Media content recommendation through chatbots

1. A method for recommending media content through intelligent automated chatting, comprising:

receiving a message in a conversation;

identifying a new topic based on the message and context of the conversation;

identifying a media content from a set of media contents based on the new topic;

generating a short video clip for the media content using a neural network model including a convolution neural network (CNN) part and a recurrent neural network (RNN) part, and wherein the generating the short video clip further comprises:

dividing the media content into a plurality of clips;

mapping the plurality of clips into a plurality of vectors through the CNN part;

selecting a part of vectors representing clips that should be remained through the RNN part; and

generating the short video clip based on the part of vectors; and

providing a recommendation of the media content in the conversation.

2. The method of claim 1 , wherein the identifying a new topic comprises:

identifying the new topic based further on a knowledge graph related to the set of media contents.

3. The method of claim 2 , wherein the knowledge graph comprises a first kind of data indicating attributes of the media contents and a second kind of data indicating similarity between media contents.

4. The method of claim 1 , wherein the identifying a media content from a set of media contents comprises:

scoring matching rates of at least part of the set of media contents based on the new topic and a user profile of the user; and

selecting the media content from the set of media contents based on the matching rates.

5. The method of claim 4 , wherein the scoring matching rates comprises scoring the matching rates based further on at least one of:

a knowledge graph related to the set of media contents;

the user's emotion in the context of the conversation; and

bidding information of at least one of the set of media contents.

6. The method of claim 1 , wherein the media content comprises at least one of a television program and a video advertisement.

7. The method of claim 6 , further comprising performing at least one of the following in response to the user's feedback to the recommendation:

playing the television program;

booking the television program; and

recording the television program.

8. The method of claim 6 , wherein the providing a recommendation of the media content comprises providing at least one of:

comments for an actor of the media content;

comments for the media content;

a representative image of the media content, or the short video clip of the media content, or the media content; and

broadcasting information related to the television program or purchasing information related to the video advertisement.

9. The method of claim 8 , further comprising at least one of:

generating the comments for the actor from a description text about the actor by using a neural network model;

generating the comments for the media content from a description text about the media content by using a neural network model; and

generating the short video clip from the media content by using the neural network model.

10. An apparatus for recommending media content through intelligent automated chatting, comprising:

at least one processor; and

memory including instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to:

receive a message in a conversation;

identify a new topic based on the message and context of the conversation;

identify a media content from a set of media contents based on the new topic;

generate a short video clip for the media content using a neural network model including a convolution neural network (CNN) part and a recurrent neural network (RNN) part, and wherein the generating the short video clip comprises:

divide the media content into a plurality of clips;

map the plurality of clips into a plurality of vectors through the CNN part;

select a part of vectors representing clips that should be remained through the RNN part; and

generate the short video clip based on the part of vectors; and

provide a recommendation of the media content in the conversation.

11. The apparatus of claim 10 , wherein the new topic is identified based further on a knowledge graph related to the set of media contents.

12. The apparatus of claim 11 , wherein the knowledge graph comprises a first kind of data indicating attributes of the media contents and a second kind of data indicating similarity between media contents.

13. The apparatus of claim 10 , wherein the media content is identified from the set of media contents by:

scoring matching rates of at least part of the set of media contents based on the new topic and a user profile of the user; and

selecting the media content from the set of media contents based on the matching rates.

14. The apparatus of claim 13 , wherein the matching rates are scored based further on at least one of:

a knowledge graph related to the set of media contents;

the user's emotion in the context of the conversation; and

bidding information of at least one of the set of media contents.

15. The apparatus of claim 10 , wherein the media content comprises at least one of a television program and a video advertisement.

16. The apparatus of claim 15 , further comprising performance of at least one of the following in response to the user's feedback to the recommendation:

playing the television program;

booking the television program; and

recording the television program.

17. The apparatus of claim 15 , wherein the recommendation provides at least one of:

comments for an actor of the media content;

comments for the media content;

a representative image of the media content, or a short video clip of the media content, or the media content; and

broadcasting information related to the television program or purchasing information related to the video advertisement.

18. The apparatus of claim 17 , further comprising at least one of:

comments for the actor generated from a description text about the actor;

the comments for the media content generated from a description text about the media content; and

the short video clip generated from the media content.

19. A computer system, comprising:

one or more processors; and

a memory storing computer-executable instructions that, when executed, cause the one or more processors to:

receive a message from a user in a conversation with the user;

identify a new topic based on the message and context of the conversation;

identify a media content from a set of media contents based on the new topic;

generate a short video clip for the media content using a neural network model including a convolution neural network (CNN) part and a recurrent neural network (RNN) part, and wherein the generating the short video clip further comprises:

divide the media content into a plurality of clips;

map the plurality of clips into a plurality of vectors through the CNN part;

select a part of vectors representing clips that should be remained through the RNN part; and

generate the short video clip based on the part of vectors; and

provide a recommendation of the media content in the conversation.