Method and device for recommending multimedia content and computing equipment

Abstract

The invention provides a method and device for recommending multimedia content, computing equipment, a computer readable storage medium and a computer program product. The method comprises the steps that multiple recommendation language vectors of multiple pieces of multimedia content matched with a user intention are obtained, each recommendation language vector is generated based on a recommendation language of the corresponding multimedia content, and the recommendation language is generated by a first target model based on associated information of the corresponding multimedia content; acquiring a plurality of recommended words corresponding to the recommended word vectors; and providing a recommendation list including the plurality of multimedia contents and recommendation reasons related to the plurality of recommendation languages to the user. In this way, according to the technical scheme, the model is used for generating the recommendation language for the multimedia content, and the recommendation reason is given while the content is recommended to the user, so that the recommendation system has higher recommendation accuracy and interpretability.

Description

Methods, apparatus, and computing devices for recommending multimedia content Technical Field

[0001] This disclosure relates to the field of computer technology, and more specifically, to a method, apparatus, computing device, computer-readable storage medium, and computer program product for recommending multimedia content. Background Technology

[0002] With the widespread adoption of the internet, a vast amount of new multimedia content (such as songs and movies) is released daily. Along with this rapid growth in multimedia content, user tastes are becoming increasingly personalized. Faced with such a massive amount of multimedia resources, accurately capturing user preferences and recommending other multimedia content with similar styles has become a major challenge that current multimedia content recommendation systems urgently need to address.

[0003] Current mainstream multimedia content recommendation methods are based on existing playlists, for example, recommending other content in the playlist containing content that a user is interested in. However, the rationale for recommending multimedia content using this method is unclear, leading to low accuracy and interpretability. Furthermore, since newly released multimedia content is only manually added to playlists by users after a period of time, this type of recommendation method also has a certain degree of lag. Improved multimedia content recommendation methods are needed to meet user needs.

[0004] Summary of the Invention

[0005] In view of this, the present disclosure provides a method, apparatus, computing device, computer-readable storage medium, and computer program product for recommending multimedia content, which can use a target model to generate recommendation text for multimedia content and then provide reasons for recommending the content to the user, thereby making the recommendation system have higher recommendation accuracy and interpretability.

[0006] According to a first aspect of this disclosure, a method for recommending multimedia content is provided, comprising: acquiring multiple recommendation vectors of multiple multimedia content matching a user's intent, wherein each recommendation vector is generated based on a recommendation for the corresponding multimedia content, and the recommendation is generated by a first target model based on the association information of the corresponding multimedia content; acquiring multiple recommendation terms corresponding to the multiple recommendation vectors; and providing the user with a recommendation list including the multiple multimedia content and recommendation reasons related to the multiple recommendation terms.

[0007] According to a second aspect of this disclosure, an apparatus for recommending multimedia content is provided, comprising: a recommendation vector acquisition unit configured to acquire multiple recommendation vectors of multiple multimedia content matching a user's intent, wherein each recommendation vector is generated based on a recommendation for the corresponding multimedia content, and the recommendation is generated by a first target model based on the corresponding multimedia content association information; a recommendation phrase acquisition unit configured to acquire multiple recommendation phrases corresponding to the multiple recommendation vectors; and a recommendation reason provision unit configured to provide the user with a recommendation list including multiple multimedia content and recommendation reasons related to the multiple recommendation phrases.

[0008] According to a third aspect of this disclosure, a computing device is provided, comprising: at least one processing unit; and at least one memory coupled to the at least one processing unit and storing instructions for execution by the at least one processing unit, the instructions, when executed by the at least one processing unit, causing the computing device to perform the method as described in the first aspect of this disclosure.

[0009] According to a fourth aspect of this disclosure, a non-transient computer storage medium is provided, including machine-executable instructions that, when executed by a device, cause the device to perform the method as described in the first aspect of this disclosure.

[0010] According to a fifth aspect of this disclosure, a computer program product is provided, including machine-executable instructions that, when executed by a device, cause the device to perform the method as described in the first aspect of this disclosure.

[0011] It should be understood that the summary section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0012] The above and other objects, features, and advantages of embodiments of the present disclosure will become more readily understood from the following detailed description with reference to the accompanying drawings. In the drawings, several embodiments of the present disclosure will be described by way of example and non-limitation, wherein:

[0013] Figure 1 illustrates a schematic diagram of an environment in which various embodiments of the present disclosure can be implemented;

[0014] Figure 2 shows a schematic block diagram of a system for recommending multimedia content according to an embodiment of the present disclosure;

[0015] Figure 3 shows a schematic flowchart of a method for recommending multimedia content according to an embodiment of the present disclosure;

[0016] Figure 4 shows a schematic block diagram of an apparatus for recommending multimedia content according to an embodiment of the present disclosure; and

[0017] Figure 5 shows a block diagram of a device capable of implementing several embodiments of the present disclosure. Detailed Implementation

[0018] The concept of this disclosure will now be described with reference to various exemplary embodiments shown in the accompanying drawings. It should be understood that these embodiments are described merely to enable those skilled in the art to better understand and further implement this disclosure, and are not intended to limit the scope of this disclosure in any way. It should be noted that similar or identical reference numerals may be used in the figures where feasible, and similar or identical reference numerals may denote similar or identical elements. Those skilled in the art will understand that alternative embodiments of the structures and / or methods described herein may be employed without departing from the principles and concept of this disclosure as described below.

[0019] In the context of this disclosure, the term "comprising" and its various variations may be understood as open-ended terms meaning "including but not limited to"; the term "based on" may be understood as "at least partially based on"; the term "one embodiment" may be understood as "at least one embodiment"; and the term "another embodiment" may be understood as "at least one other embodiment". Other terms that may appear but are not mentioned herein should not be interpreted or limited in a manner contrary to the concept on which the embodiments of this disclosure are based, unless expressly stated otherwise.

[0020] Currently, with the massive growth of multimedia content (e.g., songs, movies, etc.), users are also exhibiting distinct personalized preferences. Current mainstream methods for recommending multimedia content have some drawbacks. Common playlist-based recommendation methods lack accurate recommendation rationale and suffer from a certain lag, failing to recommend the latest multimedia content to users. Some methods use generative artificial intelligence (AI) models to provide recommendation rationale, but the generated rationale is usually based solely on the media name and is not accurate enough.

[0021] To address or mitigate the aforementioned problems and / or other potential issues, embodiments of this disclosure propose a method for recommending multimedia content. This method generates recommendation text and corresponding vectors by mining relevant information (e.g., lyrics, descriptions, melodies, dialogue, etc.) within the multimedia content. When recommending content to a user, the generated vectors can be used to match the user's intent, resulting in a recommendation list of corresponding multimedia content. Furthermore, the reasons for the recommendations are derived from the pre-generated recommendation texts, thereby ensuring the recommendations have good accuracy and interpretability.

[0022] The basic principles and implementation of this disclosure are illustrated below with reference to the accompanying drawings. It should be understood that the exemplary embodiments given are merely intended to enable those skilled in the art to better understand and implement the embodiments of this disclosure, and are not intended to limit the scope of this disclosure in any way.

[0023] Figure 1 illustrates a schematic diagram of an environment 100 in which various embodiments of the present disclosure can be implemented. As shown in Figure 1, environment 100 includes a user terminal 101 operable by a user and a network server 105. Optionally, the user terminal 101 may specifically be a smartphone, tablet computer, laptop computer, smart TV, in-vehicle computer, wearable device (e.g., smart bracelet, smartwatch), etc., with a display function. The user terminal 101 may have a browser or various applications (including system applications and third-party applications, such as music applications, video applications, etc.). The applications on the user terminal 101 may have a favorites list to represent personal preferences and historical information. The user terminal can obtain information through applications, applets, web pages, etc., and display it on the user terminal's screen. The user terminal 101 may support text input, voice input, etc.

[0024] Network server 105 can be an independent physical network server, a network server cluster or distributed system composed of multiple physical network servers, or a cloud network server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms.

[0025] User terminal 101 and network server 105 can communicate via a network. The network between user terminal 101 and network server 105 can be a wired network or a wireless network, such as a local area network (LAN), metropolitan area network (MAN), wide area network (WAN), cellular data communication network, or other electronic networks capable of information exchange.

[0026] As shown in Figure 1, user terminal 101 can transmit data, information, and services to network server 105 via the network. In this example, the user transmits user input 102 requesting multimedia content recommendations to network server 105 through user terminal 101, and then network server 105 returns a recommendation list 103 and recommendation reasons 104 to user terminal 101.

[0027] For example, when using a music application, a user might want the system to recommend some new music and input "Hi, please recommend some music" via voice or text. Accordingly, the user terminal 101 transmits user input 102 to the network server 105. After receiving the user input 102, the network server 105 returns a recommendation list 103 and a recommendation reason 104 to the user terminal 101. For example, the recommendation reason 104 could be "A's song 'Song Title A' records the starting point of his performing career," "B's song 'Song Title B' describes the pain of losing a lover," or "C's song 'Song Title C' conveys nostalgia for past love," etc. The generation of the recommendation list 103 and recommendation reason 104 can be implemented according to embodiments of this disclosure, which will be described in detail below.

[0028] Figure 2 shows a schematic block diagram of a system 200 for recommending multimedia content according to an embodiment of the present disclosure. As shown in Figure 2, system 200 may include an offline environment and an online environment. Users can interact with the online environment and retrieve data resources stored in the offline environment to obtain final multimedia content recommendations.

[0029] In an offline environment, associated information 211 can be extracted from the multimedia content in the multimedia library 210. For example, for a song, associated information 211 may include lyrics, descriptive information (such as a brief introduction), or melody, or optionally, information such as song title, singer, lyricist, and composer. For a movie, associated information 211 may include dialogue, descriptive information, or optionally, information such as movie title, director, and lead actors.

[0030] In an offline environment, generative models, such as Large Language Models (LLMs), can also be used to generate summaries of the associated information 211 as recommendations 212 for each multimedia content, and each recommendation 212 can be added to a recommendation corpus 222. Large Language Models refer to natural language processing models with a large number of parameters and complex structures. These models are trained using deep learning techniques and are capable of understanding and generating natural language text. Large Language Models are commonly used for various natural language processing tasks, such as text generation, machine translation, and sentiment analysis. They can handle large-scale corpora and capture the complex structure and semantic relationships of language during training. For example, for the song "A," a large language model can be used to analyze its lyrics and generate the following recommendation 212: "The song 'A' is a love song sung by singer A. The lyrics review his career and talk about how he became famous in just one year."

[0031] Example prompts for a large language model are as follows:

[0032] In an offline environment, a corresponding recommendation vector 213 can also be generated based on the recommendation 212. For example, a text-to-vector model that can transform text information into vectors that can express the semantics of the text can be used. For example, the BAAI General Embedding Large model (BGE-Large) can be used to generate the recommendation vector 213 based on the recommendation 212. It is understood that other models are also applicable, such as Word2Vec, GloVe, BERT, One-Hot Model, Bag of Words Model, etc., and this disclosure does not limit them.

[0033] In some implementations, when recommending 212 directly to generate recommending vector 213, if recommending 212 includes information such as multimedia names (e.g., song titles) or artist names (e.g., singer names), the generated recommending vector 213 will be significantly affected by this information. In subsequent online recommendation processes, this can lead to recommendations focusing on multimedia titles and works by the same artist. To improve the diversity of recommendations, the large language model can be instructed to avoid using multimedia names or artist names when generating recommending 212. In some implementations, the large language model may not strictly follow the prompts when generating recommending 212. In this case, information such as multimedia names or artist names can be additionally removed from the recommending 212 generated by the large language model, and the recommending vector 213 can be generated using the recommended 212 with this information removed.

[0034] In an offline environment, the similarity between each recommendation vector 213 and all other recommendation vectors 213 can also be calculated. Optionally, the cosine similarity between recommendation vectors 213 can be calculated to find several recommendation vectors that are closest to each recommendation vector, and the recommendation vectors 213 and their similarity scores can be associated and saved in the recommendation vector library 223.

[0035] Therefore, a recommendation lexicon 222 and a recommendation vector 223 are constructed in an offline environment. Based on this, users can obtain recommendation lists and corresponding recommendation reasons in an online environment. Specifically, this interacts with the multimedia application 220 to provide users with their needs. The intent analysis unit 221 can utilize LLM to perform intent analysis and extraction on user input. For example, if the user input specifies which multimedia content to recommend, the intent analysis unit 221 can recommend multimedia content based on the user's explicit needs; or if the user input does not include explicit recommendation needs, such as "Hi, play some music," the intent analysis unit 221 can search the user's music playback history and make recommendations based on that history.

[0036] In some embodiments, the recommender system 224 can search the recommender lexicon 222 and the recommender vector lexicon 223 based on the intent analysis results of the intent analysis unit 221 to determine the final recommender list. Then, the recommender reasoning unit 225 can obtain the recommender's phrase corresponding to each multimedia content in the recommender list and use the second target model to generate a recommender reason and provide it to the multimedia application 220. The process of providing recommended content and recommender reasons to users in an online environment is further described below with reference to FIG3.

[0037] Figure 3 illustrates a flowchart of a method 300 for recommending multimedia content according to some embodiments of the present disclosure. In some embodiments, method 300 may be implemented by, for example, a web server 105 as shown in Figure 1. It should be understood that method 300 may also include additional actions not shown and / or actions shown may be omitted, and the scope of the present disclosure is not limited in this respect.

[0038] As shown in Figure 3, in box 310, method 300 can obtain multiple recommendation vectors for multiple multimedia content sets matching the user's intent. In some embodiments, the multimedia content can be a song, and its associated information can be lyrics, descriptive information, and / or melody. In some embodiments, the multimedia content can also be a movie, and its associated information can be dialogue and / or descriptive information. This disclosure does not limit the type of multimedia content.

[0039] Referring to Figure 2, in an offline environment, a recommendation is generated based on the association information of multimedia content using a first target model (e.g., a generative model), and a recommendation vector is generated based on the generated recommendation using a text-to-vector model. The recommendation vector and the recommendation can be distributed and stored in a recommendation database 222 and a recommendation vector database 223.

[0040] In some embodiments, the cosine of the angle between the recommendation vectors 213 can be used as the similarity. The recommendation vector library 223 can store the vectors most similar to the recommendation vector of each multimedia content. In some embodiments, other multimedia content can also be sorted based on similarity for each multimedia content, and the relationship between each multimedia content and the top N most similar other multimedia content can be saved in the recommendation vector library 223, where N is a positive integer. Table 1 below shows exemplary similarity entries between songs stored in the recommendation vector library 223, each entry including the five most similar songs, sorted from highest to lowest similarity.

[0041] Table 1. List of Song Similarities

[0042] In some embodiments, the user intent may indicate at least one multimedia content, and obtaining multiple recommendation vectors includes: determining multiple matching recommendation vectors 213 based on the similarity between the recommendation vector 213 of at least one multimedia content and other recommendation vectors 213. In some embodiments, the user intent may involve multiple multimedia content, in which case a certain number of recommendation vectors (i.e., the corresponding multimedia content) can be obtained for each multimedia content, and the content to be recommended can be determined after integration (e.g., sorting). In some embodiments, the user intent may also be a piece of text (e.g., "I want to listen to sentimental songs"), from which a vector representing its semantics can be obtained, and thus this vector can be used to search for similar vectors in the vector library 223.

[0043] In box 320, method 300 can obtain multiple recommendation phrases corresponding to multiple recommendation phrase vectors. In some embodiments, recommendation phrase vectors 213 can be searched one by one in the recommendation phrase library 222 to obtain the corresponding recommendation phrases 212.

[0044] In box 330, method 300 can provide the user with a recommendation list including multiple multimedia content items and recommendation reasons associated with multiple recommendation statements. Thus, the recommendation list and the recommendation reasons for each multimedia content item can be displayed on the user's terminal.

[0045] The recommendation list may include a large amount of multimedia content, for example, dozens of items. If the recommendation text 212 of each multimedia content is simply concatenated and combined, the final recommendation reason text will be too long, making it inconvenient to display on small-screen devices. In some embodiments, each recommendation text obtained from the recommendation text library 222 can be input into a second target model to generate recommendation reasons for presentation to the user. For example, the output results regarding recommendation reasons can be limited to a specified number of characters in the input prompts of the second target model to facilitate quick browsing by the user.

[0046] To facilitate understanding, exemplary recommendation reasons derived from embodiments of this disclosure are given below. Assume the user intends to receive recommendations for "sentimental" playlists. It can be seen that the recommendation reasons incorporate lyric information, resulting in more accurate recommended playlists.

[0047] "Here are some deeply moving songs for you, such as Singer A's 'Song A,' which records the beginning of his performing career; Singer B's 'Song B,' which tells the story of the pain of losing a lover; Singer C's 'Song C,' which depicts the struggle and disappointment in love; and Singer D's 'Song D,' which conveys the longing for past love. These songs are full of emotion, either telling stories or evoking memories, allowing the music to take you through time and feel the resonance deep in your heart."

[0048] Additionally, a user might be interested in certain content and click or tap on a recommendation reason for that content. In this case, in response to the user's choice of recommendation reason, a detailed recommendation message can be requested from the server, and the detailed information of the recommendation message corresponding to the selected recommendation reason can be displayed on the user's terminal.

[0049] In some embodiments, the recommendation lexicon and recommendation vector lexicon are updated in response to the addition of new multimedia content to the multimedia lexicon. Compared to recommendation methods based on manual playlists, this approach can promptly recommend the latest multimedia content to users, avoiding the lag problem commonly found in current recommendation systems.

[0050] Exemplary embodiments of the present disclosure have been described above with reference to Figures 1 to 3. Compared to existing multimedia content recommendation schemes, the multimedia content recommendation scheme of the present disclosure can use a model to generate recommendation text for multimedia content, and then provide reasons for recommending the content to the user, thereby making the recommendation system have higher recommendation accuracy and interpretability.

[0051] Figure 4 shows a schematic block diagram of an apparatus 400 for recommending multimedia content according to an embodiment of the present disclosure. As shown in Figure 4, the apparatus 400 includes: a recommendation vector acquisition unit 410, a recommendation acquisition unit 420, and a recommendation reason provision unit 430.

[0052] In some embodiments, the recommendation vector acquisition unit 410 is configured to acquire multiple recommendation vectors for multiple multimedia content that match the user's intent, wherein each recommendation vector is generated based on the recommendation of the corresponding multimedia content, and the recommendation is generated by a first target model based on the association information of the corresponding multimedia content; the recommendation vector acquisition unit 420 is configured to acquire multiple recommendation words corresponding to the multiple recommendation vectors; and the recommendation reason providing unit 430 is configured to provide the user with a recommendation list including multiple multimedia content and recommendation reasons related to the multiple recommendation words.

[0053] It should be noted that further actions or steps shown in Figures 1 to 3 can be implemented using the device 400 shown in Figure 4. For example, device 400 may include more modules or units to implement the actions or steps described above, or some of the units or modules shown in Figure 4 may be further configured to implement the actions or steps described above. This will not be repeated here.

[0054] Figure 5 shows a schematic block diagram of an example device 500 that can be used to implement embodiments of the present disclosure. As shown, device 500 includes a computing unit 501, which can perform various appropriate actions and processes according to computer program instructions stored in read-only memory (ROM) 502 or loaded from storage unit 506 into random access memory (RAM) 503. Various programs and data required for the operation of device 500 may also be stored in RAM 503. The computing unit 501, ROM 502, and RAM 503 are interconnected via bus 504. Input / output (I / O) interface 605 is also connected to bus 504.

[0055] Multiple components in device 500 are connected to I / O interface 505, including: input unit 506, such as keyboard, mouse, etc.; output unit 507, such as various types of monitors, speakers, etc.; storage unit 508, such as disk, optical disk, etc.; and communication unit 509, such as network card, modem, wireless transceiver, etc. Communication unit 509 allows device 500 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0056] The computing unit 501 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 501 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 501 performs the various methods and processes described above, such as method 300. For example, in some embodiments, method 300 may be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 508. In some embodiments, part or all of the computer program may be loaded and / or installed on device 500 via ROM 502 and / or communication unit 509. When the computer program is loaded into RAM 503 and executed by the computing unit 501, one or more steps of method 300 described above may be performed. Alternatively, in other embodiments, the computing unit 501 may be configured to perform method 300 by any other suitable means (e.g., by means of firmware).

[0057] In some embodiments, the methods and processes described above can be implemented as a computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of this disclosure.

[0058] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination thereof. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0059] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper cables, fiber optic cables, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to computer-readable storage media within the respective computing / processing device.

[0060] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​and conventional procedural programming languages. The computer-readable program instructions may execute entirely on a user's computer, partially on a user's computer, as a standalone software package, partially on a user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0061] These computer-readable program instructions can be provided to a processing unit of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processing unit of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner. Thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0062] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0063] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of devices, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0064] The various embodiments of this disclosure have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or technical improvements to the technology in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A method for recommending multimedia content, comprising: obtaining a plurality of recommendation phrase vectors of a plurality of multimedia contents matching a user intent, wherein each recommendation phrase vector is generated based on a recommendation phrase of a corresponding multimedia content, the recommendation phrase being generated by a first target model based on associated information of the corresponding multimedia content; obtaining a plurality of recommendation phrases corresponding to the plurality of recommendation phrase vectors; and providing, to the user, a recommendation list including the plurality of multimedia contents and a recommendation reason related to the plurality of recommendation phrases. 2.The method of claim 1, further comprising generating a recommendation phrase library, wherein comprising: for a multimedia content in a multimedia library, generating, as the recommendation phrase, a summary of the associated information of the multimedia content using the first target model; and saving the recommendation phrase into the recommendation phrase library. 3.The method of claim 2, further comprising: removing at least one of a multimedia name, an author name or an artist name from the summary of the associated information of the multimedia content. 4.The method of claim 2 or 3, further comprising generating a recommendation vector library, wherein comprising: for a recommendation phrase in the recommendation phrase library, generating a recommendation phrase vector; calculating a similarity between the recommendation phrase vector of the recommendation phrase and recommendation phrase vectors of other recommendation phrases; and saving the recommendation phrase vector and the similarity into the recommendation vector library in association. 5.The method of claim 4, further comprising: in response to adding a new multimedia content in the multimedia library, updating the recommendation phrase library and the recommendation vector library. 6.The method of claim 1, wherein the user intent indicates at least one multimedia content, and the obtaining the plurality of recommendation phrase vectors comprises: determining a plurality of matched recommendation phrase vectors based on a similarity between the recommendation phrase vector of the at least one multimedia content and other recommendation phrase vectors. 7.The method of claim 1, further comprising: generating the recommendation reason based on the plurality of recommendation phrases using a second target model. 8.The method of claim 7, further comprising: in response to a user selection of the recommendation reason, displaying detailed information of a recommendation phrase corresponding to the selected recommendation reason. 9.The method of claim 1, wherein the multimedia content comprises a song, and the associated information comprises at least one of lyrics, description information and melody. 10.An apparatus for recommending multimedia content, comprising: a recommendation phrase vector obtaining unit configured to obtain a plurality of recommendation phrase vectors of a plurality of multimedia contents matching a user intent, wherein each recommendation phrase vector is generated based on a recommendation phrase of a corresponding multimedia content, the recommendation phrase being generated by a first target model based on associated information of the corresponding multimedia content; a recommendation phrase obtaining unit configured to obtain a plurality of recommendation phrases corresponding to the plurality of recommendation phrase vectors; and a recommendation reason providing unit configured to provide, to the user, a recommendation list including the plurality of multimedia contents and a recommendation reason related to the plurality of recommendation phrases. 11.A computing device, comprising: at least one processing unit; ​ ​ ​ ​ at least one memory coupled to the at least one processing unit and storing instructions for execution by the at least one processing unit, the instructions, when executed by the at least one processing unit, cause the computing device to perform the method of any of claims 1 to 9.

12. A non-transitory computer storage medium comprising machine executable instructions that, when executed by a device, cause the device to perform the method of any of claims 1 to 9.

13. A computer program product comprising machine executable instructions that, when executed by a device, cause the device to perform the method of any of claims 1 to 9.

Images