Information generation method and apparatus, device and storage medium

Abstract

Provided are an information generation method and apparatus, a device and a storage medium. The method comprises: at least on the basis of reference information of a target user, generating a prompt word input for a machine learning model (210); providing the prompt word input to the machine learning model to obtain a model output of the machine learning model, wherein the model output at least indicates a set of media items of a media item type and description information associated with the set of media items (220); and providing the set of media items and the description information to the target user (230).

Description

Method, apparatus, device and storage medium for information generation TECHNICAL FIELD

[0001] Exemplary implementations of the present disclosure generally relate to the field of computers, and particularly relate to a method, apparatus, device and computer-readable storage medium for information generation. BACKGROUND

[0002] With the rapid development of computer technology, more and more applications and platforms are currently designed to provide various services to users. For example, a user can publish, browse, view media item content (which can also be referred to as media items, content items, media data, etc.) of various media item types (such as video types, audio types, picture types, etc.) in an application. Multiple media items are often output to the user in the form of an information stream, so that the user can browse and view them in sequence.

[0003] SUMMARY

[0004] In a first aspect of the present disclosure, a method for information generation is provided. The method comprises: generating a prompt word input for a machine learning model based at least on reference information of a target user, the reference information comprising at least behavior data related to a media item type obtained with authorization of the target user, the prompt word input being used to guide the machine learning model to recommend media items corresponding to the media item type for the target user; obtaining a model output of the machine learning model by providing the prompt word input to the machine learning model, the model output at least indicating a set of media items of the media item type and description information associated with the set of media items; and providing the set of media items and the description information to the target user.

[0005] In a second aspect of the present disclosure, an apparatus for information generation is provided. The apparatus comprises: a prompt word generation module configured to generate a prompt word input for a machine learning model based at least on reference information of a target user, the reference information comprising at least behavior data related to a media item type obtained with authorization of the target user, the prompt word input being used to guide the machine learning model to recommend media items corresponding to the media item type for the target user; an output acquisition module configured to obtain a model output of the machine learning model by providing the prompt word input to the machine learning model, the model output at least indicating a set of media items of the media item type and description information associated with the set of media items; and an information providing module configured to provide the set of media items and the description information to the target user.

[0006] In a third aspect of the disclosure, an electronic device is provided. The electronic device includes 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 cause the electronic device to perform the method according to the first aspect of the disclosure.

[0007] In a fourth aspect of the disclosure, a computer-readable storage medium is provided, having stored thereon a computer program which, when executed by a processor, causes the processor to implement the method according to the first aspect of the disclosure.

[0008] According to a fifth aspect of the disclosure, a computer program product is provided, comprising a computer program, wherein the computer program, when executed by a processor, implements the method according to the first aspect of the disclosure.

[0009] It is to be understood that the details set forth herein are not intended to limit the key or critical features of the implementations of the disclosure, nor are they intended to limit the scope of the disclosure. Other features of the disclosure will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0010] The above and other features, advantages, and aspects of the implementations of the disclosure will become more apparent by describing in detail some implementations thereof with reference to the annexed drawings. In the drawings:

[0011] FIG. 1 shows a schematic diagram of an example environment in which embodiments of the disclosure can be implemented;

[0012] FIG. 2 shows a flowchart of a process for information generation according to some embodiments of the disclosure;

[0013] FIGS. 3A and 3B show schematic diagrams of examples of system prompt words according to some embodiments of the disclosure;

[0014] FIG. 4 shows an example model output according to some embodiments of the disclosure;

[0015] FIG. 5 shows a schematic structural block diagram of an apparatus for information generation according to some embodiments of the disclosure; and

[0016] FIG. 6 shows a block diagram of an electronic device in which one or more embodiments of the disclosure can be implemented. DETAILED DESCRIPTION

[0017] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the drawings, it is understood that the present disclosure can be implemented in various forms and should not be interpreted as being limited to the embodiments set forth herein, but rather, these embodiments are provided so as to more completely and thoroughly understand the present disclosure. It is understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0018] In the description of embodiments of the present disclosure, the term "comprising" and its conjugations should be understood to encompass the meaning of "including but not limited to". The term "based on" should be understood as "based at least in part on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The term "some embodiments" should be understood as "at least some embodiments". Other explicit and implicit definitions can also be included below.

[0019] In this document, unless explicitly stated, performing a step "in response to A" does not mean that the step is performed immediately after A, but can include one or more intermediate steps.

[0020] It can be understood that the data involved in the technical solutions of the present disclosure (including but not limited to the data itself, the obtaining or use of the data) should comply with the requirements of relevant laws and regulations and relevant provisions.

[0021] It can be understood that before using the technical solutions disclosed in the embodiments of the present disclosure, the type of personal information involved in the present disclosure, the scope of use, the use scenario, etc. should be informed to the user and the authorization of the user should be obtained through appropriate means according to relevant laws and regulations.

[0022] For example, in response to receiving the active request of the user, a prompt information is sent to the user to explicitly prompt the user that the operation requested to be performed will require obtaining and using the personal information of the user, so that the user can voluntarily choose whether to provide the personal information to the software or hardware such as electronic device, application, server or storage medium, etc. performing the operation of the technical solutions of the present disclosure according to the prompt information.

[0023] As an optional but non-limiting implementation manner, in response to receiving the active request of the user, the manner of sending the prompt information to the user can be, for example, the manner of pop-up window, and the prompt information can be presented in the form of text in the pop-up window. In addition, the pop-up window can also carry selection controls for the user to select "agree" or "disagree" to provide personal information to the electronic device.

[0024] It can be understood that the above notification and user authorization process is only illustrative and does not limit the implementation manner of the present disclosure, and other manners meeting the relevant laws and regulations can also be applied to the implementation manner of the present disclosure.

[0025] As used in this paper, the term "model" refers to a model that learns the relationship between inputs and outputs from training data, enabling it to generate corresponding outputs for a given input after training. Model generation can be based on machine learning techniques. Deep learning is a machine learning algorithm that processes inputs and provides corresponding outputs using multiple layers of processing units. A neural network model is an example of a deep learning-based model. In this paper, "model" may also be referred to as a "machine learning model," "learning model," "machine learning network," or "learning network," and these terms are used interchangeably.

[0026] A neural network is a machine learning network based on deep learning. A neural network processes input and provides a corresponding output, typically consisting of an input layer, an output layer, and one or more hidden layers between the input and output layers. Neural networks used in deep learning applications often include many hidden layers, thus increasing the network's depth. The layers of a neural network are connected sequentially, so that the output of the previous layer is provided as the input to the next layer. The input layer receives the input to the neural network, while the output layer's output serves as the final output. Each layer of a neural network includes one or more nodes (also called processing nodes or neurons), each node processing the input from the layer above.

[0027] Machine learning typically comprises three phases: training, testing, and application (also known as inference). In the training phase, a given model is trained using a large amount of training data, iteratively updating parameter values ​​until the model can consistently generate inferences that meet the expected goals from the training data. Through training, the model can be considered to have learned the relationship between inputs and outputs (also known as an input-output mapping) from the training data. The parameter values ​​of the trained model are determined. In the testing phase, test inputs are applied to the trained model to test whether it can provide the correct output, thus determining the model's performance. The testing phase can sometimes be integrated into the training phase. In the application or inference phase, the trained model can be used to process actual model inputs based on the trained parameter values ​​to determine the corresponding model output.

[0028] Figure 1 illustrates a schematic diagram of an example environment 100 in which embodiments of the present disclosure can be implemented. As shown in Figure 1, the example environment 100 may include a terminal device 110.

[0029] In the example environment 100, a terminal device 110 has an application 115 installed therein. The application 115 can be any suitable type of application capable of providing media content, examples of which can include, but are not limited to, a social application, an audio / video application, a media item playing application, a broadcasting application, and the like, embodiments of the present disclosure are not limited in this respect. A user 140 can interact with the application 115 via the terminal device 110 and / or its attached devices. In the environment 100 of FIG. 1, the terminal device 110 can render an interaction page via the application 115 if the application 115 is in an active state. The interaction page can be any suitable type of page that can support the user 140 to input any suitable type of data and present any suitable type of media item to the user 140.

[0030] In some embodiments, the terminal device 110 communicates with a server 130 to implement the provision of services for the application 115. The terminal device 110 can be any type of mobile terminal, fixed terminal, or portable terminal including a mobile handset, a desktop computer, a laptop computer, a notebook computer, a netbook computer, a tablet computer, a media computer, a multimedia tablet, a palmtop computer, a portable gaming terminal, a VR / AR device, a Personal Communication System (PCS) device, a personal navigation device, a Personal Digital Assistant (PDA), an audio / video player, a digital camera / camcorder, a positioning device, a television receiver, a radio broadcast receiver, an electronic book device, a game device, or any combination thereof, including accessories and peripherals of these devices, or any combination thereof. In some embodiments, the terminal device 110 can also support any type of interface to the user (such as "wearable" circuitry, etc.).

[0031] The server 130 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud service, cloud database, cloud computing, cloud function, cloud storage, network service, cloud communication, middleware service, domain name service, security service, content distribution network, and basic cloud computing services such as big data and artificial intelligence platform. The server 130 may, for example, include a computing system / server, such as a mainframe, an edge computing node, a computing device in a cloud environment, and the like. The server 130 can provide background services for the application 115 in the terminal device 110 that supports a virtual scene.

[0032] A communication connection can be established between the server 130 and the terminal device 110. The communication connection can be established in a wired manner or a wireless manner. The communication connection can include, but is not limited to, a Bluetooth connection, a mobile network connection, a Universal Serial Bus (USB) connection, a Wireless Fidelity (WiFi) connection, and the like, and embodiments of the present disclosure are not limited in this regard. In embodiments of the present disclosure, the server 130 and the terminal device 110 can implement signaling interaction through the communication connection therebetween.

[0033] In some embodiments, to improve the user experience of the user browsing the media items, the terminal device 110 can also determine the media items 152 to be provided to the user 140 and the description information 154 associated with the media items 152 by means of the machine learning model 120. It can be understood that the media items 152 can include a plurality of media items, and the description information 154 can include a plurality of description information associated with the plurality of media items. The machine learning model 120 can be deployed locally on the terminal device 110, or on other electronic devices (such as the server 130). If the machine learning model 120 is deployed on other electronic devices, the terminal device 110 can invoke the machine learning model 120 to determine the media items 152 and the description information 154 via a communication connection between the terminal device 110 and the other electronic devices. The description information 154 can be used to introduce the media items 152, for example, so that the user can know the summary information, background information, and the like of the media items 152.

[0034] It should be understood that the structure and function of the various elements in the environment 100 are described for illustrative purposes only, without implying any limitation on the scope of the present disclosure.

[0035] When providing a certain type of media content, such as recommended music, to a user, it is desirable to provide description information about the music at the same time. Traditionally, a playlist needs to be manually constructed and a pre-defined human commentary needs to be inserted between multiple pieces of music in the playlist to achieve the introduction of the music to the user while the music is being played, and to achieve continuous music playing with interlude commentary. This relies on a real human anchor to perform the human commentary, which requires a large human cost. In addition, the commentary of the human commentary is often fixed, and the user cannot interrupt or change it, which can affect the user's interactive experience of multiple media items.

[0036] Therefore, according to an embodiment of the present disclosure, an improved scheme of information generation is proposed. In the scheme, a prompt word input for a machine learning model is generated based on at least reference information of a target user. The reference information at least includes behavior data related to a media item type, which is obtained with authorization of the target user. The prompt word input is used to guide the machine learning model to recommend media items of the media item type to the target user. By providing the prompt word input to the machine learning model, a model output of the machine learning model is obtained, which at least indicates a set of media items of the media item type and description information associated with the set of media items. The set of media items and the description information are provided to the target user.

[0037] In this way, the set of media items and the description information associated with the set of media items can be personalized for the user based on the reference information of the user, which helps to improve the interactivity of the user in browsing the media items and the user experience, enhance the richness and attractiveness of the description information, and improve the flexibility and adaptability of generating the description information and the set of media items.

[0038] Some example embodiments of the present disclosure will be described below with continuous reference to the drawings.

[0039] FIG. 2 shows a schematic diagram of an example process 200 of information playing according to some embodiments of the present disclosure. The process 200 can be implemented at the terminal device 110. For ease of discussion, the process 200 will be described with reference to the environment 100 of FIG. 1. It should be noted that the operations performed by the aforementioned terminal device 110 and the operations subsequently described as performed by the terminal device 110 can be specifically performed by a related application (e.g., the application 115) installed on the electronic device 110 and / or the server 130.

[0040] At block 210, the terminal device 110 generates a prompt word input for a machine learning model based on at least reference information of a target user. The reference information at least includes behavior data related to a media item type, which is obtained with authorization of the target user. The prompt word input is used to guide the machine learning model to recommend media items of the media item type to the target user. In some examples, the media item type can include a music type. In some examples, the media item type of interest can also include a video type, an image type, and the like. For ease of description, only the music type will be exemplarily described below.

[0041] The reference information of the target user can be obtained through a data collection process. The main goal of the data collection is to obtain sufficient user preference information, including but not limited to music types, singers, bands, and user context needs (such as music types preferred when exercising, working, or relaxing). The collection process should follow the principles of user privacy protection and data security to ensure the safety and privacy of user data are not infringed.

[0042] In some embodiments, the terminal device 110 can obtain information about the media item type collected from the target user, and determine the reference information or a part of the reference information based on the obtained information. In some examples, the reference information is obtained through direct interaction of the agent with the user. This includes asking the user about music listening requirements, music selection in specific situations, etc. For example, the terminal device 110 can interact with the target user by providing a questionnaire to the target user, having a conversation with the target user via a conversation window, a feedback mechanism, etc. The information about the media item type can indicate behavior data of the target user on the media item type.

[0043] In some embodiments, alternatively or additionally, the terminal device 110 can obtain historical interaction information of the target user on at least one media item of the media item type, and determine the reference information or a part of the reference information based on the obtained information.

[0044] In some embodiments, the terminal device 110 can obtain historical playback records, historical like / comment / favorite records, etc. of the target user on at least one media item of the media item type, and determine the historical interaction information of the target user on at least one media item of the media item type by analyzing these records. For example, the terminal device 110 can obtain music playback records of multiple musics of the target user, and determine the historical interaction information of the target user on at least one music by analyzing the music playback records. Based on the playback history records or other interaction information, the historical interaction information can indicate the most frequently played media item or corresponding style (e.g. music style) of the target user, the user (e.g. singer, composer, etc.) associated with the most frequently played media item, the frequency of playing, etc., the recently favorite media item or corresponding style of the target user, etc.

[0045] In some embodiments, alternatively or additionally, the terminal device 110 can obtain third-party reference information of the target user from a target application platform based on authorization of the target user, and determine the reference information or a part of the reference information based on the obtained information. The terminal device 110 can obtain the third-party reference information of the target user from the target application platform based on authorization of the target user. The target application platform here can be an application different from the application 115, which can also be any appropriate type of application. The target application platform and the application 115 can be the same type of application, e.g. can both be media item playback applications. The terminal device 110 can obtain music playback information of the target user from the target application platform based on authorization of the target user, which can indicate public preferences of the target user on music and additional dimensions of music-related activities.

[0046] In some embodiments, alternatively or additionally, the terminal device 110 determines the reference information based on at least one of the information related to the environment context of the target user, and determines the reference information based on the obtained information or a part of the reference information. The terminal device 120 can also obtain the environment context information related to the target user, for example, with the authorization of the target user. The environment context information can include time information, location information, activity information, etc. The environment context information can indicate the behavior data of the media item in a specific scenario of the target user.

[0047] In some embodiments, to ensure the quality of the obtained information and thus the quality of the reference information, the terminal device 110 can further process the obtained information to remove irrelevant data of the media item and error data. In some embodiments, after obtaining the above information, the terminal device 110 can determine the reference information in any appropriate manner. For example, the terminal device 110 can determine the reference information based on the obtained information by means of machine learning and data analysis techniques.

[0048] The reference information at least includes the behavior data related to the media item type with the authorization of the target user. For example, the reference information can include the behavior data associated with the media item of the music type. In some embodiments, the reference information can also include the data related to the playing scenario of the media item type. For example, the data related to the playing scenario of the media item type can indicate that the media item of the music type is played in the traffic scenario. In some embodiments, the reference information can also include the data related to the playing device of the media item type. For example, the data related to the playing device of the media item type can indicate the type of the device playing the media item of the music type. In some embodiments, the reference information can also include the data related to the playing time of the media item type. For example, the data related to the playing time of the media item type can indicate the time period of playing the media item of the music type by the user. In some embodiments, the reference information can also include the data related to the creation background of the media item type. For example, the data related to the creation background of the media item type can indicate the creation background (e.g., the composer, the lyricist, the singer, the band, etc.) of the media item of the music type.

[0049] In some embodiments, the behavior data can indicate a total content style of the media items that the target user is interested in. Taking the media items as music for example, the total content style can be folk, jazz, classical, rock, etc. In some embodiments, the behavior data can also indicate content styles of the media items that the target user is interested in respectively within different scenarios. The different scenarios can include, for example, a sports scenario, a work scenario, a pre-sleep scenario, etc. In some embodiments, the behavior data can also indicate historical playback behavior of the target user on the media items. The historical playback behavior can indicate, for example, a playback time, a playback frequency, a playback device type, a playback system / version, etc. of the target user on the media items.

[0050] In some embodiments, the behavior data can also indicate a requirement of the target user on the description information of the media items. Taking the media items as music for example, the requirement of the target user on the description information of the media items can include, for example, an expectation of the user to provide a brief description about the music, to provide an introduction about the creation background or the creator, to provide an explanation about the creation style, etc.

[0051] At block 220, the terminal device 110 obtains a model output of the machine learning model 120 by providing a prompt input to the machine learning model 120, the model output indicating at least a set of media items of a media item type and description information associated with the set of media items. The machine learning model 120 herein can be based on any suitable model structure, including but not limited to a Transformer model, a convolutional neural network (CNN), a recurrent neural network (RNN), a deep neural network (DNN), etc. In some embodiments, the machine learning model 120 can be a language model (LM) or a large language model (LLM). Such a machine learning model 120 is capable of performing content generation through the guidance of the prompt input to handle different types of tasks. In embodiments of the present disclosure, the machine learning model 120 is capable of at least assisting in generating a media item recommendation for the target user and providing description information about the recommended media item.

[0052] In some embodiments, the terminal device 110 can obtain a predetermined prompt input template for the machine learning model 120 and generate the prompt input based on the prompt input template and the reference information. The terminal device 110 can generate the prompt input for the machine learning model 120 by filling in the reference information of the target user that needs to be recommended into the prompt input template, for example.

[0053] In some embodiments, the terminal device 110 can extract first guidance information in the prompt word input from the behavior data in the reference information, the first guidance information including a recommendation requirement for the media item and / or a generation requirement for the description information of the media item. Based on such guidance information, the machine learning model 120 can determine to recommend the media item that meets the user's interest to the user, and also can output the description information that satisfies the user at the same time.

[0054] Alternatively or additionally, in some embodiments, the terminal device 110 can also extract second guidance information in the prompt word input from the data related to the playing scene in the reference information, the second guidance information including a playing scene requirement for the recommended media item. The terminal device 110 can also extract third guidance information in the prompt word input from the data related to the playing device in the reference information, the third guidance information including a playing device requirement for the recommended media item. The terminal device 110 can also extract fourth guidance information in the prompt word input from the data related to the playing time in the reference information, the fourth guidance information including a playing time requirement for the recommended media item. The terminal device 110 can also extract fifth guidance information in the prompt word input from the data related to the creation background in the reference information, the fifth guidance information including a generation requirement for the description information of the media item. These guidance information can further assist the machine learning model 120 to determine the media item that meets the specific playing scene, specific playing device and / or playing time. In this way, through the guidance information in the prompt word input, the machine learning model 120 can determine the media item that matches the specific context for recommendation to the user under the conditions of different times, scenes and / or devices.

[0055] In some embodiments, the prompt word input includes a system prompt word and a content generation prompt word for the media item type. Accordingly, the prompt word template can include the system prompt word, and also include the configuration rule for the content generation prompt word. The system prompt word may, for example, guide the use of the reference information by the machine learning model 120 and the generation requirement of the model output. FIG. 3A shows an example 300A of part of the content corresponding to the system prompt word according to some embodiments of the present disclosure. Based on the system prompt word, the machine learning model 120 can be guided to know how to utilize the subsequent information to perform the model processing.

[0056] The content generation prompt word is mainly generated based on the reference information of the target user. Accordingly, the configuration rule for the content generation prompt word can include the guidance information extracted from the reference information. Specifically, the configuration rule for the content generation prompt word may, for example, include the first type of guidance information for the machine learning model extracted from the behavior data in the reference information. For example, if the behavior data is associated with a music type media item, the terminal device 110 can extract the first guidance information for the machine learning model based on the total content style, historical playback behavior, etc. information of the music type media item indicated by the behavior data as at least part of the prompt word input, wherein the first guidance information includes the recommendation requirement for the media item and / or the generation requirement for the description information of the media item.

[0057] The configuration rule for the content generation prompt word may, for example, also include the second type of guidance information for the machine learning model extracted from the data related to the playback scene in the reference information. For example, if the media item type is a music type, the terminal device 110 can extract the second guidance information for the machine learning model associated with the playback scene of the target user playing music from the data related to the playback scene as part of the prompt word input.

[0058] The configuration rule for the content generation prompt word may, for example, also include the third type of guidance information for the machine learning model extracted from the data related to the playback device in the reference information. For example, if the media item type is a music type, the terminal device 110 can extract the third guidance information for the machine learning model associated with the device of the target user playing music from the data related to the playback device as part of the prompt word input.

[0059] The configuration rule for the content generation prompt word may, for example, also include the fourth type of guidance information for the machine learning model extracted from the data related to the playback time in the reference information. For example, if the media item type is a music type, the terminal device 110 can extract the fourth guidance information for the machine learning model associated with the time of the target user playing music from the data related to the playback time as part of the prompt word input.

[0060] The configuration rule for the content generation prompt word may, for example, also include the fifth type of guidance information for the machine learning model extracted from the data related to the creation background in the reference information. For example, if the media item type is a music type, the terminal device 110 can extract the fifth guidance information for the machine learning model associated with the creation background of the music from the data related to the creation background as part of the prompt word input.

[0061] The content generation prompt words can be used to guide the machine learning model 120 to provide the desired model output based on the reference information, for example. FIG. 3B shows an example 300B of content generation prompt words for content generation (i.e., media item recommendation and description information generation of media items) according to some embodiments of the present disclosure. It should be understood that the examples given in FIG. 3A and FIG. 3B are merely examples of description information, and do not imply any limitation on the scope of embodiments of the present disclosure. Any appropriate prompt word input for a language model can be designed as needed in actual applications.

[0062] As to the specific way of obtaining the model output, if the machine learning model 120 is deployed locally at the terminal device 110, the terminal device 110 can directly provide the prompt word input to the machine learning model 120, and obtain the model output for the prompt word input from the machine learning model 120. If the machine learning model 120 is deployed at other electronic devices (e.g., the server 130), the terminal device 110 can send the prompt word input to the server 130 via a communication connection between the terminal device 110 and the other electronic devices (e.g., the server 130). Alternatively or additionally, in some embodiments, the terminal device 110 can also directly send the reference information to the server 130, and the server 130 generates the prompt word input for the machine learning model 120 based on the reference information. The server 130 provides the prompt word input to the machine learning model 120, and obtains the model output for the prompt word input from the machine learning model 120. The terminal device 110 in turn obtains the model output of the machine learning model 120 from the server 130 for output to the user.

[0063] Taking the music type as an example of the media item type, the set of media items of the media item type indicated by the model output can be a set of songs recommended to the user, and the description information associated with the set of media items can include description information associated with the set of songs. The description information is intended to help the user understand the played songs and achieve a seamless effect between the switching of different songs.

[0064] In some embodiments, the description information can include a lead information (also referred to as an intro) provided before the playing of the set of media items. The lead information may, for example, be used to introduce the set of media items, to hint the target user that the set of media items is about to be provided to the target user, etc. In some embodiments, alternatively or additionally, the description information can include a tie-in information (also referred to as a tie-in) provided between the playing intervals of the individual media items in the set of media items, and / or an introduction information for the individual media items in the set of media items. The introduction information can be used to introduce the relevant content of the corresponding media item (e.g., the background of the creation, the creator, the relevant story, etc.), the tie-in information can be used to summarize the previous media item and lead to the next media item, and the tie-in information and the introduction information can also be collectively referred to as the transition information associated with the set of media items. In some embodiments, alternatively or additionally, the description information can include a wrap-up information provided after the playing of the set of media items. The wrap-up information can be used to hint the user that the provision of the set of media items is complete, etc.

[0065] Referring to FIG. 4, FIG. 4 shows a schematic diagram of an example 400 of the description information, according to some embodiments of the present disclosure. As shown in FIG. 4, the description information can include a lead information 401, transition information 402 (including transition information 402-1, 402-2, 402-3, etc.), and a wrap-up information 404. It should be understood, of course, that the example given in FIG. 4 is merely an example of the description information, and does not imply any limitation on the scope of embodiments of the present disclosure.

[0066] At block 230, the terminal device 110 provides the set of media items and the description information to the target user. The set of media items can have an order. The terminal device 110 can play the set of media items for the user based on the order of the set of media items, and output the description information for the set of media items during the playing of the set of media items. The order of the set of media items here can be default, or determined by the machine learning model 120.

[0067] Taking the set of media items as a set of songs for example, the terminal device 110 may, for example, play the set of songs in the order of the playing of the set of songs, and play the description information for the set of songs during the playing of the set of songs. With continued reference to FIG. 4, the terminal device 110 may, for example, play the lead information 401 for the set of songs before playing the set of songs 403 (including songs 403-1, 403-2, 403-3, etc.), play the corresponding transition information 402 between every two songs during the playing of the set of songs, and play the wrap-up information 404 after the playing of the set of songs. In some embodiments, the description information can be output in a textual manner, or output to the user in a voice manner through a text-to-speech conversion technique.

[0068] In some embodiments, the terminal device 110 can further obtain feedback information of the target user for the model output. The feedback information may, for example, include a score, an evaluation, an interaction record (the interaction record may, for example, include a record of the target user skipping a certain media item and description information) of the target user for the model output, and the like. The feedback information can indicate a degree of satisfaction of the target user for the model output. For example, the higher the score of the target user, the higher the degree of satisfaction of the target user for the model output.

[0069] The terminal device 110 may, for example, update the prompt word input or the prompt word template corresponding to the prompt word input based on the feedback information of the target user for the model output. In some embodiments, if the feedback information does not reach a threshold value, the terminal device 110 can update the prompt word input or the prompt word template based on the feedback information. For example, if the score does not reach a threshold score, the terminal device 110 can determine that the target user is not satisfied with the model output, and the terminal device 110 can update the prompt word input or the prompt word template. The terminal device 110 may, for example, employ any appropriate manner to analyze the feedback information to determine whether the behavior data of the target user related to the media item type has changed and whether the target user has new needs. The terminal device 110 may, for example, update the prompt word input or the prompt word template based on the analysis result of the feedback information. It can be understood that the terminal device 110 can generate an updated prompt word input based on the updated prompt word template. This helps to make the updated prompt word input more in line with user preferences. The terminal device 110 may, in turn, obtain an updated model output by providing the updated prompt word input to the machine learning model 120. The terminal device 110 can determine a set of updated media items indicated by the updated model output and description information associated with the set of updated media items. The terminal device 110 may, in turn, provide the set of updated media items and the description information associated with the set of updated media items to the target user. In this way, the model output can be made more and more in line with user preferences by repeatedly obtaining feedback information and updating prompt word inputs.

[0070] In summary, according to embodiments of the present disclosure, a set of media items and description information associated with the set of media items can be provided for a user based on reference information of the user, which helps to improve the interactivity of the user browsing media items and the user experience, enhance the richness and attractiveness of the description information, and improve the flexibility and adaptability of generating the description information and the set of media items.

[0071] Embodiments of the present disclosure also provide a corresponding apparatus for implementing the above method or process. FIG. 5 shows a schematic structural block diagram of an apparatus 500 for information generation according to some embodiments of the present disclosure. The apparatus 500 can be implemented as or included in the terminal device 110. Various modules / components in the apparatus 500 can be implemented by hardware, software, firmware, or any combination thereof.

[0072] As shown in FIG. 5, the apparatus 500 includes a prompt word generation module 510 configured to generate a prompt word input for a machine learning model based at least on reference information of a target user, the reference information including at least behavior data related to a media item type obtained with authorization of the target user, the prompt word input being used to guide the machine learning model to recommend media items of the media item type to the target user. The apparatus 500 further includes an output acquisition module 520 configured to obtain a model output of the machine learning model by providing the prompt word input to the machine learning model, the model output indicating at least a set of media items of the media item type and description information associated with the set of media items. The apparatus 500 further includes an information providing module 530 configured to provide the set of media items and the description information to the target user.

[0073] In some embodiments, the reference information further includes at least one of: data related to a playing scene of the media item type, data related to a playing device of the media item type, data related to a playing time of the media item type, data related to a creation background of the media item type.

[0074] In some embodiments, the behavior data related to the media item type indicates at least one of: a total content style of media items of interest to the target user, content styles of media items of interest to the target user in different scenarios respectively, historical playing behaviors of the target user on media items, requirements of the target user on description information of media items.

[0075] In some embodiments, the prompt word generation module 510 includes a first guide information extraction module configured to extract first guide information in the prompt word input from the behavior data in the reference information, the first guide information including a recommendation requirement on media items and / or a generation requirement on description information of media items.

[0076] In some embodiments, the prompt word generation module 510 further comprises at least one of: a second guidance information extraction module configured to extract second guidance information in the prompt word input from data related to a play scene in the reference information, the second guidance information comprising a play scene requirement for the recommended media item; a third guidance information extraction module configured to extract third guidance information in the prompt word input from data related to a play device in the reference information, the third guidance information comprising a play device requirement for the recommended media item; a fourth guidance information extraction module configured to extract fourth guidance information in the prompt word input from data related to a play time in the reference information, the fourth guidance information comprising a play time requirement for the recommended media item; and a fifth guidance information extraction module configured to extract fifth guidance information in the prompt word input from data related to a creation background in the reference information, the fifth guidance information comprising a generation requirement for description information of the media system.

[0077] In some embodiments, the prompt word generation module 510 comprises: a template acquisition module configured to acquire a prompt word template for the target model; and an input generation module configured to generate the prompt word input based at least on the prompt word template and the reference information.

[0078] In some embodiments, the prompt word input comprises a system prompt word and a content generation prompt word for a media item type, and the prompt word template comprises: the system prompt word, the system prompt word guiding use of the reference information by the machine learning model and a generation requirement of the model output, and a configuration rule for the content generation prompt word, the content generation prompt word being used to guide the machine learning model to provide the model output based on the reference information.

[0079] In some embodiments, the configuration rule for the content generation prompt word comprises at least one of: first type guidance information for the machine learning model extracted from behavior data in the reference information, second type guidance information for the machine learning model extracted from data related to a play scene in the reference information, third type guidance information for the machine learning model extracted from data related to a play device in the reference information, fourth type guidance information for the machine learning model extracted from data related to a play time in the reference information, and fifth type guidance information for the machine learning model extracted from data related to a creation background in the reference information.

[0080] In some embodiments, the information providing module 530 comprises: a play module configured to play the set of media items based on an order of the set of media items; and an output module configured to output description information of the set of media items during the playing of the set of media items.

[0081] In some embodiments, the description information further comprises at least one of: lead information provided before the playing of the set of media items, interlude information provided between playing intervals of respective media items in the set of media items, introduction information of respective media items in the set of media items, and epilogue information provided after the playing of the set of media items.

[0082] In some embodiments, the reference information is determined by at least one of: information about the media item type collected from the target user, historical interaction information of at least one media item of the media item type by the target user, third-party reference information of the target user obtained from the target application platform based on authorization of the target user, or environmental context information related to the target user.

[0083] In some embodiments, the apparatus 500 further comprises a feedback information obtaining module configured to obtain feedback information of the target user for the model output, and an updating module configured to update the prompt word input or the prompt word template corresponding to the prompt word input based on the feedback information of the target user for the model output.

[0084] In some embodiments, the apparatus 500 further comprises an output updating module configured to obtain an updated model output by providing the updated prompt word input to the machine learning model.

[0085] In some embodiments, the media item type at least comprises a music type, and / or wherein the machine learning model comprises a language model.

[0086] The units and / or modules included in the apparatus 500 can be implemented in various ways, including software, hardware, firmware, or any combination thereof. In some embodiments, one or more units and / or modules can be implemented using software and / or firmware, e.g., machine-executable instructions stored on a storage medium. In addition or as an alternative, part or all of the units and / or modules in the apparatus 500 can be implemented at least partially by one or more hardware logic components. As an example and not by way of limitation, example types of hardware logic components that can be used include Field-Programmable Gate Arrays (FPGAs), Application-Specific Integrated Circuits (ASICs), Application-Specific Standard Products (ASSPs), System-on-a-Chip (SOCs), Complex Programmable Logic Devices (CPLDs), etc.

[0087] It should be understood that one or more steps in the above methods can be performed by an appropriate electronic device or combination of electronic devices. Such an electronic device or combination of electronic devices may, for example, include the terminal device 110 in FIG. 1.

[0088] FIG. 6 illustrates a block diagram of an electronic device 600 in which one or more embodiments of the disclosure can be implemented. It should be understood that the electronic device 600 illustrated in FIG. 6 is merely exemplary and should not be construed as limiting the functionality and scope of the embodiments described herein. The electronic device 600 illustrated in FIG. 6 can be used to implement the terminal device 110 of FIG. 1 or the apparatus 500 of FIG. 5.

[0089] As illustrated in FIG. 6, the electronic device 600 is in the form of a general electronic device. Components of the electronic device 600 can include, but are not limited to, one or more processors or processing units 610, a memory 620, a storage device 630, one or more communication units 640, one or more input devices 650, and one or more output devices 660. The processing unit 610 can be a real or virtual processor and capable of executing various processing according to programs stored in the memory 620. In a multi-processor system, multiple processing units execute computer-executable instructions in parallel to improve parallel processing capability of the electronic device 600.

[0090] The electronic device 600 typically includes a number of computer storage media. Such media can be any available media that is accessible by the electronic device 600 and includes both volatile and non-volatile media, removable and non-removable media. The memory 620 can be a volatile memory (e.g., registers, cache, random access memory (RAM)), a non-volatile memory (e.g., read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory), or some combination thereof. The storage device 630 can be a removable or non-removable media and can include a machine-readable medium, such as a flash drive, a magnetic disk, or any other medium that can be used to store information and / or data and that can be accessed by the electronic device 600.

[0091] The electronic device 600 can further include additional removable / non-removable, volatile / non-volatile storage media. Although not shown in FIG. 6, a disk drive and a disk drive interface can be provided for reading from or writing to a removable, non- volatile magnetic disk (e.g., a "hard disk that can be used for storing software and / or data) and an optical disk drive and an optical disk drive interface can be provided for reading from or writing to a removable, non-volatile optical disk (such as a CD-ROM or other optical medium). In these instances, each drive can be connected to the bus (not shown) by one or more data media interfaces. The memory 620 can include a computer program product 625 having one or more program modules configured to carry out the various methods or actions of the various embodiments of the present disclosure.

[0092] The communication unit 640 enables communication through communication media with other electronic devices. Additionally, the functionality of the components of the electronic device 600 can be implemented in a single computing cluster or a plurality of computer machines that are capable of communicating over a communication connection. As such, the electronic device 600 can operate in a networked environment using logical connections to one or more other servers, network personal computers (PCs), or another network node.

[0093] The input device 650 can be one or more input devices, such as a mouse, a keyboard, a trackball, etc. The output device 660 can be one or more output devices, such as a display, a speaker, a printer, etc. The electronic device 600 can also communicate with one or more external devices (not shown), such as a storage device, a display device, etc., through the communication unit 640, as necessary, with one or more devices that enable a user to interact with the electronic device 600, or with any device (e.g., a network card, a modem, etc.) that enables the electronic device 600 to communicate with one or more other electronic devices. Such communication can be carried out via an input / output (I / O) interface (not shown).

[0094] According to an example implementation of the present disclosure, there is provided a computer-readable storage medium having computer-executable instructions stored thereon, where the computer-executable instructions are executed by a processor to implement the method described above. According to an example implementation of the present disclosure, there is also provided a computer program product tangibly stored on a non-transitory computer-readable medium and including computer-executable instructions, where the computer-executable instructions are executed by a processor to implement the method described above.

[0095] Various aspects of the disclosure are now described with reference to the drawings. In general, the drawings described below are diagrammatic and schematic representations of actual or conceptual structures and processes, and are not limiting of the scope of the present disclosure. In the drawings, the same reference numerals are used to represent similar or like items.

[0096] These computer readable program instructions can be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks. These computer readable program instructions can also be stored in a computer readable storage medium that can include a non-transitory computer readable medium, thereby making a computer program product, i.e., an article of manufacture, according to the definitions and teachings of 35 U.S.C. § 101. The computer readable storage medium can include a computer readable storage medium alone or in combination with another computer readable storage medium from which the computer program instructions can be read. The computer readable storage medium can be a computer readable storage medium having a plurality of program instructions logically stored (e.g., as one or more files). Examples of computer readable storage media include random access memory (RAM), read only memory (ROM), programmable read only memory (PROM), erasable programmable read only memory (EPROM), electrically erasable programmable read only memory (EEPROM), compact disk read only memory (CD-ROM), compact disk read / write (CD-R / W), digital versatile disk (DVD), flash memory, and / or any other volatile or non-volatile computer readable storage medium.

[0097] The computer readable program instructions can 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 such that the instructions which execute on the computer, other programmable data processing apparatus, or other device implement the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0098] The flow diagrams and the block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various implementations of the present disclosure. In this regard, each block in the flow diagrams and the block diagrams can represent a module, segment, or portion of instructions, which comprises one or more executable instructions for implementing the specified logic functions (s). In some alternative implementations, the functions noted in the blocks can occur out of the order noted in the figures. For example, two blocks shown in succession may, in some cases, be executed substantially concurrently, or the blocks can sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and / or flowchart illustrations, and combinations thereof, can be implemented by a dedicated hardware-based system that performs the specified functions or acts, or combinations of hardware and software.

[0099] implementations. Numerous modifications and adaptations will be apparent to those skilled in the art without departing from the scope and spirit of the described implementations. The herein disclosed subject matter is to be considered merely illustrative in nature and its details are therefore not to be taken in a limiting sense. The terminology used in the detailed description of the various implementations presented is to be taken only in a descriptive, and not a limiting, sense.

Claims

1. A method for information generation, comprising: generating, based at least on reference information of a target user, a prompt input for a machine learning model, the reference information comprising at least behavior data related to a media item type, the prompt input being used to guide the machine learning model to recommend media items of the media item type to the target user; obtaining a model output of the machine learning model by providing the prompt input to the machine learning model, the model output indicating at least a set of media items of the media item type and description information associated with the set of media items; and providing the set of media items and the description information to the target user.

2. The method of claim 1, wherein the reference information further comprises at least one of: data related to a play scenario of the media item type, data related to a play device of the media item type, data related to a play time of the media item type, data related to a creation background of the media item type.

3. The method of claim 1, wherein the behavior data related to the media item type indicates at least one of: a total content style of media items that the target user is interested in, content styles of media items that the target user is interested in respectively in different scenarios, a historical play behavior of the target user on media items, a requirement of the target user on description information of media items.

4. The method of claim 1, wherein generating the prompt input comprises: extracting first guide information in the prompt input from the behavior data in the reference information, the first guide information comprising a recommendation requirement on media items and / or a generation requirement on description information of media items.

5. The method of claim 2, wherein generating the prompt input further comprises at least one of: extracting second guide information in the prompt input from the data related to the play scenario in the reference information, the second guide information comprising a play scenario requirement on recommended media items; extracting third guide information in the prompt input from the data related to the play device in the reference information, the third guide information comprising a play device requirement on the recommended media items; extracting fourth guide information in the prompt input from the data related to the play time in the reference information, the fourth guide information comprising a play time requirement on the recommended media items; extracting fifth guide information in the prompt input from the data related to the creation background in the reference information, the fifth guide information comprising a generation requirement on description information of media items.

6. The method of claim 1, wherein generating the prompt input comprises: obtaining a prompt template for the target model; and generating the prompt input based at least on the prompt template and the reference information.

7. The method of claim 6, wherein the prompt input comprises a system prompt and a content generation prompt for the media item type, and wherein the prompt template comprises: ​ the system prompt word guiding the use of the reference information by the machine learning model and the generation requirement of the model output, and a configuration rule of the content generation prompt word for guiding the machine learning model to provide the model output based on the reference information.

8. The method of claim 7, wherein the configuration rule of the content generation prompt word comprises at least one of: extracting first type of guidance information to the machine learning model from the behavior data in the reference information, extracting second type of guidance information to the machine learning model from the data related to the play scene in the reference information, extracting third type of guidance information to the machine learning model from the data related to the play device in the reference information, extracting fourth type of guidance information to the machine learning model from the data related to the play time in the reference information, extracting fifth type of guidance information to the machine learning model from the data related to the creation background in the reference information.

9. The method of claim 1, wherein providing the set of media items and the description information to the target user comprises: playing the set of media items based on the order of the set of media items; and outputting the description information of the set of media items during the playing of the set of media items.

10. The method of claim 1, wherein the description information further comprises at least one of: preamble information provided before the playing of the set of media items, string information provided between the play interval of each media item in the set of media items, introduction information of each media item in the set of media items, conclusion information provided after the playing of the set of media items.

11. The method of claim 1, wherein the reference information is determined by at least one of: information about the media item type collected from the target user, historical interaction information of at least one media item of the media item type by the target user, third-party reference information of the target user obtained from a target application platform based on the authorization of the target user, or environmental context information related to the target user.

12. The method of claim 1, further comprising: obtaining feedback information of the target user for the model output; and updating the prompt word input or the prompt word template corresponding to the prompt word input based on the feedback information of the target user for the model output.

13. The method of claim 12, further comprising: obtaining an updated model output by providing the updated prompt word input to the machine learning model.

14. The method of claim 1, wherein the media item type at least comprises a music type, and / or wherein the machine learning model comprises a language model.

15. An apparatus for information generation, comprising: ​ ​ ​ The prompt word generation module is configured to generate a prompt word input for a machine learning model based on reference information of a target user, the reference information comprising at least behavior data related to a media item type obtained with authorization of the target user, the prompt word input being used to guide the machine learning model to recommend media items of the media item type to the target user; and The output obtaining module is configured to obtain a model output of the machine learning model by providing the prompt word input to the machine learning model, the model output indicating at least a set of media items of the media item type and description information associated with the set of media items; and The information providing module is configured to provide the set of media items and the description information to the target user.

16. An electronic device, comprising: at least one processing unit; and at least one memory that is coupled to the at least one processing unit and stores instructions for execution by the at least one processing unit, the instructions, when executed by the at least one processing unit, cause the electronic device to perform the method according to any one of claims 1 to 14.

17. A computer readable storage medium having stored thereon a computer program, the computer program being executable by a processor to implement the method according to any one of claims 1 to 14.

18. A computer program product comprising a computer program, wherein the computer program, when executed by a processor, implements the method according to any one of claims 1 to 14.

Images