Program selection device, program selection method, and program selection program

Abstract

To provide a broadcast program selection device, a broadcast program selection method and a broadcast program selection program, capable of expressing interest in a broadcast program as an individual character for each robot and a change thereof.SOLUTION: A broadcast program selection device 1 includes: a keyword extraction unit 11 that extracts a predetermined kind of keyword from broadcast program information in a predetermined period, acquired by a robot; a keyword calculation unit 12 that counts the number of times of appearance of the keyword in the broadcast program information in the predetermined period and the number of times of broadcast programs in which the keyword appears, for each of the keywords; an interest keyword extraction unit 13 that extracts an interest keyword on the basis of an index of an interest degree that is larger as the number of times of appearance is larger and the number of times of broadcast programs is smaller; an interest broadcast program extraction unit 21 that extracts the broadcast program including the interest keyword as an interest broadcast program of the robot from a broadcast program information database; and a robot interest broadcast program communication unit 30 that outputs information on the interest broadcast program for presenting a viewing request of the interest broadcast program to a user.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a control device for a communication robot that watches TV programs with a person.

Background Art

[0002] Conventionally, it has been desired to realize a communication robot that watches TV with a person and has a conversation. For example, in Patent Document 1, a robot has been proposed that extracts keywords from video, audio, and subtitle data of a TV program and generates utterance sentences. As a result, the robot can make utterances related to the program, and the robot and the person can have a conversation while watching TV.

[0003] Also, in Patent Document 2, a method has been proposed in which an interest area and an interest model that characterizes the interest area are learned from the viewing program information of the viewer, and the interest keywords for each user are extracted from the keywords and viewing situation of the program being viewed using the parameters of the interest model. Furthermore, in Patent Document 3, a method has been proposed in which a user updates the action data in the internal memory area of the robot to provide behavior that does not give boredom due to repeated same actions.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Patent Document 2

Patent Document 3

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, in the technology of Patent Document 1, since keywords are extracted from the video or subtitle data of the TV program being viewed and the robot makes a speech, even if the user discovers a program that the user wants to newly view or the program preference changes from the content of the robot's speech, it does not represent the individuality of the robot itself. Also, the technology of Patent Document 2 is also a technology for extracting the user's interest keywords and does not represent the interest of the robot itself.

[0006] In the technology of Patent Document 3, even if the growth of the robot can be expressed by updating the action data of the robot, since the update is performed by the user, the robot cannot change its actions by itself and grow.

[0007] An object of the present invention is to provide a program selection device, a program selection method, and a program selection program that can express the interest in programs and its changes as the individuality of each robot.

Means for Solving the Problems

[0008] The program selection device according to the present invention includes a keyword extraction unit that extracts predetermined types of keywords from program information for a predetermined period acquired by the robot, a keyword calculation unit that counts the number of appearances of each of the keywords in the program information for the predetermined period and the number of program times in which the keyword appears, an interest keyword extraction unit that extracts interest keywords based on an interest index that increases as the number of appearances increases and decreases as the number of program times decreases, an interest program extraction unit that extracts a program including the interest keywords from a program information database as an interest program of the robot, and a robot interest program communication unit that outputs information of the interest program in order to present a viewing request of the interest program to the user.

[0009] The program selection device may output the changed interest program by repeating the extraction of the interest keywords and the extraction of the interest programs at a predetermined timing.

[0010] The keyword extraction unit may adjust the length of the predetermined period according to the amount of the program information.

[0011] The keyword extraction unit may acquire text information from video or audio in addition to subtitle data as the program information.

[0012] The program selection device includes a satisfaction calculation unit that acquires information on programs viewed by the robot and varies the satisfaction with respect to the interesting keywords of the interesting program when the program is the interesting program, and a satisfaction update unit that updates and stores the satisfaction for each interesting keyword. The robot interesting program communication unit may output the interesting program including the interesting keyword with relatively low satisfaction.

[0013] The satisfaction calculation unit may vary the satisfaction while decreasing the marginal utility according to the number of times the interesting program is viewed.

[0014] The marginal utility may be set for each robot.

[0015] The marginal utility may be set according to an index of the degree of interest.

[0016] When the degree of increase in the satisfaction becomes equal to or less than a predetermined value, the satisfaction update unit may request the keyword extraction unit to extract new interesting keywords.

[0017] The keyword extraction unit may exclude the interesting keywords extracted in the past from the extraction target.

[0018] The program selection method according to the present invention includes a keyword extraction step of extracting predetermined types of keywords from program information for a predetermined period acquired by a robot, a keyword calculation step of counting, for each of the keywords, the number of occurrences of the keyword in the program information for the predetermined period and the number of programs in which the keyword appears, an interesting keyword extraction step of extracting interesting keywords based on an index of degree of interest that increases as the number of occurrences increases and decreases as the number of programs decreases, an interesting program extraction step of extracting, from a program information database, programs including the interesting keywords as interesting programs of the robot, and a robot interesting program communication step of outputting information on the interesting programs to present a viewing request for the interesting programs to a user, which are executed by a computer.

[0019] The program selection program according to the present invention is for causing a computer to function as the program selection device.

Effects of the Invention

[0020] According to the present invention, it is possible to express the interest in programs and its changes as the individuality of each robot.

Brief Description of the Drawings

[0021]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Mode for Carrying Out the Invention

[0022] Hereinafter, an example of an embodiment of the present invention will be described. The program selection device of this embodiment is a device that finds a TV program that a robot autonomously wants to watch, finds keywords that the robot itself is interested in from the keywords in the watched program, and expresses the individuality by regarding the TV program including these keywords as the program that the robot itself wants to watch. Furthermore, the program selection device gives a sense of growth so that the robot becomes interested in another program by watching a TV program.

[0023] FIG. 1 is a block diagram showing the functional configuration of the program selection device 1 in this embodiment. The program selection device 1 is an information processing device (computer) having a control unit and a storage unit, which is incorporated into a robot or communicatively connected to output program information as an extraction result.

[0024] The control unit of the program selection device 1 operates as the following respective functional units by reading and executing software (program selection program) stored in the storage unit. In addition to the program selection program, various databases (DBs) are stored in the storage unit.

[0025] Specifically, the control unit includes an interest data unit 10, a program data unit 20, a robot interest program communication unit 30, and a satisfaction data unit 40. In addition, the storage unit includes an interest data DB 50, a robot interest program DB 60, a program data DB 70, and a satisfaction data DB 80.

[0026] The interest data unit 10 includes a keyword extraction unit 11, a keyword calculation unit 12, and an interest keyword extraction unit 13, and outputs keyword data indicating the robot's own interests to the interest data DB 50.

[0027] The keyword extraction unit 11 acquires the text information of the programs viewed by the robot within a predetermined period (for example, the most recent one week), and extracts predetermined types of keywords from the text information. The text information of the program may be, for example, subtitle data or electronic program guide (EPG) data. In addition, for example, text information may be extracted from video or audio using existing methods. Also, the keywords to be extracted may be nouns. For example, in addition to proper nouns, words included in a predetermined dictionary are extracted.

[0028] The keyword calculation unit 12 counts, for each keyword extracted by the keyword extraction unit 11, the number of occurrences (appearance count) in the text information of the programs within a predetermined period and the number of programs in which the keyword appears (program count).

[0029] FIG. 2 is a diagram showing an example of extracting keyword information by the keyword extraction unit 11 and the keyword calculation unit 12 in the present embodiment. In this example, first, words included in the keyword dictionary are sequentially extracted as keywords from the subtitle data of the programs viewed for one week.

[0030] These keywords are extracted multiple times repeatedly, such as "Hokkaido" and "wine". Therefore, the appearance count and the program count for each extracted keyword are counted. For example, "Hokkaido" appears in only one program within one week of programs and appears 30 times in this program. Also, "wine" appears 15 times in total in two programs.

[0031] Note that the target period for keyword extraction and counting may be appropriately set for each user. Also, the length of the target period may be adjusted according to the amount of program information (for example, subtitle data). For example, when the subtitle data is less than a predetermined amount, the target period may be set longer.

[0032] The interest keyword extraction unit 13 extracts keywords that appear more frequently and in fewer program episodes from among the keywords extracted by the keyword extraction unit 11, and stores them in the interest data DB50 as the robot's interest keywords.

[0033] Figure 3 is a diagram showing an example of extraction of interest keywords by the interest keyword extraction unit 13 in the present embodiment. Based on the number of appearances and the number of program episodes counted for each keyword by the keyword calculation unit 12, keywords with a large number of appearances and a small number of program episodes are extracted.

[0034] Specifically, using the interest level z = number of appearances / number of program episodes as an index, the top keywords with a large interest level z are extracted. Here, "Hokkaido" has the maximum interest level z = 30 and is stored in the interest data DB50 as the extraction result of the interest keyword. Note that a plurality of interest keywords may be extracted. For example, a predetermined number of top keywords with an interest level z or keywords with an interest level z equal to or higher than a threshold value may be extracted.

[0035] The program data unit 20 includes an interest program extraction unit 21 and outputs data indicating programs that the robot itself is interested in to the robot interest program DB60. The interest program extraction unit 21 extracts programs that include the interest keywords extracted by the interest data unit 10 from an external program information DB such as EPG information, and stores the extracted programs in the robot interest program DB60 as programs that the robot itself is interested in (robot interest programs).

[0036] The robot interest program communication unit 30 acquires data of the robot interest programs from the robot interest program DB60 in order for the robot to make a statement such as "I want to watch '○○○ (robot interest program)' " to the user and present a viewing request, and transmits it to the robot's speech system.

[0037] Figure 4 is a flowchart showing the processing flow by the program selection device 1 in the present embodiment. In step S1, the keyword extraction unit 11 acquires keywords from subtitle data of TV programs viewed during the specified period or the like.

[0038] In step S2, the keyword calculation unit 12 counts the number of appearances and the number of programs for each of the keywords acquired in step S1.

[0039] In step S3, the interesting keyword extraction unit 13 extracts interesting keywords with a large number of appearances and a small number of programs counted in step S2.

[0040] In step S4, the interesting keyword extraction unit 13 stores the extracted interesting keywords in the interest data DB50.

[0041] In step S5, the interesting program extraction unit 21 acquires the interesting keywords stored in the interest data DB50 in step S4, and extracts programs containing these interesting keywords from the program information DB as robot interesting programs.

[0042] In step S6, the interesting program extraction unit 21 stores the information of the extracted robot interesting programs in the robot interesting program DB60.

[0043] In step S7, the robot interesting program communication unit 30 acquires the information of the robot interesting programs stored in the robot interesting program DB60 in step S6, and transmits it to the speech system.

[0044] Here, the extraction of interesting keywords by the interest data unit 10 and the extraction of robot interesting programs by the program data unit 20 are repeatedly executed at a predetermined timing, so that the robot interesting programs also change and are output as the viewed programs change. Note that the timing for extracting interesting keywords and robot interesting programs may be a fixed cycle, but may be appropriately set according to changes (for example, decreases) in the satisfaction calculated by the satisfaction data unit 40, or at any early timing.

[0045] The satisfaction data unit 40 includes a satisfaction calculation unit 41 and a satisfaction update unit 42, calculates the satisfaction of the robot according to the viewing status of the robot's favorite programs, and outputs it to the satisfaction data DB 80.

[0046] The satisfaction calculation unit 41 acquires the data of the programs viewed by the robot from the program data DB 70, and acquires the data of the robot's favorite programs from the robot favorite program DB 60. Note that in the program data DB 70, information on programs viewed by the robot, that is, programs presented during the operation of the robot, at least information capable of matching with the robot's favorite programs, is extracted from EPG or subtitle data and stored. When the program data acquired by the satisfaction calculation unit 41 from the program data DB 70 matches the robot's favorite program, that is, when the robot views a favorite program, the satisfaction with respect to the robot's interest keywords is changed.

[0047] At this time, the satisfaction calculation unit 41 changes (increases or decreases) the satisfaction while decreasing the marginal utility according to the number of viewing times of the robot's favorite programs. The initial value and the degree of decrease of the marginal utility may be constant, but may also be set for each robot or according to an index of the degree of interest for each interest keyword.

[0048] Also, in order to express that the timing of getting bored varies depending on the program being viewed, the marginal utility may be adjusted according to the program, for example, based on the degree of match between the interest keyword and the robot's favorite program. Furthermore, in order to express the fading of interest over time, etc., the marginal utility may be adjusted according to the passage of time since the interest keyword was extracted.

[0049] The satisfaction update unit 42 updates the satisfaction calculated for each interest keyword by the satisfaction calculation unit 41 and stores it in the satisfaction data DB 80. When a plurality of interesting keywords are extracted, the satisfaction data DB80 stores the satisfaction for each of these interesting keywords. The robot interesting program communication unit 30 may output, for example, a robot interesting program including an interesting keyword with relatively low satisfaction.

[0050] In addition, when the viewing of the robot interesting program is repeated and the degree of increase in satisfaction (increase rate or increase amount) decreases to a predetermined level or less due to the diminishing marginal utility, for example, when the satisfaction decreases, the satisfaction update unit 42 requests the interesting keyword extraction unit 13 to extract new interesting keywords in order to extract a new robot interesting program. At this time, the satisfaction update unit 42 notifies the interesting keyword extraction unit 13 of the past interesting keyword information recorded in the satisfaction data DB80, and the interesting keyword extraction unit 13 may exclude the interesting keywords extracted in the past from the extraction targets. Note that the interesting keyword extraction unit 13 may exclude all past interesting keywords from the targets, but this is not limited thereto. For example, interesting keywords whose satisfaction has turned downward or interesting keywords for which a predetermined time has elapsed since extraction may be excluded from the targets.

[0051] FIG. 5 is a diagram illustrating a method for calculating the satisfaction when the robot in the present embodiment views an interesting program. In this example, the initial value of the marginal utility z1 = 40, and when the corresponding robot interesting program is viewed for the first time for a certain interesting keyword, the satisfaction y1 = 40. Also, the marginal utility decreases, for example, by half each time the robot interesting program is viewed.

[0052] In this case, by viewing the robot interesting program for the second time, the marginal utility z2 = z1 / 2 = 20, and the satisfaction y2 = y1 × z2 = 40 × 20 = 800. Similarly, each time the robot interesting program is viewed, the satisfaction increases as y3 = 8000, y4 = 40000, y5 = 100000, y6 = 125000, but the increase rate decreases, and by the seventh viewing, y7 = 78125, and the satisfaction turns downward.

[0053] For example, after the seventh viewing when the satisfaction level turns downward, the interest keywords are updated and new robot-interest programs are extracted. Alternatively, for example, after the sixth viewing when the rate of increase in satisfaction level is less than 30%, new robot-interest programs may be extracted.

[0054] Note that the rate of decrease in marginal utility is not limited to 1 / 2, and a fixed amount may be subtracted instead of the rate. Also, the formula for calculating satisfaction level is just an example, and for example, marginal utility may be added.

[0055] FIG. 6 is a flowchart showing the processing flow by the satisfaction level data section 40 in the present embodiment. In step S11, the satisfaction level calculation section 41 acquires robot-interest programs from the robot-interest program DB 60.

[0056] In step S12, the satisfaction level calculation section 41 acquires data of the programs viewed by the robot from the program data DB 70.

[0057] In step S13, the satisfaction level calculation section 41 calculates the satisfaction level for each interest keyword according to the number of matches between the programs viewed by the robot and the robot-interest programs.

[0058] In step S14, the satisfaction level update section 42 determines whether the satisfaction level has decreased. If this determination is YES, the process proceeds to step S15; if the determination is NO, the process ends.

[0059] In step S15, in order to acquire new robot-interest programs, the satisfaction level update section 42 requests the interest keyword extraction section 13 to extract new interest keywords.

[0060] According to this embodiment, the program selection device 1 extracts predetermined types of keywords from program information such as subtitle data for a predetermined period acquired by the robot, and based on an index of degree of interest that increases as the number of appearances increases and decreases as the number of programs decreases, extracts interesting keywords. Then, the program selection device 1 outputs a program including these interesting keywords as the robot's interesting program, thereby presenting the viewing request for the interesting program from the robot to the user. Therefore, when the keyword appears frequently, it indicates that the robot is interested in this keyword. Also, if the number of programs in which this keyword appears is small, by setting that the robot wants to watch more programs in which the interesting keyword appears, the robot can extract the programs it is interested in by itself and present them to the user. As a result, it is possible to express the interest in programs and the sense of growth as the individuality of each robot according to the change in the viewed programs.

[0061] In this way, when the robot shows its own interest and the state of growth, for the user, it becomes an experience substituting for watching TV with people with different values, and it can be expected that the interaction during TV viewing with the robot will become more enjoyable. Also, since the program in which the robot is interested is extracted based on the content of the program that the user watched together with the robot, this program is considered to be close to the user's interest, leading to empathy with the robot and the discovery of new interests of the user himself / herself.

[0062] The program selection device 1 repeats the extraction of interesting keywords and the extraction of the robot's interesting programs at a predetermined timing, thereby changing and outputting the robot's interesting programs, and thus it is possible to realize a robot whose object of interest changes over time and does not bore the user.

[0063] The program selection device 1 adjusts the period of the program for extracting interesting keywords according to the amount of program information, thereby reducing the variation in the index of degree of interest due to the viewing situation and being able to extract appropriate interesting keywords.

[0064] The program selection device 1, as program information, in addition to subtitle data, acquires text information from video or audio, thereby reducing the influence of variations in the amount of subtitle data for each program, and further increasing the amount of information so that the degree of interest for each keyword can be appropriately calculated.

[0065] The program selection device 1 acquires information on programs viewed by the robot, and when the program is a robot's interesting program, it varies the satisfaction level for the corresponding interesting keyword, and outputs a robot's interesting program that includes an interesting keyword with a relatively low satisfaction level. Thereby, when the robot views an interesting program, the satisfaction level of the robot increases, and the relative change in the satisfaction level between keywords can express the change in the robot's interest in the program.

[0066] The program selection device 1 varies the satisfaction level while decreasing the marginal utility according to the number of times the robot's interesting program is viewed. That is, each time the robot views an interesting program, the degree of increase in the satisfaction level decreases, so that a more realistic satisfaction level can be expressed.

[0067] This marginal utility can be set for each robot, thereby expressing the individuality of each robot. Also, the marginal utility can be set according to an index of the degree of interest in a keyword, thereby expressing the difference in the change in the satisfaction level according to the degree of interest.

[0068] When the degree of increase in the satisfaction level becomes equal to or less than a predetermined value, for example, when the satisfaction level decreases, the program selection device 1 extracts new interesting keywords and the robot's interesting programs. Thereby, the program selection device 1 can express that the robot has become bored with the robot's interesting program and has shifted to a new interest.

[0069] The program selection device 1 can suppress unnatural changes such as a regression of interest by excluding the interesting keywords extracted in the past from the extraction target, and can express an appropriate change.

[0070] The embodiments of the present invention have been described above. However, the present invention is not limited to the above-described embodiments. Further, the effects described in the above embodiments are merely an enumeration 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.

[0071] In this embodiment, the configuration and operation of the program selection device 1 have been mainly described. However, the present invention is not limited thereto, and it may be configured as a method or program for causing a communication robot to request a user to watch a program, including each component.

[0072] Furthermore, a program for realizing the functions of the program selection device 1 may be recorded on a computer-readable recording medium, and the program recorded on this recording medium may be read into a computer system and executed to be realized.

[0073] The "computer system" referred to here includes hardware such as an OS and peripheral devices. Further, the "computer-readable recording medium" refers to a portable medium such as a flexible disk, a magneto-optical disk, a ROM, a CD-ROM, or a storage device such as a hard disk incorporated in a computer system.

[0074] Furthermore, the "computer-readable recording medium" also includes a communication line that dynamically holds a program for a short 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 volatile memory inside a computer system that becomes a server or a client in that case and holds a program for a certain period of time. Also, the above program may be for realizing a part of the above-described functions, and may further be realized in combination with a program already recorded in the computer system for realizing the above-described functions.

Explanation of Reference Numerals

[0075] 1 Program selection device 10 Interest Data Section 11 Keyword Extraction Section 12 Keyword Calculation Section 13 Interesting Keyword Extraction Section 20 Program Data Section 21 Interesting Program Extraction Section 30 Robot Interesting Program Communication Section 40 Satisfaction Data Section 41 Satisfaction Calculation Section 42 Satisfaction Update Section 50 Interest Data DB 60 Robot Interesting Program DB 70 Program Data DB 80 Satisfaction Data DB

Claims

1. A keyword extraction unit that extracts keywords of a predetermined type from program information for a predetermined period acquired by a robot; A keyword calculation unit that counts, for each of the keywords, the number of occurrences of the keyword in the program information for the predetermined period and the number of programs in which the keyword appears; An interesting keyword extraction unit that extracts interesting keywords based on an index of interestingness that increases as the number of occurrences increases and decreases as the number of programs decreases; An interesting program extraction unit that extracts, from a program information database, programs including the interesting keywords as interesting programs of the robot; A robot interesting program communication unit that outputs information on the interesting programs in order to present a viewing request for the interesting programs to a user, the program selection device comprising:

2. The program selection device according to claim 1, wherein the extraction of the interesting keywords and the extraction of the interesting programs are repeated at a predetermined timing to change and output the interesting programs.

3. The program selection device according to claim 1 or claim 2, wherein the keyword extraction unit adjusts the length of the predetermined period according to the amount of the program information.

4. The program selection device according to any one of claims 1 to 3, wherein the keyword extraction unit acquires text information from video or audio in addition to subtitle data as the program information.

5. A satisfaction calculation unit that acquires information on programs viewed by the robot and varies the satisfaction of the interesting programs with respect to the interesting keywords when the programs are the interesting programs; A satisfaction update unit that updates and stores the satisfaction for each of the interesting keywords, The robot interesting program communication unit outputs the interesting programs including the interesting keywords with relatively low satisfaction, the program selection device according to any one of claims 1 to 4.

6. The program selection device according to claim 5, wherein the satisfaction calculation unit varies the satisfaction while decreasing the marginal utility according to the number of times the interesting program is viewed.

7. The program selection device according to claim 6, wherein the marginal utility is set for each robot.

8. The program selection device according to claim 6 or claim 7, wherein the marginal utility is set according to the index of interestingness.

9. The program selection device according to any one of claims 5 to 8, wherein the satisfaction update unit requests the keyword extraction unit to extract new interesting keywords when the degree of increase in the satisfaction becomes equal to or less than a predetermined value.

10. The program selection device according to any one of Claims 1 to 9, wherein the interest keyword extraction unit excludes the interest keywords extracted in the past from the extraction target.

11. A keyword extraction step of extracting keywords of a predetermined type from program information for a predetermined period acquired by a robot; A keyword calculation step of counting, for each of the keywords, the number of occurrences of the keyword in the program information for the predetermined period and the number of programs in which the keyword has appeared; An interest keyword extraction step of extracting interest keywords based on an interest index that increases as the number of occurrences increases and decreases as the number of programs decreases; An interesting program extraction step of extracting, from a program information database, a program including the interest keyword as an interesting program of the robot; A robot interesting program communication step of outputting information on the interesting program in order to present a viewing request for the interesting program to a user, and a program selection method executed by a computer.

12. A program selection program for causing a computer to function as the program selection device according to any one of Claims 1 to 10.

Images