together with a person to speak related to the program, thereby enabling a dialogue between the robot and the person while watching television.

[0003] Furthermore, Patent Document 2 proposes a technique that allows multiple robots to have different personalities while referring to information obtained from other robots. Furthermore, Patent Document 3 proposes a technology in which a robot expresses its individuality by identifying programs that it would like to watch as programs containing keywords extracted from programs that it has viewed, and by changing the level of interest in each keyword, it expresses a sense of the robot's growth. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2019-185400 A [Patent Document 2] JP 2018-101249 A [Patent Document 3] JP 2022-162742 A Summary of the Invention [Problem to be solved by the invention]

[0005] However, the personality of a robot expressed by the conventional technology was based on the personal information of the user who uses the robot, such as the television programs the user is watching, or was limited to adding common information to this. For this reason, the personality of the robot was likely to be similar to or common to the user, and was not sufficient to be called the personality of the robot itself.

[0006] An object of the present invention is to provide an interesting program acquisition device and an interesting program acquisition program that can appropriately express the personality of a robot. [Means for solving the problem]

[0007] The interest program acquisition device of the present invention comprises a viewing data extraction unit which extracts keywords from programs viewed by a user during a specified period, a score generation unit which calculates a score indicating the frequency of appearance of the keywords extracted from programs viewed by other users other than the user, a score adding unit which calculates a weighting value indicating the frequency of appearance of the keywords extracted from programs viewed by the user, an interest score unit which calculates an interest score for each keyword by adjusting the score in accordance with the weighting value, and a program determination unit which selects an interest program based on the interest score for the keywords extracted from information on each viewable program.

[0008] The score adding unit may subtract an updated score corresponding to the amount of fluctuation in the frequency of appearance of the keyword in the past from the weighting value each time a keyword for which the weighting value has been calculated appears in a new program viewed by the user.

[0009] The viewing data extraction unit may further extract program genres along with the keywords, the score generation unit may further calculate a score indicating the frequency of occurrence of the genres extracted from the programs viewed by the other users, the score addition unit may further calculate a weighting value indicating the frequency of occurrence of the genres extracted from the programs viewed by the user, and the program determination unit may determine the programs of interest based on the interest scores for the keywords and genres extracted from information about each of the viewable programs.

[0010] The score adding unit may subtract an updated score from the weighting value based on the amount of fluctuation in the frequency of appearance of the keyword or genre in the past each time the keyword or genre for which the weighting value was calculated appears in a new program watched by the user.

[0011] The program determination section may compare the sum of the interest scores for each of the viewable programs, and select the program with the largest sum of the interest scores.

[0012] The keywords may be extracted from subtitle data and electronic program guide data for each program.

[0013] The genre may be extracted from electronic program guide data for each program.

[0014] The interesting program acquisition program according to the present invention is for causing a computer to function as the interesting program acquisition device. Effect of the Invention

[0015] According to the present invention, the robot's personality is appropriately expressed by the robot's own programs of interest. [Brief description of the drawings]

[0016] [Figure 1] 1 is a block diagram showing a functional configuration of an interest program acquisition device according to an embodiment; [Diagram 2] 5 is a diagram illustrating processing content of a viewing data extraction unit in the embodiment. FIG. [Diagram 3] 11 is a diagram illustrating a process of a score generating unit according to an embodiment. FIG. [Figure 4] 11 is a diagram illustrating a processing content of a weighting value update unit in the embodiment. FIG. [Diagram 5] 13 is a flowchart showing a processing procedure of an adding score unit in the embodiment. [Figure 6] 13 is a diagram illustrating an example of the processing content of an interest score unit in the embodiment. FIG. [Figure 7] 11 is a diagram illustrating a process performed by a program determination unit in the embodiment. FIG. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0017] An example of an embodiment of the present invention will now be described. The interesting program acquisition device of this embodiment is a robot that watches content such as television programs together with a person, and acquires program content (interesting programs) that the robot itself wants to watch, and uses this in communication with the user. As the personality of the robot, an interest keyword and genre are initially set based on the viewing history of people other than the user of the robot, and this personality is weighted based on the viewing history of the user, thereby changing the robot's interest programs.

[0018] FIG. 1 is a block diagram showing a functional configuration of an interesting program acquisition device 1 according to the present embodiment. The interesting program acquisition device 1 is an information processing device (computer) equipped with a control unit 10 and a storage unit 20, and is incorporated in a robot or communicatively connected to the robot to output program information acquired as the robot's own interesting program. The interesting program acquisition device 1 also receives input from a television (television receiver) of subtitle data of a program being played on the television and data included in an electronic program guide (EPG) for this program. The interesting program acquisition device 1 is also connected via a network to a database (EPG information DB) that stores EPG data and is provided in an external server or the like, and data stored in the EPG information DB is input to the interesting program acquisition device 1.

[0019] The control unit 10 operates as each of the following functional units by reading and executing software (interesting program acquisition program) stored in the storage unit 20. In addition, the storage unit 20 stores various databases (DBs) in addition to the interesting program acquisition program.

[0020] Specifically, the control unit 10 includes a viewing data extraction unit 11, a score generation unit 12, a score addition unit 13, an interest score unit 14, and a program determination unit 15. The storage unit 20 also includes a viewing DB 21, a unique score DB 22, a weighting value DB 23, and an interest score DB 24.

[0021] The viewing data extraction section 11 includes a keyword extraction section 111 and a genre extraction section 112, and stores information about programs in the viewing DB 21 according to the user's television program viewing history.

[0022] The keyword extraction unit 111 extracts keywords from the subtitle data and electronic program guide (EPG) data of a program viewed by a user. Here, the keywords are words and phrases such as proper nouns contained in the subtitle data and the EPG data, and those registered in a dictionary prepared in advance may be extracted.

[0023] The genre extraction unit 112 extracts the genre of a program viewed by the user from the EPG data of the program. The keywords extracted by the keyword extraction unit 111 and the genres extracted by the genre extraction unit 112 are stored in the viewing DB 21. Here, the genre is one or more categories that are set for each program according to its contents, and for example, a medium category (sub-genre) in the EPG data may be extracted.

[0024] FIG. 2 is a diagram illustrating the processing contents of the viewing data extracting section 11 in this embodiment. The subtitle data includes time information and text data for displaying the subtitles. The EPG data also includes time information for each program, as well as codes for classifying the contents of the programs (major classification, medium classification, etc.), and the program title or introductory text.

[0025] The keyword extraction unit 111 extracts a plurality of keywords from the text data. The genre extraction unit 112 extracts genres, such as "0x4f" (variety / other) and "0x01" (weather), from the EPG data. Note that, for example, in "0x4f", "4" is a code indicating a major category and "f" is a code indicating a medium category.

[0026] The score generation unit 12 includes a keyword score generation unit 121 and a genre score generation unit 122, and calculates scores indicating the frequency of occurrence of each of the keywords and genres extracted from the viewing history of users other than the user using the target robot, and stores the scores in the unique score DB22.

[0027] Here, the viewing history of other users (viewing DB21) is selected from the viewing histories of all robot users shared on the network by a plurality of robots of the same model, and is set as the personality of the robot. If the average viewing history of all users were used, the robot would have average interests, and users would find it difficult to sense individuality in the robot. Therefore, the score generation unit 12 selects the viewing histories of the individual users. In addition, if the viewing history of an individual whose hobbies and tastes are different from those of the user is selected, the user will be less likely to feel that his or her viewing history is affecting the creation of the robot's personality. Therefore, the score generation unit 12 may select the viewing history of an individual who matches, for example, the age or gender of the user, rather than being completely random.

[0028] The keyword score generating unit 121 calculates, for example, a TF value as a score indicating the appearance frequency of each keyword. The genre score generating unit 122 calculates the viewing ratio as a score indicating the frequency of appearance of the genre. These scores, which indicate the frequency of occurrence of keywords and genres, are stored in the unique score DB 22 .

[0029] FIG. 3 is a diagram illustrating the process contents of the score generating unit 12 in this embodiment. The keyword score generating unit 121 converts information on the subtitle data and EPG data of all programs viewed by a user v (another user) during a predetermined period (for example, one week) into a single document d. v Document d v Keyword in w i Number of occurrences of n w_i,d_v , the sum of the frequency of occurrence of all keywords Σ j n w_i,d_v Using the keyword w i TF value (tf w_i,d_v ) is calculated as follows:

number

[0030] For example, if the number of times the genre "weather" appears is 140 and the total number of genres extracted from all programs is 2000, the genre score generation unit 122 calculates that the viewing ratio of "weather" is 140÷2000=0.07 (7%).

[0031] The score addition unit 13 includes a keyword score generation unit 131, a genre score generation unit 132, and a weighting value update unit 133, and uses the user's viewing history, i.e., the viewing DB21, as input, updates the weighting values ​​to be added to each score in the unique score DB22, and stores them in the weighting value DB23.

[0032] The keyword score generation unit 131 and the genre score generation unit 132, like the keyword score generation unit 121 and the genre score generation unit 122 provided in the score generation unit 12, calculate scores indicating the frequency of occurrence of each keyword and genre as weighting values ​​based on the user's viewing history stored in the viewing DB21.

[0033] The weighting value update unit 133 calculates weighting values ​​for the scores of each keyword and genre stored in the unique score DB 22 as follows, and stores the calculated weighting values ​​in the weighting value DB 23.

[0034] FIG. 4 is a diagram illustrating the processing contents of the weighting value update unit 133 in this embodiment. Here, the procedure for calculating the weighting value of a keyword is exemplified, but the same applies to the update score of a genre.

[0035] First, the weighting value update unit 133 counts the number of occurrences of a certain keyword w in the user's viewing history (viewing DB 21) for each predetermined period (for example, monthly), and obtains the difference x x (1 < i ≤ n) as an absolute value. Then, the average value of the differences in the number of occurrences is i (1 < i ≤ n) is obtained as an absolute value. Then, the average value of the differences in the number of occurrences is

Equation

[0036] Next, the subtitle data and EPG data of all the viewed programs are regarded as one document d v and, when the sum of the number of occurrences of all keywords in the document d v is Σ j n w_j,d_v the update score l x of the keyword w w_x,d_v is defined as follows.

Equation

[0037] The method for calculating the update score for a genre is the same as that for a keyword. That is, the weighting value update unit 133 first calculates the number of occurrences of a genre, for example, monthly, and obtains the difference in the number of occurrences per month. Subsequently, the weighting value update unit 133 calculates the average value of all the differences and uses the value obtained by dividing it by the number of occurrences of all keywords as the update score.

[0038] Here, the update score is a value that is subtracted from the score every time the corresponding keyword or genre appears once in the user's viewing history. That is, the new interest originally weighted to the robot's interest is updated using the law of diminishing marginal utility every time a program related to the corresponding keyword or genre is actually viewed. As a result, the way interest increases will differ depending on whether you see an interesting keyword or genre for the first time (1st time) or multiple times (for example, 100th time). The more times you view something, the more the weighting value for that keyword or genre will decrease by the update score, indicating a decrease in interest (boredom).

[0039] FIG. 5 is a flowchart showing the processing procedure of the score adding unit 13 in this embodiment. In step S1, score adding section 13 extracts from viewing DB 21 the viewing history of a user for a predetermined period (for example, one week). In step S2, the score adding unit 13 calculates the TF value for each keyword and the viewing ratio for each genre, using the keyword score generating unit 131 and the genre score generating unit 132, as scores indicating the frequency of appearance, respectively.

[0040] In step S3, the score adding unit 13 causes the weighting value updating unit 133 to calculate a weighting value for each keyword and genre. In step S4, the score adding unit 13 stores the calculated weighting value in the weighting value DB 23.

[0041] The interest score unit 14 includes an interest keyword score generation unit 141 and an interest genre score generation unit 142, and calculates interest keyword scores and interest genre scores by adding the weighting values ​​in the weighting value DB23 to the scores of each keyword and genre stored in the unique score DB22, and stores them in the interest score DB24.

[0042] FIG. 6 is a diagram illustrating the processing contents of the interest score unit 14 in this embodiment. The interest keyword score generation unit 141 calculates an interest keyword score indicating the robot's current interests, which is the robot's original interests (unique score DB22) plus the user's interests (weighting value DB23), by adding the respective weighting values ​​to the score (TF value) of each keyword.

[0043] The interest genre score generating unit 142, like the interest keyword score generating unit 141, calculates an interest genre score by adding a weighting value to the score (viewing ratio) of each genre.

[0044] Here, the interest keyword score and the interest genre score are, for example, the maximum value x max and the minimum value x min may be normalized to a value between 0 and 1 as follows:

number

[0045] The degree of weighting can be changed as appropriate, and the score (TF value or viewing ratio) in the unique score DB22 and the score (TF value or viewing ratio) in the weighting value DB23 may be added equally (1:1), or, for example, the score (TF value or viewing ratio) in the weighting value DB23 may be reduced to a predetermined ratio (1:0.1, etc.).

[0046] The program determination unit 15 determines programs of interest to the robot based on the scores of each keyword and genre stored in the interest score DB 24. Specifically, the program determination unit 15 first obtains one week's worth of EPG data from the EPG information DB, and extracts keywords and genres from the introductions of each program. Next, the program determination unit 15 calculates the sum of the interest scores for each program based on the interest scores stored in the interest score DB 24, and sets the program with the largest sum of interest scores as the program in which the robot itself is interested.

[0047] FIG. 7 is a diagram illustrating the process contents of the program determination unit 15 in this embodiment. The interest score DB 24 stores interest scores for keywords 1, 2, 3, . . . and interest scores for genres 1, 2, 3, . At this time, genres 1 and 3 and keywords 1, 5, ... are extracted from program 1 obtained from the EPG information DB. By summing up the interest scores of these genres and keywords, a score of 0.102 for program 1 and a score of 0.1358 for program 2 are obtained. In this example, program 2, which has the largest sum of interest scores, is determined to be the interesting program.

[0048] In this way, the program determination unit 15 selects a program that matches the robot's current interests based on the interest score DB 24. The interest score DB24 is a result of weighting the original personality of the robot (unique score DB22) with the interest based on the user's viewing history (weighting value DB23), and is updated, for example, at the following times:

[0049] First, the viewing DB21, which serves as the source of weighting values ​​and is provided as the personality of other robots, is updated in real time each time a user (robot) watches a program, or at a predetermined interval (for example, daily, weekly, etc.).

[0050] The unique score DB22 set as the personality of the robot may not be updated after the initial setting, but may be updated at predetermined intervals (e.g., one year, half a year, etc.). In this case, in order to avoid excessively large changes, the viewing history of the same user as the previous time may be used.

[0051] On the other hand, the weighting value DB23 based on the user's viewing history is updated every month, which is the calculation unit for the update score, and is further updated each time a keyword or genre stored in the database appears in a newly viewed program. The interest score DB24 is updated whenever the weighting value DB23 is updated, or at a fixed cycle of one month, so that the interests of the robot itself change.

[0052] According to this embodiment, the interesting program acquisition device 1 calculates a score indicating the frequency of appearance of keywords and genres extracted from programs viewed by users other than the user of the robot, and regards this as the robot's original interest. Furthermore, the interesting program acquisition device 1 calculates a weighting value indicating the frequency of appearance of keywords and genres extracted from programs viewed by the user of the robot, thereby weighting the robot's original interest. This allows the creation of robots with different interests depending on the user, and expresses the individuality of the robot. Furthermore, the robot's original interests are updated based on the user's viewing history, so the robot expresses a sense of growth by tuning in to the user's interests while maintaining its individuality. As a result, the interest program acquisition device 1 allows the robot to autonomously select a program that the robot would like to watch based on the interest score that expresses the individuality and sense of growth, and the robot's own interest programs can appropriately express the robot's individuality and sense of growth.

[0053] The interesting program acquisition device 1 may determine an interesting program only by keywords, without using an interest score related to the genre of the program. Here, the amount of information in the EPG data used when selecting an interesting program varies depending on the program. Therefore, a program with a short introduction in the EPG has a small number of extracted keywords, and is therefore unlikely to be selected as an interesting program. As a result, when the interesting program acquisition device 1 determines an interesting program only by keywords, it tends to select a program with a long introduction. Therefore, the interesting program acquisition device 1 can more appropriately select interesting programs by calculating and using the interest scores for the genres that are assigned to all programs, in addition to the keywords.

[0054] The interest program acquisition device 1 may subtract an updated score according to the amount of fluctuation in the frequency of appearance of the keyword or genre in the past from the weighting value each time the keyword or genre for which a weighting value has been calculated appears in a new program viewed by the user. The amount of variation may be, for example, the average value of the difference in the number of occurrences per month. This amount of variation represents the speed at which a user becomes bored with a keyword or genre, and the difference in this amount for each robot further represents individuality.

[0055] The interesting program acquisition device 1 compares the sum of the interest scores for each viewable program and selects the program with the largest sum. This allows the interesting program acquisition device 1 to appropriately and easily determine the degree of interest in each program.

[0056] The interest program acquisition device 1 extracts keywords for each program from the subtitle data and EPG data, and extracts the genre from the EPG data, thereby easily setting an interest score and rating each program. In addition, when the video or audio of the program can be acquired, the interesting program acquisition device 1 may extract the keyword and the genre by image analysis or audio recognition, etc. Also, the interesting program acquisition device 1 may extract the keyword and the genre by searching the program related information on the Internet.

[0057] Although the embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments. Furthermore, the effects described in the above-described embodiments are merely a list of the most preferable effects resulting from the present invention, and the effects of the present invention are not limited to those described in the embodiments.

[0058] In this embodiment, the configuration and operation of the interest program acquisition device 1 have been mainly described, but the present invention is not limited to this and may be configured as a method or program having each component for acquiring programs of interest to the robot itself.

[0059] Furthermore, the functions of the interest program acquisition device 1 may be realized by recording a program for realizing the functions of the interest program acquisition device 1 on a computer-readable recording medium, and reading and executing the program recorded on the recording medium into a computer system.

[0060] The term "computer system" here includes hardware such as the OS and peripheral devices. Additionally, "computer-readable recording media" refers to portable media such as flexible disks, optical magnetic disks, ROMs, and CD-ROMs, as well as storage devices such as hard disks built into computer systems.

[0061] Furthermore, the term "computer-readable recording medium" may include a recording medium that dynamically stores a program for a short period of time, such as a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line, and a recording medium that stores a program for a certain period of time, such as a volatile memory in a computer system that serves as a server or client in such a case. The above program may be a program for realizing part of the above-mentioned functions, or may be a program that can realize the above-mentioned functions in combination with a program already recorded in the computer system. [Explanation of symbols]

[0062] 1. Device for acquiring programs of interest 10 Control section 11 Viewing data extraction unit 12 Score Generation Section 13 Additive Score Section 14 Interest Score Section 15 Program Decision Department 20 Memory section 21 Viewing DB 22 Unique score DB 23 Weighting Value DB 24 Interest Score DB 111 Keyword Extraction Unit 112 Genre Extraction Section 121 Keyword score generation unit 122 Genre Score Generator 131 Keyword score generation unit 132 Genre Score Generator 133 Weighting value update unit 141 Interest Keyword Score Generation Unit 142 Interest genre score generation unit

Claims

1. a viewing data extraction unit that extracts keywords from programs viewed by a user during a predetermined period; a score generating unit that calculates a score indicating an appearance frequency of a keyword extracted from a program viewed by a user other than the user; a score adding unit that calculates a weighting value indicating an appearance frequency of a keyword extracted from a program viewed by the user; an interest score unit that calculates an interest score for each keyword by adjusting the score according to the weighting value; and a program determination unit that selects an interesting program based on the interest score for the keyword extracted from information on each viewable program.

2. The interest program acquisition device of claim 1, wherein the score adding unit subtracts an updated score from the weighting value each time a keyword for which the weighting value has been calculated appears in a program newly viewed by the user, the updated score corresponding to the amount of fluctuation in the frequency of occurrence of the keyword in the past.

3. The viewing data extraction unit further extracts a program genre together with the keyword, The score generation unit further calculates a score indicating an appearance frequency of the genre extracted from the programs viewed by the other users, The score adding unit further calculates a weighting value indicating an appearance frequency of the genre extracted from the program viewed by the user, The interesting program acquisition device according to claim 1 , wherein the program determination unit determines the interesting programs based on the interest scores for keywords and genres extracted from information on each of the viewable programs.

4. The interest program acquisition device of claim 3, wherein the score adding unit subtracts an updated score from the weighting value each time a keyword or genre for which the weighting value has been calculated appears in a new program watched by the user, the updated score corresponding to the amount of fluctuation in the frequency of occurrence of the keyword or genre in the past.

5. 5. The interesting program acquisition device according to claim 1, wherein the program determination unit compares the sum of the interest scores for each of the viewable programs and selects the program with the largest sum of the interest scores.

6. 5. The apparatus for acquiring an interest in a program according to claim 1, wherein the keyword is extracted from subtitle data and electronic program guide data for each program.

7. 5. The apparatus for acquiring an interest-providing program according to claim 3, wherein the genre is extracted from electronic program guide data for each program.

8. An interesting program acquisition program for causing a computer to function as the interesting program acquisition device according to any one of claims 1 to 4.

Images