Video recommendation method and device, electronic equipment and storage medium

Abstract

The invention relates to a video recommendation method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining a test question text and at least one candidate test question text, and enabling the at least one candidate test question text to have a corresponding candidate answer video; for any candidate test question text, determining the similarity between the candidate test question text and the test question text; according to the similarity, screening a target test question text from the at least one candidate test question text; and recommending a target answer video corresponding to the target test question text. In this way, the corresponding answer video is recommended based on the similarity between the test question texts, generation in the modes of manual recording, editing, editing and the like can be avoided, time is saved, and efficiency is improved.

Description

Technical Field

[0001] The present application relates to the field of Internet technology, and in particular to a video recommendation method, device, electronic device, and storage medium. Background Art

[0002] With the popularization of Internet technology, online education has gradually become a trend. Online education breaks the time and space limitations of traditional education, allowing learners to receive education anytime and anywhere, and can meet learners' needs for immersive and interactive learning.

[0003] In the related art, online education usually requires answer videos of test questions, which are generally generated by manual recording, editing, and editing, which is time-consuming and inefficient. Summary of the invention

[0004] In order to solve the technical problem that the answer videos for the above-mentioned test analysis are generally generated by manual recording, editing, etc., which takes a lot of time and is inefficient, this application provides a video recommendation method, device, electronic device, and storage medium. The specific technical solution is as follows:

[0005] In a first aspect, the present application provides a video recommendation method, the method comprising:

[0006] Obtaining a test question text and at least one candidate test question text, wherein at least one of the candidate test question texts has a corresponding candidate answer video;

[0007] For any of the candidate test question texts, determining the similarity between the candidate test question text and the test question text;

[0008] According to the similarity, selecting a target test question text from at least one of the candidate test question texts;

[0009] Recommend a target answer video corresponding to the target test question text.

[0010] In an optional implementation, the determining the similarity between the candidate test question text and the test question text includes:

[0011] Preprocessing the candidate test question text, and converting the preprocessed candidate test question text into a candidate test question vector;

[0012] Preprocessing the test question text, and converting the preprocessed test question text into a test question vector;

[0013] The candidate test question vector and the test question vector are input into a pre-trained deep learning model to obtain the similarity between the candidate test question text and the test question text.

[0014] In an optional implementation, the selecting a target test question text from at least one of the candidate test question texts according to the similarity comprises:

[0015] Selecting a maximum similarity from the similarities, and determining whether the maximum similarity is greater than or equal to a first preset similarity threshold;

[0016] When the maximum similarity is greater than or equal to the first preset similarity threshold, the candidate test question text corresponding to the maximum similarity is determined as the target test question text.

[0017] In an optional embodiment, the method further comprises:

[0018] When the maximum similarity is less than the first preset similarity threshold, determining whether the maximum similarity is greater than or equal to a preset second similarity threshold;

[0019] Wherein, the preset first similarity threshold is greater than the preset second similarity threshold;

[0020] When the maximum similarity is greater than or equal to the second preset similarity threshold, the candidate test question text corresponding to the maximum similarity is determined as the target test question text.

[0021] In an optional implementation, the recommending a target answer video corresponding to the target test question text includes:

[0022] Obtaining a target answering video corresponding to the target test question text, and determining the difference between the test question text and the target test question text;

[0023] According to the difference, the target answer video is adjusted, and the adjusted target answer video is recommended.

[0024] In an optional embodiment, before executing the method, the method further includes:

[0025] Obtain at least one pair of matching existing test question text and video text, and perform the following processing on any matching existing test question text and video text:

[0026] Establishing a correspondence between the existing test question text and the existing answer video corresponding to the video text;

[0027] Preprocessing the existing test question text, and converting the preprocessed existing test question text into an existing test question vector;

[0028] Preprocessing the video text, and converting the preprocessed video text into a video text vector;

[0029] Inputting the existing test question vector and the video text vector into a deep learning model to obtain a predicted similarity;

[0030] Determine a loss value between the predicted similarity and a matching label between the matched existing test question text and the video text;

[0031] The deep learning model is trained according to the loss value, and when the loss value converges, the training is stopped to obtain a pre-trained deep learning model.

[0032] In an optional implementation, the obtaining of at least one pair of matching existing test question text and video text includes:

[0033] Obtain at least one existing test question text and at least one existing answer video;

[0034] For any of the existing answer videos, determining whether there is a corresponding subtitle text for the existing answer video;

[0035] In the case where the subtitle text exists, decomposing the existing answer video into a frame sequence;

[0036] Extracting the subtitle text corresponding to the existing answering video from the frame sequence, and determining the subtitle text as the video text corresponding to the existing answering video;

[0037] In the absence of the subtitle text, performing speech recognition on the audio corresponding to the existing answer video to obtain a speech recognition text;

[0038] Determine the speech recognition text as the video text corresponding to the existing answer video;

[0039] At least one pair of matching existing test question texts and video texts is determined from at least one of the existing test question texts and at least one of the video texts.

[0040] In a second aspect, the present application provides a video recommendation device, the device comprising:

[0041] A text acquisition module, used to acquire a test question text and at least one candidate test question text, wherein at least one of the candidate test question texts has a corresponding candidate answer video;

[0042] A similarity determination module, for determining the similarity between the candidate test question text and the test question text for any of the candidate test question texts;

[0043] A text screening module, used for screening a target test question text from at least one of the candidate test question texts according to the similarity;

[0044] The video recommendation module is used to recommend the target answer video corresponding to the target test question text.

[0045] In a third aspect, an electronic device is provided, comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus;

[0046] Memory, used to store computer programs;

[0047] The processor is used to implement any video recommendation method described in the first aspect when executing the program stored in the memory.

[0048] In a fourth aspect, a storage medium is further provided, wherein instructions are stored in the storage medium, and when the instructions are executed on a computer, the computer executes any of the video recommendation methods described in the first aspect.

[0049] In a fifth aspect, a computer program product comprising instructions is also provided, which, when executed on a computer, enables the computer to execute any of the above-mentioned video recommendation methods.

[0050] The above-mentioned technical scheme provided by the embodiment of the present application has the following advantages compared with the prior art: the video recommendation method provided by the embodiment of the present application obtains a test question text and at least one candidate test question text, wherein at least one candidate test question text each has a corresponding candidate answer video, and for any candidate test question text, determines the similarity between the candidate test question text and the test question text, and based on the similarity, selects a target test question text from at least one candidate test question text, and recommends a target answer video corresponding to the target test question text.

[0051] By determining the similarity between the test question text and at least one candidate test question text, a target test question text is screened from at least one candidate test question text based on the similarity, and a target answer video corresponding to the target test question text is recommended. In this way, based on the similarity between the test question texts, the corresponding answer video is recommended, which can avoid the generation through manual recording, clipping, editing, etc., saving time and improving efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0053] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0054] One or more embodiments are exemplarily described by pictures in the corresponding drawings, and these exemplified descriptions do not constitute limitations on the embodiments. Elements with the same reference numerals in the drawings represent similar elements, and unless otherwise stated, the figures in the drawings do not constitute proportional limitations.

[0055] Figure 1 A schematic diagram of an implementation flow of a video recommendation method provided in an embodiment of the present application;

[0056] Figure 2 A schematic diagram of an implementation flow of another video recommendation method provided in an embodiment of the present application;

[0057] Figure 3 A schematic diagram of an implementation flow of a model training method provided in an embodiment of the present application;

[0058] Figure 4 A schematic diagram of the structure of a video recommendation device provided in an embodiment of the present application;

[0059] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0061] The disclosure below provides many different embodiments or examples to realize the different structures of the present application. In order to simplify the disclosure of the present application, the parts and settings of specific examples are described below. Of course, they are only examples, and the purpose is not to limit the present application. In addition, the present application can repeat reference numbers and / or letters in different examples. This repetition is for the purpose of simplification and clarity, and does not itself indicate the relationship between the various embodiments and / or settings discussed.

[0062] like Figure 1FIG. 1 is a schematic diagram of an implementation flow of a video recommendation method provided in an embodiment of the present application. The method may specifically include the following steps:

[0063] S101, obtaining a test question text and at least one candidate test question text, wherein each of the at least one candidate test question text has a corresponding candidate answer video.

[0064] In an embodiment of the present application, a test question text is obtained. For example, the test question text may be about an animal keeper distributing nuts to monkeys and squirrels. Each monkey receives 4 nuts and each squirrel receives 2 nuts. The animal keeper distributes a total of 100 nuts. There are 35 squirrels and monkeys in total, and the question is how many squirrels and monkeys there are respectively.

[0065] The test text may be input by the user, thereby obtaining the test text input by the user. Alternatively, the test document may be scanned to obtain the test text in the test document, which is not limited in the present embodiment.

[0066] In addition to obtaining the test text as described above, it is also necessary to obtain at least one candidate test text, which refers to an existing test text. For each candidate test text, there is a corresponding candidate answer video, which means that for each existing test text, there is a corresponding existing answer video.

[0067] It should be noted that the candidate answer video corresponding to the candidate test question text is a video of answering the candidate test question text, which can be generated by manual recording, clipping, editing, etc., or by AI, and the number of candidate answer videos can be multiple, which is not limited to the embodiments of the present application.

[0068] S102: for any candidate test question text, determine the similarity between the candidate test question text and the test question text.

[0069] S103, screening a target test question text from at least one candidate test question text according to the similarity.

[0070] In an embodiment of the present application, for at least one candidate test question text, for any candidate test question text, the similarity between the candidate test question text and the above test question text is determined, so that there is a corresponding similarity between the above test question text and each candidate test question text.

[0071] For at least one candidate test question text, a target test question text can be selected from the at least one candidate test question text according to the similarity between the test question text and each candidate test question text. It can be understood that the similarity between the target test question text and the test question text is the greatest.

[0072] It should be noted that, for the similarity between the above-mentioned candidate test texts and the test text, the cosine distance or the Euclidean distance can be calculated to determine the cosine distance or the Euclidean distance as the similarity between the above-mentioned candidate test texts and the test text, and the embodiments of the present application are not limited to this.

[0073] S104, recommending a target answer video corresponding to the target test question text.

[0074] In an embodiment of the present application, for at least one candidate test question text, after the above screening, a target test question text can be obtained, and the target test question text has a corresponding target answer video, thereby the target answer video corresponding to the target test question text can be recommended.

[0075] In this way, based on the similarity between the test texts, the corresponding answer videos are recommended, which can avoid the generation through manual recording, clipping, editing, etc., saving time and improving efficiency.

[0076] It should be noted that, for the target answer video corresponding to the target test question text, the number of the target answer videos can be multiple, which means that multiple target answer videos corresponding to the target test question text can be recommended, or a certain target answer video corresponding to the target test question text can be recommended based on the score, click-through rate / click volume, etc. The embodiments of the present application are not limited to this.

[0077] Through the above description of the technical solution provided in the embodiment of the present application, a test question text and at least one candidate test question text are obtained, wherein at least one candidate test question text each has a corresponding candidate answer video, and for any candidate test question text, the similarity between the candidate test question text and the test question text is determined, and based on the similarity, a target test question text is screened from at least one candidate test question text, and a target answer video corresponding to the target test question text is recommended.

[0078] By determining the similarity between the test question text and at least one candidate test question text, a target test question text is screened from at least one candidate test question text based on the similarity, and a target answer video corresponding to the target test question text is recommended. In this way, based on the similarity between the test question texts, the corresponding answer video is recommended, which can avoid the generation through manual recording, clipping, editing, etc., saving time and improving efficiency.

[0079] like Figure 2 FIG. 2 is a schematic diagram of an implementation flow of another video recommendation method provided in an embodiment of the present application. The method may specifically include the following steps:

[0080] S201, obtaining a test question text and at least one candidate test question text, wherein each of the at least one candidate test question texts has a corresponding candidate answer video.

[0081] In the embodiment of the present application, this step is similar to the above-mentioned step S101, and the embodiment of the present application will not be described one by one here.

[0082] S202, for any candidate test question text, preprocess the candidate test question text, and convert the preprocessed candidate test question text into a candidate test question vector.

[0083] S203, preprocessing the test question text, and converting the preprocessed test question text into a test question vector.

[0084] S204, inputting the candidate test question vector and the test question vector into a pre-trained deep learning model to obtain the similarity between the candidate test question text and the test question text.

[0085] In an embodiment of the present application, for at least one candidate test question text, for any candidate test question text, the candidate test question text is preprocessed to obtain a preprocessed candidate test question text, and the preprocessed candidate test question text is converted into a candidate test question vector.

[0086] Similarly, for the above-mentioned test text, the test text is preprocessed to obtain a preprocessed test text, and the preprocessed test text is converted into a test vector. The preprocessed candidate test text and the preprocessed test text can be converted into corresponding candidate test vectors and test vectors through a text vectorization model, both of which are numerical vectors, and the embodiments of the present application do not limit this.

[0087] After the above steps, the candidate question vector and the question vector can be obtained, and the candidate question vector and the question vector can be input into the pre-trained deep learning model to obtain the similarity between the candidate question text and the question text. Among them, the pre-trained deep learning model can be any type of deep learning model currently on the market, which can be used to predict similarity, and the embodiments of the present application are not limited to this.

[0088] It should be noted that preprocessing refers to word segmentation, stop word removal, stem extraction or word type restoration, etc., so as to convert the preprocessed candidate test question text into a candidate test question vector and convert the preprocessed test question text into a test question vector. The embodiments of the present application do not limit this.

[0089] In addition, the above pre-trained deep learning model needs to be trained before it can be obtained. Figure 3 FIG. 1 is a schematic diagram of an implementation flow of a model training method provided in an embodiment of the present application, and the method may specifically include the following steps:

[0090] S301, obtaining at least one pair of matching existing test question text and video text, and performing the following processing for any matching existing test question text and video text.

[0091] In the embodiment of the present application, at least one pair of matching existing test question text and video text is obtained, where the video text refers to the text corresponding to the existing answer video. Thus, for any matching existing test question text and video text, the following steps S302 to S307 are performed.

[0092] Among them, for at least one pair of matching existing test question text and video text, a large number of existing test question texts and existing answer videos on the market can be collected to obtain at least one pair of matching existing test question text and video text.

[0093] Specifically, at least one existing test question text and at least one existing answer video are obtained. For any existing answer video, it is determined whether there is a corresponding subtitle text for the existing answer video. If there is a subtitle text for the existing answer video, the existing answer video is decomposed into a frame sequence, that is, decomposed into frames of images to form a frame sequence, and the subtitle text corresponding to the existing answer video is extracted from the frame sequence, and the subtitle text is determined as the video text corresponding to the existing answer video.

[0094] For an existing answer video, if the existing answer video does not have a subtitle text, then the audio corresponding to the existing answer video can be subjected to speech recognition to obtain a speech recognition text, and the speech recognition text is determined as the video text corresponding to the existing answer video. After the above steps, the corresponding video text is determined for each existing answer video.

[0095] Based on this, at least one video text and at least one existing test question text can be obtained, and at least one pair of matching existing test question text and video text can be determined from the at least one existing test question text and the at least one video text.

[0096] Among them, NLP technology can be used to deeply understand the meaning of existing test texts, and NLP technology can be used to deeply understand the meaning of video texts, and the meaning of existing test texts can be matched with the meaning of video texts. If they match, then they can be determined as a pair of matching existing test texts and video texts. This embodiment of the present application is not limited to this.

[0097] S302, establishing a correspondence between the existing test question text and the existing answer video corresponding to the video text.

[0098] In an embodiment of the present application, for matching existing test texts and video texts, a correspondence can be established between the existing test texts and the existing answer videos corresponding to the video texts. In this way, when a test text similar to the existing test text (i.e., the above-mentioned candidate test text) is encountered subsequently, the existing answer video (i.e., the above-mentioned candidate answer video) corresponding to the existing test text can be recommended.

[0099] S303, preprocessing the existing test question text, and converting the preprocessed existing test question text into an existing test question vector.

[0100] In an embodiment of the present application, the existing test question text can be preprocessed to obtain the preprocessed existing test question text, where the so-called preprocessing refers to word segmentation, stop word removal, stem extraction or word form restoration, etc., and then the preprocessed existing test question text can be converted into an existing test question vector, which is a numerical vector.

[0101] S304, preprocessing the video text, and converting the preprocessed video text into a video text vector.

[0102] In an embodiment of the present application, the video text can be preprocessed to obtain preprocessed video text. The so-called preprocessing refers to word segmentation, stop word removal, stem extraction or word form restoration, etc., and the preprocessed video text can be converted into a video text vector, which is a numerical vector.

[0103] S305, input the existing test question vector and the video text vector into the deep learning model to obtain the predicted similarity.

[0104] In an embodiment of the present application, for the existing test question vector and video text vector obtained above, the existing test question vector and video text vector can be input into a deep learning model to obtain the predicted similarity, that is, the predicted similarity between the existing test question text and the video text.

[0105] S306, determining the loss value between the predicted similarity and the matching labels between the matched existing test question text and the video text.

[0106] S307, training the deep learning model according to the loss value, and stopping the training to obtain a pre-trained deep learning model when the loss value converges.

[0107] In an embodiment of the present application, for the matching existing test question text and video text, there are corresponding matching tags, for example, the matching tag can be 1, so that the predicted similarity and the loss value between the matching matching tags between the existing test question text and the video text can be determined.

[0108] For the loss value, the deep learning model can be trained according to the loss value, and when the loss value converges, the training is stopped to obtain the pre-trained deep learning model. The convergence of the loss value can be, for example, that the loss value is less than a certain threshold, which is not limited in the embodiments of the present application.

[0109] After the above model training, the final pre-trained deep learning model can be obtained. For at least one candidate test question text, for any candidate test question text, the candidate test question text is pre-processed to obtain the pre-processed candidate test question text, and the pre-processed candidate test question text is converted into a candidate test question vector. Similarly, for the above test question text, the test question text is pre-processed to obtain the pre-processed test question text, and the pre-processed test question text is converted into a test question vector. The candidate test question vector and the test question vector can be input into the pre-trained deep learning model to obtain the similarity between the candidate test question text and the test question text.

[0110] S205: Select a maximum similarity from the similarities, and determine whether the maximum similarity is greater than or equal to a first preset similarity threshold.

[0111] S206: When the maximum similarity is greater than or equal to the first preset similarity threshold, the candidate test question text corresponding to the maximum similarity is determined as the target test question text.

[0112] S207, recommending a target answer video corresponding to the target test question text.

[0113] In an embodiment of the present application, for any candidate test text, the similarity between it and the test text can be determined, thereby selecting the maximum similarity from these similarities and judging whether the maximum similarity is greater than or equal to a first preset similarity threshold, for example, the first preset similarity threshold is 100%.

[0114] When the maximum similarity is greater than or equal to the first preset similarity threshold, it means that the candidate test question text corresponding to the maximum similarity matches the test question text. At this time, the candidate test question text corresponding to the maximum similarity can be determined as the target test question text, and the target answer video corresponding to the target test question text is recommended.

[0115] In addition, when the maximum similarity is less than the first preset similarity threshold, it means that the candidate test text corresponding to the maximum similarity does not match the test text, and at this time, it can be determined whether the maximum similarity is greater than or equal to the preset second similarity threshold. The preset first similarity threshold is greater than the preset second similarity threshold, for example, the preset second similarity threshold is 95%.

[0116] When the maximum similarity is less than the preset first similarity threshold and greater than or equal to the second preset similarity threshold, it means that the candidate test question text corresponding to the maximum similarity is similar to the test question text. At this time, the candidate test question text corresponding to the maximum similarity can be determined as the target test question text. The target answer video corresponding to the target test question text needs to be adjusted before it can be recommended.

[0117] Specifically, the target answer video corresponding to the target test question text is obtained, and the difference between the test question text and the target test question text is determined. According to the difference, the target answer video is adjusted, and the adjusted target answer video is recommended. The difference is generally reflected in the difference of keywords, and the target answer video can be adjusted by replacing keywords, as shown below.

[0118] For example, the above test text is about an animal keeper distributing nuts to monkeys and squirrels. Each monkey gets 4 nuts and each squirrel gets 2 nuts. The animal keeper distributes a total of 100 nuts. There are 35 squirrels and monkeys in total. The question is how many squirrels and monkeys there are respectively.

[0119] After the above processing, similar to the target test text (chickens and rabbits in the same cage problem), for the target answer video corresponding to the target test text (chickens and rabbits in the same cage problem), the chickens and rabbits can be modified into squirrels and monkeys based on the difference between the test text and the target test text. Then, 4 nuts correspond to 4 feet of the rabbit, 2 nuts correspond to 2 feet of the chicken, 100 nuts correspond to 100 feet, and 35 squirrels and monkeys correspond to the total number of chickens and rabbits. Then modify the corresponding content in the target answer video, and the adjusted target answer video is recommended.

[0120] For the above-mentioned recommended target answering videos, or the recommended adjusted target answering videos, user feedback can be obtained, that is, whether the recommended target answering videos, or the recommended adjusted target answering videos are helpful to the users. For the target answering videos or the adjusted target answering videos that are helpful to the users, the test question text and the target test question text can be used as samples to optimize the pre-trained deep learning model, and the target answering videos or the adjusted target answering videos that are not helpful to the users do not need to be processed.

[0121] Corresponding to the above method embodiment, the present application embodiment also provides a video recommendation device, such as Figure 4 As shown, the device may include: a text acquisition module 410 , a similarity determination module 420 , a text screening module 430 , and a video recommendation module 440 .

[0122] The text acquisition module 410 is used to acquire the test question text and at least one candidate test question text, wherein at least one of the candidate test question texts has a corresponding candidate answer video;

[0123] A similarity determination module 420 is used to determine the similarity between the candidate test question text and the test question text for any of the candidate test question texts;

[0124] A text screening module 430, configured to screen a target test question text from at least one of the candidate test question texts according to the similarity;

[0125] The video recommendation module 440 is used to recommend the target answer video corresponding to the target test question text.

[0126] The present application also provides an electronic device, such as Figure 5 As shown, it includes a processor 51, a communication interface 52, a memory 53 and a communication bus 54, wherein the processor 51, the communication interface 52, and the memory 53 communicate with each other through the communication bus 54.

[0127] A memory 53, for storing computer programs;

[0128] The processor 51 is used to execute the program stored in the memory 53 to implement the following steps:

[0129] Obtain a test question text and at least one candidate test question text, wherein at least one of the candidate test question texts has a corresponding candidate answer video; for any of the candidate test question texts, determine the similarity between the candidate test question text and the test question text; based on the similarity, filter a target test question text from at least one of the candidate test question texts; and recommend a target answer video corresponding to the target test question text.

[0130] The communication bus mentioned in the above electronic device can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0131] The communication interface is used for communication between the above air conditioner and other devices.

[0132] The memory may include a random access memory (RAM) or a non-volatile memory, such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0133] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0134] In another embodiment provided by the present application, a storage medium is further provided. The storage medium stores instructions. When the storage medium is executed on a computer, the computer executes the video recommendation method described in any one of the above embodiments.

[0135] In another embodiment provided by the present application, a computer program product including instructions is also provided, and when the computer is run on the computer, the computer is enabled to execute the video recommendation method described in any one of the above embodiments.

[0136] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a storage medium, or transmitted from one storage medium to another storage medium, for example, the computer instructions may be transmitted from a website site, a computer, a server or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center. The storage medium may be any available medium that a computer can access or a data storage device such as a server or a data center that includes one or more available media integrations. The available medium may be a magnetic medium, (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive Solid State Disk (SSD)), etc.

[0137] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0138] Each embodiment in this specification is described in a related manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0139] The above description is only a preferred embodiment of the present application and is not intended to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application are included in the protection scope of the present application.

Claims

1. A video recommendation method, characterized in that: The method comprises: Obtaining a test question text and at least one candidate test question text, wherein at least one of the candidate test question texts has a corresponding candidate answer video; For any of the candidate test question texts, determining the similarity between the candidate test question text and the test question text; According to the similarity, selecting a target test question text from at least one of the candidate test question texts; Recommend a target answer video corresponding to the target test question text.

2. The method according to claim 1, characterized in that Determining the similarity between the candidate test question text and the test question text includes: Preprocessing the candidate test question text, and converting the preprocessed candidate test question text into a candidate test question vector; Preprocessing the test question text, and converting the preprocessed test question text into a test question vector; The candidate test question vector and the test question vector are input into a pre-trained deep learning model to obtain the similarity between the candidate test question text and the test question text.

3. The method according to claim 1, characterized in that The step of selecting a target test question text from at least one of the candidate test question texts according to the similarity comprises: Selecting a maximum similarity from the similarities, and determining whether the maximum similarity is greater than or equal to a first preset similarity threshold; When the maximum similarity is greater than or equal to the first preset similarity threshold, the candidate test question text corresponding to the maximum similarity is determined as the target test question text.

4. The method according to claim 3, characterized in that The method further comprises: When the maximum similarity is less than the first preset similarity threshold, determining whether the maximum similarity is greater than or equal to a preset second similarity threshold; Wherein, the preset first similarity threshold is greater than the preset second similarity threshold; When the maximum similarity is greater than or equal to the second preset similarity threshold, the candidate test question text corresponding to the maximum similarity is determined as the target test question text.

5. The method according to claim 4, characterized in that The target answering video corresponding to the target test question text is recommended, including: Obtaining a target answering video corresponding to the target test question text, and determining the difference between the test question text and the target test question text; According to the difference, the target answer video is adjusted, and the adjusted target answer video is recommended.

6. The method according to claim 2, characterized in that Before executing the method, the method further includes: Obtain at least one pair of matching existing test question text and video text, and perform the following processing on any matching existing test question text and video text: Establishing a correspondence between the existing test question text and the existing answer video corresponding to the video text; Preprocessing the existing test question text, and converting the preprocessed existing test question text into an existing test question vector; Preprocessing the video text, and converting the preprocessed video text into a video text vector; Inputting the existing test question vector and the video text vector into a deep learning model to obtain a predicted similarity; Determine the predicted similarity and the loss value between the matching labels of the existing test question text and the video text; The deep learning model is trained according to the loss value, and when the loss value converges, the training is stopped to obtain a pre-trained deep learning model.

7. The method according to claim 6, characterized in that The step of obtaining at least one pair of matching existing test question text and video text includes: Obtain at least one existing test question text and at least one existing answer video; For any of the existing answer videos, determining whether there is a corresponding subtitle text for the existing answer video; In the case where the subtitle text exists, decomposing the existing answer video into a frame sequence; Extracting the subtitle text corresponding to the existing answering video from the frame sequence, and determining the subtitle text as the video text corresponding to the existing answering video; In the absence of the subtitle text, performing speech recognition on the audio corresponding to the existing answer video to obtain a speech recognition text; Determine the speech recognition text as the video text corresponding to the existing answer video; At least one pair of matching existing test question texts and video texts is determined from at least one of the existing test question texts and at least one of the video texts.

8. A video recommendation device, characterized in that: The device comprises: A text acquisition module, used to acquire a test question text and at least one candidate test question text, wherein at least one of the candidate test question texts has a corresponding candidate answer video; A similarity determination module, for determining the similarity between the candidate test question text and the test question text for any of the candidate test question texts; A text screening module, used for screening a target test question text from at least one of the candidate test question texts according to the similarity; The video recommendation module is used to recommend the target answer video corresponding to the target test question text.

9. An electronic device, characterized in that: It includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus; Memory, used to store computer programs; A processor, for implementing any of the methods described in claims 1-7 when executing a program stored in a memory.

10. A storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images