CN116932936A

CN116932936A - Information processing method, information processing device, computer equipment and storage medium

Info

Publication number: CN116932936A
Application number: CN202210320195.2A
Authority: CN
Inventors: 刘静; 郑昊; 马中彪
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2022-03-29
Filing date: 2022-03-29
Publication date: 2023-10-24

Abstract

The embodiment of the application discloses an information processing method, an information processing device, computer equipment and a storage medium, wherein the method comprises the following steps: displaying a first feedback page of query information, wherein the first feedback page comprises: the information content of the feedback information corresponding to the query information in the first browsing mode; if the mode switching operation is detected, displaying a second feedback page of the query information; the second feedback page comprises: the information content of the feedback information in the second browsing mode; in the first feedback page and the second feedback page, the information content in one feedback page is the whole information content of the feedback information, and the information content in the other feedback page is the simplified information content obtained based on the whole information content. The embodiment of the application can improve the diversity of browsing modes, thereby improving the convenience of browsing information and further improving the use viscosity of the application.

Description

Information processing method, information processing device, computer equipment and storage medium

Technical Field

The present application relates to the field of internet technologies, and in particular, to the field of computer technologies, and in particular, to an information processing method, an information processing device, a computer device, and a storage medium.

Background

With the development of internet technology and computer technology, more and more Applications (APP) may provide information query services for objects (i.e., users), for example, web browsing applications may provide query services for various information such as encyclopedia terms (dictionary editing collected terms or words, and corresponding paraphrasing), work-related problems, and the like. After the query information (such as terms in the encyclopedia entry) is input to the object, the application can output and display feedback information (such as paraphrasing in the encyclopedia entry) corresponding to the query information for the object based on the information query service. At present, how to display feedback information corresponding to query information so as to improve the use viscosity of the application becomes a research hotspot.

Disclosure of Invention

The embodiment of the application provides an information processing method, an information processing device, computer equipment and a storage medium, which can improve the diversity of browsing modes, thereby improving the convenience of browsing information, enabling an object to obtain good consumption experience when the content is consumed, and further improving the use viscosity of the application.

In one aspect, an embodiment of the present application provides an information processing method, including:

displaying a first feedback page of query information, wherein the first feedback page comprises: the information content of the feedback information corresponding to the query information in the first browsing mode;

If the mode switching operation is detected, displaying a second feedback page of the query information; the second feedback page comprises: the information content of the feedback information in the second browsing mode;

in the first feedback page and the second feedback page, the information content in one feedback page is the whole information content of the feedback information, and the information content in the other feedback page is the simplified information content obtained based on the whole information content.

In another aspect, an embodiment of the present application provides an information processing apparatus, including:

the first display unit is used for displaying a first feedback page of query information, and the first feedback page comprises: the information content of the feedback information corresponding to the query information in the first browsing mode;

the second display unit is used for displaying a second feedback page of the query information if the mode switching operation is detected; the second feedback page comprises: the information content of the feedback information in the second browsing mode;

In yet another aspect, an embodiment of the present application provides a computer device, including an input interface and an output interface, the computer device further including:

a processor adapted to implement one or more instructions; the method comprises the steps of,

a computer storage medium storing one or more instructions adapted to be loaded by the processor and to perform the steps of:

In yet another aspect, embodiments of the present application provide a computer storage medium storing one or more instructions adapted to be loaded by a processor and to perform the steps of:

In yet another aspect, embodiments of the present application provide a computer program product comprising a computer program; the computer program, when being executed by a processor, implements the above mentioned information processing method.

After the feedback information corresponding to the query information is obtained, the embodiment of the application can provide a plurality of browsing modes such as the first browsing mode, the second browsing mode and the like, so that the diversity of the browsing modes can be improved. The first browsing mode can support the object to browse all information content or reduced information content of feedback information through the first feedback page; correspondingly, the second browsing mode can support the object to browse the reduced information content or the whole information content of the feedback information through the second feedback page. Therefore, through the first browsing mode and the second browsing mode, the object can comprehensively know the query information through all information contents, the comprehensiveness of information transmission is improved, the object can rapidly and effectively know the query information through simplifying the information contents, and the effectiveness of information transmission is improved. Further, by supporting the object to freely switch different browsing modes according to actual demands, convenience of object browsing information can be improved, good object experience can be obtained when the object consumes content, and accordingly object viscosity is improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for the description of the embodiments will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present application, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

Fig. 1 is a schematic diagram of interaction between a terminal and a server according to an embodiment of the present application;

fig. 2 is a schematic flow chart of an information processing method according to an embodiment of the present application;

FIG. 3a is a schematic diagram showing a second feedback page according to an embodiment of the present application;

FIG. 3b is a schematic diagram of another embodiment of the present application for displaying a second feedback page;

FIG. 3c is a schematic diagram of a second feedback page according to an embodiment of the present application;

FIG. 3d is a schematic diagram of a content sharing card according to an embodiment of the present application;

FIG. 3e is a schematic diagram of another content sharing card according to an embodiment of the present application;

FIG. 3f is a schematic diagram of updating a browsing identifier according to an embodiment of the present application;

FIG. 3g is a diagram of an embodiment of the present application for updating the history praise amount;

FIG. 3h is a schematic diagram of a redisplay of a reference feedback page according to an embodiment of the present application;

FIG. 4 is a schematic flow chart of obtaining a target abstract text according to an embodiment of the application;

FIG. 5a is a schematic diagram of a target abstract extraction model according to an embodiment of the present application;

fig. 5b is a schematic diagram of a working principle of a decoder in a target digest extraction model according to an embodiment of the present application;

fig. 6 is a schematic structural view of an information processing apparatus according to an embodiment of the present application;

fig. 7 is a schematic structural diagram of a terminal according to an embodiment of the present application.

Detailed Description

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application.

The embodiment of the application provides an information processing method, which provides a mode of reorganizing and simplifying the content structure of feedback information by considering that the preference of an object in content consumption has a emphasis, thereby realizing the provision of a plurality of browsing modes such as a long-tail consumption mode, a simplified consumption mode and the like for the object (namely a user). Wherein, the long tail consumption mode refers to: a browsing mode for displaying the entire information content can be supported; it should be understood that in the long-tail consumption mode, if the terminal screen is smaller and the complete information content cannot be displayed at one time, the display of the complete information content can be realized by adopting a scrolling display mode in the terminal screen. The reduced consumption mode refers to: the browsing mode for displaying the simplified information content can be supported, and the simplified information content is obtained by simplifying all the information content; the reduced information content may be able to summarize the concept or idea of the feedback information in fewer words, e.g. the reduced information content may comprise target summary text corresponding to the entire information content. On the basis of providing browsing modes such as a long-tail consumption mode and a simple consumption mode, the information processing method can also support the object to freely switch different browsing modes according to actual demands so as to know all information content or simple information content of feedback information corresponding to query information through the corresponding browsing modes, so that convenience of object browsing information can be improved, good consumption experience can be obtained when the object consumes content, and accordingly, application viscosity and page stay time of an application are improved, wherein the page stay refers to that the object stays on a page and no operation (such as a sliding operation and a sliding operation) is performed.

In a specific implementation, the information processing method provided by the embodiment of the application can be executed by a terminal device (hereinafter, simply referred to as a terminal); the terminals mentioned herein may include, but are not limited to: smart phones, computers (such as tablet computers, notebook computers, desktop computers, etc.), smart wearable devices (such as smart watches, smart glasses), smart voice interaction devices, smart home appliances (such as smart televisions), vehicle terminals or aircrafts, etc. Further, various applications (i.e., APP) may be installed and run in the terminal, such as an information query application (an application that may provide an information query service), a social application, an audio/video playing application, and so on; in this case, the information processing method proposed by the embodiment of the present application may also be executed by the information inquiry application. It should be understood that if the social application or the audio-video playing application may also provide the information query service, the method may also be performed by the social application or the audio-video playing application. Optionally, the information processing method provided by the embodiment of the present application may also be executed by an information query applet (i.e. an applet that can provide an information query service), where the applet refers to an application that can run in an installed APP without installing and downloading.

In another specific implementation, the information processing method provided by the embodiment of the application can be executed by the terminal and the server together; in this case, the terminal and the server may communicate via a wired network or a wireless network, which is not limited. The server mentioned herein may be an independent physical server, or may be a server cluster or a distributed system formed by a plurality of physical servers, or may be a cloud server that provides cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDNs (Content Delivery Network, content delivery networks), basic cloud computing services such as big data and artificial intelligence platforms, and the like. Specifically, see fig. 1: after acquiring query information input by the object, the terminal can send the query information to the server, the server is responsible for determining feedback information corresponding to the query information, and performing reduction processing on all information content of the feedback information to obtain reduced information content, so that the reduced information content and all information content of the feedback information are issued to the terminal, and the terminal can output information content corresponding to a browsing mode selected by the display object for the object based on the reduced information content and all information content. It should be understood, of course, that in order to improve the timeliness of information feedback, a large amount of feedback information corresponding to query information may be stored in advance in the server, and reduced information content corresponding to each feedback information may be generated and stored in advance, so that after query information input by the object is obtained, corresponding feedback information and corresponding reduced information content may be directly searched from the storage space and sent to the terminal.

The information processing method according to the embodiment of the present application is further described below with reference to the flowchart shown in fig. 2; for convenience of explanation, the embodiment of the application mainly takes the terminal to execute the information processing method as an example. Referring to fig. 2, the information processing method may include the following steps S201 to S202:

s201, displaying a first feedback page of query information, wherein the first feedback page comprises: and inquiring information content of feedback information corresponding to the information in the first browsing mode.

The query information refers to information to be queried input by a target object (a user using a terminal for executing the method), and the feedback information corresponding to the query information is information for feeding back the query information. For example, the query information may be terms or thesaurus in the encyclopedia entry entered by the target object, and the corresponding feedback information may be paraphrasing in the encyclopedia entry; if the target object inputs the term "whale", the feedback information is a relevant paraphrase explaining what is whale. For another example, the query information may be any question input by the target object, and then the corresponding feedback information may be an answer corresponding to the question; if the target object inputs "what the principle of machine learning is," the feedback information is a related article about the principle of machine learning.

In an embodiment of the present application, the first browsing mode may be the long tail consumption mode mentioned above; in this case, the information content of the feedback information corresponding to the query information in the first browsing mode refers to the entire information content of the feedback information. Alternatively, the first browsing mode may be the reduced consumption mode mentioned above; in this case, the information content of the feedback information corresponding to the query information in the first browsing mode refers to the reduced information content of the feedback information.

S202, if the mode switching operation is detected, displaying a second feedback page of the query information; the second feedback page includes: and feeding back information content of the information in the second browsing mode.

It should be understood that, if the first browsing mode mentioned in the aforementioned step S201 is the long-tail consuming mode, the second browsing mode mentioned herein is the reduced consuming mode; in this case, the information content of the feedback information in the second browsing mode refers to the reduced information content of the feedback information, and the mode switching refers to switching from the long-tail consumption mode to the reduced consumption mode. If the first browsing mode mentioned in the aforementioned step S201 is the reduced consumption mode, the second browsing mode mentioned here is the long-tail consumption mode; in this case, the information content of the feedback information in the second browsing mode refers to the entire information content of the feedback information, and the mode switching refers to switching from the reduced consumption mode to the long-tail consumption mode.

That is, in the first feedback page and the second feedback page, there is a case where the information content in one feedback page is the entire information content of the feedback information, and the information content in the other feedback page is the reduced information content obtained based on the entire information content. It should be understood that after the second feedback page is displayed, the target object may be further supported to trigger the terminal to execute mode switching again by inputting a new mode switching operation, so as to display the first feedback page; namely, the embodiment of the application can support the target object to continuously switch modes between the first browsing mode and the second browsing mode.

In a specific implementation, any feedback page may include: a first mode component corresponding to the first browsing mode, and a second mode component corresponding to the second browsing mode; the first mode component and the second mode component may be located in a bottom navigation bar (bottom bartab) of the feedback page, or may be located at other positions, which is not limited. When any feedback page is displayed, the mode components corresponding to the corresponding browsing modes are in a selected state, and other mode components are in an unselected state; that is, when the first feedback page is displayed, the first mode component is in the selected state, the second mode component is in the unselected state, and when the second feedback page is displayed, the first mode component is in the unselected state, and the second mode component is in the selected state. In this case, the mode switching operation may include: the operation of the second mode component is selected in the first feedback page. Optionally, the mode switching operation may be an operation of inputting a specified gesture, where the specified gesture may be set according to an actual requirement, for example, the specified gesture may be a gesture sliding horizontally to the left; alternatively, the mode switching operation may also be an operation of inputting a language password for instructing to switch from the first browsing mode to the second browsing mode; still alternatively, the mode switching operation may be an operation of pressing or touching a physical component (e.g., a volume key, a power key) on the terminal, or the like.

Further, the manner in which the terminal displays the second feedback page of the query information may be any of the following: and switching from the first feedback page to the second feedback page in a page switching mode so as to display the second feedback page. Or, directly outputting and displaying a second feedback page on the first feedback page; the second feedback page in this case may be a mask page (i.e. a page that is located on the first feedback page and has a certain transparency), or may be a page that does not have transparency, which is not limited. Or, suspending and displaying a window on the first feedback page, and displaying a second feedback page in the window; in this case, the second feedback page may be understood as a sub-page that is displayed independently of the first feedback page.

Based on the description, taking query information as 'whale' in encyclopedia entry, the terminal displays a second feedback page in a page switching mode, wherein a first browsing mode corresponding to the first feedback page is a long-tail consumption mode, and a second browsing mode corresponding to the second feedback page is a simplified consumption mode. The schematic diagram of the target object triggering the terminal to display the second feedback page by selecting the second mode component in the first feedback page can be seen in fig. 3 a; the schematic diagram of the target object triggering the terminal to display the second feedback page by inputting a gesture sliding horizontally to the left can be seen in fig. 3 b. It should be understood that, limited to the display size of the terminal screen, the first feedback page in fig. 3a and fig. 3b only displays a part of the information content of the whole information content, and the target object may trigger the terminal to scroll through the first feedback information in a manner of sliding up and down to display the remaining information content that is not displayed, so as to realize browsing the whole information content through the first feedback page. Also, the two components "all encyclopedia" and "reduced encyclopedia" in fig. 3a and 3b are merely exemplary and not limiting as to the manner in which the first mode component and the second mode component are respectively represented; wherein "full encyclopedia" corresponds to a first schema component and "reduced encyclopedia" corresponds to a second schema component. In addition, in the scenario shown in fig. 3b, the first mode component and the second mode component may not be included in both pages.

It should be noted that, in the case that the first feedback page includes all information contents, the information contents in the second feedback page are reduced information contents; the reduced information content may be obtained by reducing the entire information content of the feedback information, or may be obtained by reducing the target information content in the entire information content, which is not limited thereto. The target information content can be information content selected by the target object from all information contents according to own requirements. The specific application scenario may be as follows: after the first feedback page is displayed, if the terminal does not detect the content selection operation for all the information content before detecting the mode switching operation, the reduced information content in the second feedback page is obtained by reducing all the information content of the feedback information. If the content selection operation is detected before the mode switching operation is detected, the reduced information content in the second feedback page is obtained by reducing the target information content (such as abstract extraction); that is, before detecting the mode switching operation, the terminal may acquire a content selection operation for all the information contents and select a target information content of all the information contents according to the content selection operation; after detecting the mode switching operation, the terminal may perform a summary extraction process on the target information content to obtain a simplified information content, and trigger the step of executing a second feedback page for displaying the query information, as shown in fig. 3 c.

For convenience of description, a feedback page displaying reduced information content among the first feedback page and the second feedback page may be referred to as a reduced feedback page; and, among the first feedback page and the second feedback page, a feedback page displaying the entire information content may be referred to as a reference feedback page. Also, the reference feedback page may also include video identifications of one or more videos related to the query information, which may include, but are not limited to: video covers, video names, and video web page links, etc.; then in the case where the video identification contains video covers, the condensed feedback page may also include video covers for one of the videos, as shown in fig. 3 a-3 c, described above.

In an alternative embodiment, the reduced feedback page may further include: a sharing component (identified by 31) for sharing the condensed information content. In this case, if the target object wants to share the reduced information content with other objects (such as other users or other groups), a trigger operation may be performed on the sharing component. Accordingly, in the process of displaying the reduced feedback page, if the terminal detects the triggering operation for the sharing component, the terminal may respond to the triggering operation for the sharing component to display the content sharing card 32 corresponding to the reduced information content on the reduced feedback page. Wherein the content sharing card 32 may include the reduced information content 321; optionally, the content sharing card 32 may further include a graphic code 322, and when the graphic code 322 is scanned by any terminal, a reduced feedback page may be output and displayed in any terminal. Further, in the case where a video cover is included in the condensed feedback page, the content sharing card may also include a thumbnail of the video cover (identified as 323). Taking the content sharing card 32 as an example, the content sharing card 32 includes the reduced information content 321, the graphic code 322 and the thumbnail, a schematic diagram of the content sharing card 32 is shown in fig. 3 d; it should be understood that, in other embodiments, the terminal may also switch from the reduced feedback page to the content sharing interface, so as to display the content sharing card in the content sharing interface.

After the content sharing card is displayed, the target object can execute sharing operation on the content sharing card; correspondingly, if the terminal detects the sharing operation for the content sharing card, the terminal can send the content sharing card to the object indicated by the sharing operation. In a specific implementation, when the terminal displays the content sharing card, one or more object identifiers 33 are also displayed synchronously, as shown in fig. 3 e; in this case, the sharing operation may include: and selecting at least one object identifier. That is, after the target object selects at least one object identifier, the terminal may be triggered to send the content sharing card to the object indicated by the sharing operation (i.e. the object indicated by each selected object identifier). In another specific implementation, after the content sharing card is displayed, the terminal can automatically store the content sharing card into the local space; in this case, the sharing operation may include: and sending the operation of the content sharing card in the session interface. That is, the target object may use the terminal to open a session interface with other objects, and send the content sharing card stored in the local space to the other objects through the session interface.

In another alternative embodiment, the reduced feedback page may further include: browse identification (identified with 34); the browse identification is used for indicating at least one of the following: a historical view amount 341 of the reduced information content, and an object identification 342 of all or a portion of the objects of the reduced information content. Correspondingly, if the terminal detects that the simplified information content is browsed by a new object in the process of displaying the simplified feedback page, the browsing identification can be updated and displayed in the simplified feedback page. The new object may be a target object or an object using other terminals, which is not limited. And, updating the display browsing identification may include at least one of: updating the historical browsing amount and updating the object identification; see, for example, fig. 3 f: if the historical browsing amount is 65, after the simplified information content is browsed by the new object, the historical browsing amount can be updated to 66, and the displayed object identifier is updated by adopting the object identifier 343 of the new object.

In another alternative embodiment, the reduced feedback page may further include: historical praise amounts (identified with 35) and praise components (identified with 36) of the condensed information content. Correspondingly, if the terminal detects that the praise component is triggered in the process of displaying the simplified feedback page, the display history praise amount can be updated in the simplified feedback page, as shown in fig. 3 g. Further, in the process of displaying the reduced feedback page, if the reduced information content is detected to be simultaneously prayed by the target object and at least one other object, the terminal can play the praying animation on the reduced feedback page according to the virtual object corresponding to the target object and the virtual objects corresponding to the other objects; and updating the history praise amount after the praise animation is played. That is, when there are at least two objects that praise the condensed information content at the same time, the praise animation can be played based on the virtual objects of the at least two objects, so that the interestingness of praise can be effectively improved.

In another alternative embodiment, the whole information content includes: video identification of one or more videos related to the query information. In this case, in the process of displaying the reference feedback page, if any video identifier in all the information contents is detected to be triggered, a video playing page can be displayed; then, the video corresponding to the triggered video identifier is played in the video playing page, and the sharing component 31 for sharing the condensed information content is displayed on the video playing page. If the sharing component 31 is triggered, the content sharing card 32 corresponding to the reduced information content is displayed on the video playing page. Further, in the process of displaying the content sharing card, if the target interaction operation is detected, the display of the video playing page and the content sharing card is canceled, and the reference feedback page is redisplayed. The target interaction operation comprises the following steps: a sliding operation along a specified direction on a video play page; alternatively, the video playing page further includes a page returning component 37, and the target interaction operation includes: triggering operations for the page return component 37. Taking the example that the target interaction operation includes a trigger operation for the page returning component 37, the process of displaying the video playing page and redisplaying the reference feedback page by the terminal can be seen in fig. 3 h.

After the feedback information corresponding to the query information is obtained, the embodiment of the application can provide multiple browsing modes such as the first browsing mode, the second browsing mode and the like for the object, so that the diversity of the browsing modes can be improved. The first browsing mode can support the object to browse all information content or reduced information content of feedback information through the first feedback page; correspondingly, the second browsing mode can support the object to browse the reduced information content or the whole information content of the feedback information through the second feedback page. Therefore, through the first browsing mode and the second browsing mode, the object can comprehensively know the query information through all information contents, the comprehensiveness of information transmission is improved, the object can rapidly and effectively know the query information through simplifying the information contents, and the effectiveness of information transmission is improved. Further, different browsing modes can be freely switched by the support object according to actual requirements, so that convenience of object browsing information can be improved, good consumption experience can be obtained when the content of the object is consumed, and application viscosity of the application is improved.

Based on the above-described related description of the method embodiment shown in fig. 2, a specific manner of how the above-mentioned reduced information content is obtained is explained below. Specifically, the terminal or the server can perform abstract extraction processing on all information contents of the feedback information to obtain a target abstract text, so that the simplified information contents are constructed by adopting the target abstract text; that is, in this case, the reduced information content includes: and carrying out abstract extraction processing on all information contents of the feedback information to obtain target abstract text. The abstract is characterized in that the data volume of the output text is much smaller than that of the input text, but can store much effective information; in a specific implementation, the embodiment of the application can adopt a strategy of an understanding type abstract (abstract) to realize the extraction of the target abstract text. Referring to fig. 4, the method for obtaining the target abstract text may include the following steps S401 to S404:

S401, constructing a word list based on all information contents of the feedback information, and splitting all information contents of the feedback information into K sentences, wherein K is a positive integer.

The word list comprises a plurality of high-frequency words, wherein the high-frequency words refer to words with word frequencies larger than a preset threshold value in feedback information, and the word frequencies refer to the frequency of occurrence of the words in the whole information content. Specifically, the way to construct the vocabulary based on the entire information content of the feedback information may be: firstly, word segmentation processing is carried out on all information contents of feedback information to obtain a plurality of words; and counting the number of times each word of the plurality of words appears in the whole information content respectively as the word frequency of each word. Then, determining words with word frequency greater than a preset threshold value as high-frequency words, and determining words with word frequency less than or equal to the word frequency as low-frequency words (the low-frequency words can be marked as UNK); then, a vocabulary can be constructed by adopting the determined high-frequency words, and the size of the vocabulary can be v, wherein v is a positive integer.

And S402, performing abstract extraction processing on each sentence in the K sentences according to the word list to obtain K initial abstract texts and the credibility corresponding to each initial abstract text.

In a specific implementation, a target abstract extraction model can be obtained, and the target abstract extraction model can adopt an understanding abstract (abstract) strategy to realize abstract text extraction; thus, the initial abstract text of each sentence and the corresponding credibility of each initial abstract text are obtained by the aid of the target abstract extraction model. Referring to fig. 5a, the target digest extraction model may include at least two parts, namely an encoder (encoder) and a decoder (decoder); wherein, the encoder can use single-layer or multi-layer RNN (Recurrent Neural Network, cyclic neural network), LSTM (Long Short-Term Memory), GRU (Gated Recurrent Unit, threshold cyclic unit) and other networks to encode each input word in the input information (such as a sentence); accordingly, the decoder can be understood as a language model, which can understand the input information based on the encoding result output by the encoder, so as to generate the abstract text corresponding to the input information. The abstract extraction model may be a seq2seq (sentence-to-sentence) model, or a seq2seq+attention model in NLP (Natural Language Processing ), for example.

For the kth sentence in K sentences, the kth sentence is segmented into M input words, K E [1, K]M is a positive integer; based on this, using X to represent any input word, the kth sentence can be represented as a word sequence X _t ：X _t ＝|x ₁ ，x ₂ ，…，x _M | a. The invention relates to a method for producing a fibre-reinforced plastic composite. The initial abstract text corresponding to the kth sentence is determined according to the target words output by the decoder in N time steps. Wherein the decoder is configured to decode the encoded data at t (t.e [1, N)]) The determination of the target word output by the time steps may include the following steps s11-s15:

and s11, at the t-th time step of the decoder, invoking an encoder to sequentially recursively encode each input word in the M words to obtain the hidden state of each input word at the t-th time step.

In a specific implementation, at the t-th time step of the decoder, a first input word of the M input words can be input into the encoder, and the encoder encodes the first input word to obtain the hidden state of the first word at the t-th time step; then, a second input word in the M input words can be input into the encoder, and the encoder encodes the second input word based on the hidden state corresponding to the first word to obtain the hidden state of the second word in the t-th time step; then, a third input word of the M input words may be input to the encoder, the encoder encodes the third input word based on the hidden state corresponding to the second word, to obtain the hidden state of the third word at the t-th time step, and so on, until the hidden state of each input word of the M input words at the t-th time step is obtained.

s12, determining the attention weight of the hidden state of each input word at the t time step according to the attention degree of each input word according to the target word required to be generated by the decoder at the t time step.

In a specific implementation process, for any input word, the attention weight of the hidden state of any input word in each historical time step can be obtained, wherein the historical time step refers to a time step before the t time step. Then, the attention weight of the hidden state of any input word in each historical time step can be summed, and the result of the summation operation is used as an attention weight decision factor. After the attention weight decision factor is obtained, the attention weight of the hidden state of any input word at the t-th time step can be determined according to the attention weight decision factor and the attention degree of any input word by the target word required to be generated by the decoder at the t-th time step. By adding the attention weights of the previous time steps together to decide the attention weight affecting the current time step (i.e. the t-th time step), continuing to consider the part that has already acquired the high weight can be avoided, so that repetition at the same location can be avoided, and further repetition of generating the same target word can be avoided, so as to alleviate the problem of generating repetition.

s13, based on the attention weight of the hidden state of each input word at the t-th time step, integrating the hidden states of the M input words at the t-th time step to obtain the semantic vector of the kth sentence at the t-th time step. The integration mode can be as follows: the weighting fusion or the weighting concatenation is not limited to this.

And S14, calculating the output probability of each high-frequency word in the output word list of the t-th time step according to the semantic vector, and obtaining a probability calculation result.

In one embodiment, RNNLM (recurrent neural network language model) may be used to calculate the output probability of each high frequency word in the output vocabulary at the t-th time step from the semantic vector, resulting in a probability calculation result. Among them, RNNLM has the following advantages: when the output probability of each high-frequency word is determined in the t-th time step, all previous upper information (namely the information of the target word output by each historical time step) can be utilized instead of only the information of the target word output by the t-1 time step, so that the accuracy of probability calculation can be improved. In this embodiment, P is used _t Representing the output probability of any high-frequency word, P _t May be specifically expressed as P (y _t |{y ₁ ，y ₂ ，…，y _t-1 }，X _t θ) this formula; wherein y is _t For the target word output in the t-th time step, X _t For the kth sentence, θ includes the semantic vector of the kth sentence at the time step of the kth sentence and any high-frequency word, and θ may be different depending on the high-frequency word, so that P _t Different.

In another embodiment, the output probability of each high-frequency word in the output vocabulary of the t-th time step can be calculated according to the semantic vector based on the working principle of the decoder shown in fig. 5b, so as to obtain a probability calculation result. Specifically, the target word (y is adopted) output by the decoder at the t-1 time step can be obtained _t-1 Representation) of the corresponding first weight W ₁ The decoder is in the hidden state at time step t-1 (using h _t-1 Representation) of the corresponding second weight W ₂ And a third weight W corresponding to the semantic vector ₃ . Then based on the first weight W ₁ Second weight W ₂ Third weight W ₃ For the target word y output by the decoder at the t-1 time step _t-1 Hidden state h of decoder at time step t-1 _t-1 And semantic vector c _t Carrying out weighted summation to obtain a first weighted summation result; and activating the first weighted sum result by using a sigmoid function (an activation function) to obtain the hidden state of the decoder at the t-th time step (using h _t Representation) and then based on the fourth weight W4 and the fifth weight W5, the hidden state h of the decoder at the t-th time step _t And semantic vector c _t Carrying out weighted summation to obtain a second weighted summation result; and performing activation processing on the second weighted summation result by adopting a softmax function (an activation function) to obtain an activation result P; then, the output probabilities of the high-frequency words in the output vocabulary of the t-th time step can be calculated based on the activation result P, respectively, to obtain a probability calculation result. For example, the calculation formulas of the hidden state ht and the activation result P of the decoder at the t-th time step can be shown in the following formulas 1.1 and 1.2, respectively:

h _t ＝sigmoid(W ₁ y _t-1 +W ₂ h _t-1 +W ₃ c _t )∈R ^d×1 1.1

P＝softmax(W ₄ h _t +W ₅ c _t ) 1.2

Where d is the number of hidden neurons. Based on the above description, the decoder needs to take the semantic vector as input when decoding each time step, instead of introducing the semantic vector only in the first time step; also, since the degree of attention of each input word in the input sequence (i.e., kth sentence) is different when generating the target word per time step due to the presence of the attention mechanism (attention mechanism), the semantic vector given by the encoder is different per time step. It is also notable that embodiments of the present application do not explicitly input y when calculating the activation result P _t-1 But only in calculating the hidden state h _t When input y _t-1 。

And s15, selecting a high-frequency word from the word list according to the probability calculation result, and using the high-frequency word as a target word output by the decoder in the t-th time step.

In one embodiment, the generated summary text may be generalized to solve a problem p (word|context) of a conditional probability, where context represents the input sentence, word represents the target word, and p represents the output probability; and then under context condition, calculating the output probability of each high-frequency word in the word list, and using the high-frequency word with the largest output probability as the output target word, thereby sequentially generating all the target words in the abstract. Based on this, the specific implementation of step s15 may be: and selecting the high-frequency word with the maximum output probability from the word list as a target word output by the decoder at the t-th time step according to the probability calculation result. In this case, the initial abstract text corresponding to the kth sentence is specifically obtained by the following method: and combining all target words output by the decoder in N time steps according to the sequence of the time steps to obtain an initial abstract text corresponding to the kth sentence.

In another embodiment, the beam search method is used to determine each target word, so as to obtain the final initial abstract text; the core logic is as follows: when the target word is determined for the first time, the first K high-frequency words in the word list are reserved as target words according to the order of the output probability from high to low; when the target word is determined for the second time, each high-frequency word in the first reserved K high-frequency words can be respectively combined with each high-frequency word in the word list to obtain Kv binary phrases and the probability corresponding to each binary phrase, the probability corresponding to any binary phrase is equal to the product of the output probabilities of two high-frequency words forming the binary phrase, so that the first K binary phrases in the Kv binary phrases are reserved according to the order of the probability from high to low, and the second high-frequency word in each binary phrase in the first K binary phrases is the target word determined for the second time; and so on, the first k target words are reserved each time, and the pruning operation greatly reduces the search space. Based on this, the specific implementation of step s15 may be: and sequencing each high-frequency word in the word list according to the sequence of the output probability from large to small, and selecting the first k high-frequency words from the sequencing result as target words output by the decoder in the t time step. In this case, the initial abstract text corresponding to the kth sentence is specifically obtained by the following method: and determining N-gram phrases corresponding to each target word in k target words output in the N-th time step, and selecting the N-gram phrase with the highest probability as an initial abstract text corresponding to the kth sentence.

Further, to avoid the OOV (unknown word) problem (i.e., the problem that the target word to be generated may not be in the vocabulary), the magnitude relation between the maximum output probability in the probability calculation result and the probability threshold may be compared before step s15 is performed. If the maximum output probability is greater than or equal to the probability threshold, it may indicate that the target word to be generated in the t-th time step is in the vocabulary, and at this time, the execution of step s15 may be triggered. If the maximum output probability is smaller than the probability threshold, it may indicate that the target word to be generated in the t time step may not be in the vocabulary, and at this time, according to the attention weight of the hidden state of each input word in the t time step, the input word corresponding to the hidden state with the maximum attention weight may be selected from the M input words and used as the target word output by the decoder in the t time step, so that the OOV problem may be effectively alleviated.

S403, selecting a target number of initial abstract texts from the K initial abstract texts according to the corresponding credibility of each initial abstract text in the order of high credibility; the target number may be set according to an empirical value, which is not limited.

S404, combining the initial abstract texts of the selected target quantity to obtain target abstract texts.

Specifically, the initial abstract texts with the selected target number can be spliced according to the arrangement sequence of sentences corresponding to the selected initial abstract texts in the feedback information, so that the target abstract texts are obtained.

Based on the description of steps S401 to S404, it should be noted that the target abstract extraction model used in the above process may be obtained by performing model training on the initial abstract extraction model based on a machine learning/deep learning technique in an artificial intelligence (Artificial Intelligence, AI) technique. Wherein, AI technology refers to: theory, methods, techniques and application systems that utilize digital computers or digital computer-controlled machines to simulate, extend and extend human intelligence, sense the environment, acquire knowledge and use knowledge to obtain optimal results. In other words, artificial intelligence is a comprehensive technique of computer science; the intelligent machine is mainly used for producing a novel intelligent machine which can react in a similar way of human intelligence by knowing the essence of the intelligence, so that the intelligent machine has multiple functions of sensing, reasoning, decision making and the like. The so-called machine learning is the core of AI, and one of the multi-domain interdisciplines involves multiple disciplines such as probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory, and the like. Specially researching how a computer simulates or realizes the learning behavior of human beings so as to acquire new knowledge or skills, and reorganizing the existing knowledge structure to continuously improve the performance of the computer; deep learning is a technique for machine learning by using a deep neural network system.

The process of training the initial abstract model to obtain the target abstract model may be approximately as follows:

first, S samples for model training may be obtained, S being a positive integer. During training, an English data set related to CNN (convolutional neural network)/DailyMail (a public database) and an encyclopedic Chinese data set maintained by the user can be adopted; based on this, for example, one or more terms may be obtained from the encyclopedia chinese database, each term obtained may be subjected to sentence splitting to obtain S sentences, each sentence in the S sentences may be manually summarized and abstracted to obtain a first abstract text, and then each sentence and the corresponding first abstract text may be respectively adopted to construct S samples, where one sample includes one sentence and one first abstract text. After S samples are obtained, the initial extraction model can be called to extract the abstract of sentences in each sample, so that second abstract text corresponding to each sample and output probability of each target word in each second abstract text are obtained.

From the foregoing, it is necessary to train the initial abstract extraction model to obtain the target abstract extraction model with better performance, because the target abstract extraction model is required to be invoked to extract the abstract from the sentence. Based on this, the input in model training may be sentences in each sample (which may be understood as a word sequence), and sentences in the ith sample may be represented as X ⁱ ＝|x ⁱ ₁ ，x ⁱ ₂ ，…，x ⁱ _M |，i∈[1，S]The method comprises the steps of carrying out a first treatment on the surface of the Where x is ⁱ ₁ 、x ⁱ ₂ X ⁱ _M Refers to the words in the sentence in the i-th sample. Correspondingly, the second abstract text output by the model can be understood as a word sequence, and the second abstract text output by the model by the sentence in the ith sample can be expressed as Y ⁱ ＝[y ⁱ ₁ ，y ⁱ ₂ ，…，y ⁱ _N ]It should be noted that, the second abstract text Y corresponding to different samples ⁱ The number of words of (a) may be different, i.e. the second abstract text Y corresponding to different samples ⁱ The value of N in (c) may be different. Since the text length of the input and output of the model is the same, the definition of the abstract is not met, and therefore the value of N corresponding to each sample can be limited to be smaller than that of M.

In the embodiment of the present application, the object of the model training may be to maximize the output probability of each word in the second abstract text; using y to represent any word in any second abstract text, the goal of model training can be represented using the following equation 1.3:

based on this, after obtaining the output probabilities of the words in each second abstract text, the penalty values may be minimized in the manner of SGD (Stochastic Gradient Descent, random gradient descent), thereby enabling an optimized update of the model parameters to complete the model training. Specifically, a model loss value can be calculated according to the output probability of each target word in each second abstract text by adopting a loss function, and then the model loss value is reversely propagated in an SGD mode to obtain a gradient value, so that the gradient value is adopted to optimize model parameters of an initial abstract extraction model, the initial abstract extraction model is converged, and the converged initial abstract extraction model is used as a target abstract extraction model. Wherein, L is used to represent the loss value, taking the output probability of each word in each second abstract text obtained by RNNLM (recurrent neural network language model) as an example, namely taking the output probability of any word in any second abstract text as P (y ⁱ _t |{y ⁱ ₁ ，y ⁱ ₂ ，…，y ⁱ _t-1 }，X ⁱ _t θ) is an example, the calculation formula of the model loss value can be shown in the following formula 1.4:

it should be noted that, the specific manner of the initial abstract extraction model for extracting the abstract of the sentence in each sample is similar to the specific manner of extracting the abstract of the whole information content of the feedback information by the target abstract extraction model, which is described in the foregoing, and is not repeated here. And is also provided withThe corresponding method mentioned above can also be employed to avoid OOV problems and to generate duplicate problems when extracting the second abstract text of the sentence in each sample. From the foregoing, it can be seen that the main principle of the correlation method for avoiding the generation of repetition is: the attention weights of the previous time steps are added together to obtain the corresponding attention weight decision factor (using c ^t Representation) to influence the decision of the attention weight of the current time step by an attention weight decision factor; based on this, it is necessary for the embodiment of the present application to add a loss to the attention weight decision factor when calculating the model loss value. Specifically, the corresponding loss value can be calculated by the following equation 1.5:

the full term of covloss may be coverage loss, which refers to a loss value calculated based on an attention weight decision factor; a, a _i ^t Representing the attention weight related to the t time step corresponding to the i sample; c _i ^t Attention weight decision factors for influencing the attention weight involved in the t-th time step corresponding to the i-th sample are shown in the following formula 1.6:

it should be noted that: all t mentioned in the above formulas 1.5-1.6 represent the time steps at the decoder side; i represents the token index (information indicating samples) at the encoder side, i e 0, 99 if there are 100 samples at the encoder side input]) The method comprises the steps of carrying out a first treatment on the surface of the Σi represents traversing each time step at the decoder side to compare the two sizes, choosing the smallest. Taking the first sample and the first time step as an example, equation 1.5 shows that from a ₁ ¹ And c ₁ ¹ Selecting the smallest value as covloss ₁ If a is ₁ ¹ And c ₁ ¹ Of interestIs equally distributed, covloss ₁ Will be large so that it will choose a different distribution, since the smallest of the two is chosen, so that covloss ₁ Will be smaller and the end purpose is to have each distribution be different, and duplicate words can be avoided. As such, coverage loss is a bounded quantity; based on this, in the above formula 1.4This part can be replaced by the following equation 1.7:

Wherein P (w) _t * ) Representing the output probability of the target word calculated by the model at the t-th time step; namely P (w) _t * ) Namely in the above formula 1.4

Based on the above description, the embodiment of the application uses the means of machine learning and artificial intelligence to improve the performance of the model, so that the finally obtained target abstract extraction model can more efficiently perform abstract extraction processing, and improve the abstract extraction efficiency. And the target abstract extraction model can accurately extract useful information, and allows new words and phrases to be generated to form abstract texts, so that experience exploration of a simplified consumption mode is formed, free consumption of objects between a long-tail consumption mode and the simplified consumption mode is supported, and different object requirements are met.

Based on the description of the related embodiments of the information processing method, the embodiment of the application further provides an information processing device; in particular, the apparatus may be a computer program (comprising program code) running in a terminal, and the apparatus may perform part of the method steps in the method flow shown in fig. 2 or fig. 4. Referring to fig. 6, the apparatus may operate as follows:

the first display unit 601 is configured to display a first feedback page of query information, where the first feedback page includes: the information content of the feedback information corresponding to the query information in the first browsing mode;

The second display unit 602 is configured to display a second feedback page of the query information if a mode switching operation is detected; the second feedback page comprises: the information content of the feedback information in the second browsing mode;

In one embodiment, any feedback page includes: a first mode component corresponding to the first browsing mode and a second mode component corresponding to the second browsing mode;

when any feedback page is displayed, the mode components corresponding to the corresponding browsing modes are in a selected state, and other mode components are in an unselected state;

wherein the mode switching operation includes: and selecting the operation of the second mode component in the first feedback page.

In another embodiment, among the first feedback page and the second feedback page, a feedback page displaying the reduced information content is referred to as a reduced feedback page; the reduced feedback page further includes: the sharing component is used for sharing the simplified information content;

Accordingly, the first display unit 601 or the second display unit 602 may further be configured to:

in the process of displaying the simplified feedback page, responding to the triggering operation aiming at the sharing component, and displaying a content sharing card corresponding to the simplified information content on the simplified feedback page;

and if the sharing operation aiming at the content sharing card is detected, the content sharing card is sent to an object indicated by the sharing operation.

In another embodiment, among the first feedback page and the second feedback page, a feedback page displaying the reduced information content is referred to as a reduced feedback page; the simplified feedback page further comprises a browsing identifier, wherein the browsing identifier is used for indicating at least one of the following: the historical browsing amount of the simplified information content and the object identification of all or part of the objects of the simplified information content in a historical browsing manner;

and in the process of displaying the simplified feedback page, if the fact that the simplified information content is browsed by a new object is detected, updating and displaying the browsing identification in the simplified feedback page.

In another embodiment, among the first feedback page and the second feedback page, a feedback page displaying the reduced information content is referred to as a reduced feedback page; the reduced feedback page further includes: the history praise amount of the simplified information content;

in the process of displaying the simplified feedback page, if the simplified information content is detected to be simultaneously subjected to the praise operation by the target object and at least one other object, a praise animation is played on the simplified feedback page according to the virtual object corresponding to the target object and the virtual objects corresponding to the other objects;

and after the praise animation playing is finished, updating the historical praise amount.

In another embodiment, among the first feedback page and the second feedback page, a feedback page displaying the entire information content is referred to as a reference feedback page; the total information content comprises the following components: a video identification of one or more videos related to the query information;

In the process of displaying the reference feedback page, if any video identifier in the whole information content is detected to be triggered, displaying a video playing page;

playing the video corresponding to the triggered video identifier in the video playing page, and displaying a sharing component for sharing the simplified information content on the video playing page;

and if the sharing component is triggered, displaying a content sharing card corresponding to the simplified information content on the video playing page.

In another embodiment, the first display unit 601 or the second display unit 602 may be further used for:

in the process of displaying the content sharing card, if the target interaction operation is detected, the video playing page and the content sharing card are canceled to be displayed, and the reference feedback page is redisplayed;

wherein the target interaction operation includes: a sliding operation along a specified direction on the video play page; or, the video playing page further comprises a page returning component, and the target interaction operation comprises: triggering operation for the page return component.

In another embodiment, the first feedback page includes the entire information content; accordingly, the first display unit 601 or the second display unit 602 may be further configured to:

Before the mode switching operation is detected, acquiring a content selection operation for all the information contents, and selecting a target information content in all the information contents according to the content selection operation;

and after the mode switching operation is detected, performing abstract extraction processing on the target information content to obtain the simplified information content, and triggering and executing a step of displaying a second feedback page of the query information.

In another embodiment, the reduced information content includes: the target abstract text obtained by abstract extracting the whole information content of the feedback information, wherein the obtaining mode of the target abstract text can be executed by the first display unit 601 or the second display unit 602, and specifically comprises the following steps:

constructing a word list based on all information contents of the feedback information, and splitting all information contents of the feedback information into K sentences, wherein K is a positive integer; the word list comprises a plurality of high-frequency words, wherein the high-frequency words refer to words with word frequencies larger than a preset threshold value in the whole information content;

performing abstract extraction processing on each sentence in the K sentences according to the word list to obtain K initial abstract texts and the credibility corresponding to each initial abstract text;

According to the sequence of the credibility from high to low, selecting a target number of initial abstract texts from the K initial abstract texts according to the corresponding credibility of each initial abstract text;

and combining the initial abstract texts of the selected target quantity to obtain target abstract texts.

In another embodiment, for the kth sentence in the K sentences, the kth sentence is segmented into M input words, K e [1, K ], M being a positive integer; the initial abstract text corresponding to the kth sentence is determined according to target words output by the decoder in N time steps;

the method for determining the target word output by the decoder at the t-th time step is as follows:

calling an encoder to carry out recursive encoding on each input word in the M words in turn at the t time step of the decoder so as to obtain the hidden state of each input word at the t time step; wherein t is [1, N ];

determining the attention weight of the hidden state of each input word in the t time step according to the attention degree of each input word of the target word required to be generated by the decoder in the t time step;

Integrating the hidden states of the M input words at the t time step based on the attention weight of the hidden states of each input word at the t time step to obtain semantic vectors of the kth sentence at the t time step;

calculating the output probability of each high-frequency word in the word list at the t-th time step according to the semantic vector, and obtaining a probability calculation result; and selecting a high-frequency word from the word list according to the probability calculation result to serve as a target word output by the decoder in the t-th time step.

In another embodiment, the first display unit 601 or the second display unit 602 may be specifically configured to, when determining the attention weight of the hidden state of each input word at the t-th time step according to the attention degree of each input word by the target word that the decoder needs to generate at the t-th time step:

for any input word, acquiring the attention weight of the hidden state of the any input word in each historical time step, wherein the historical time step is the time step before the t-th time step;

carrying out summation operation on the attention weight of the hidden state of any input word in each historical time step, and taking the summation operation result as an attention weight decision factor;

And determining the attention weight of the hidden state of any input word in the t time step according to the attention weight decision factor and the attention degree of the target word to be generated by the decoder in the t time step on any input word.

In another embodiment, when the first display unit 601 or the second display unit 602 is configured to select a high-frequency word from the vocabulary according to the probability calculation result, the high-frequency word may be specifically used as the target word output by the decoder at the t-th time step:

according to the probability calculation result, selecting a high-frequency word with the maximum output probability from the word list as a target word output by the decoder in the t-th time step;

or ordering all the high-frequency words in the word list according to the probability calculation result in order of the output probability from large to small, and selecting the first k high-frequency words from the ordering result as target words output by the decoder in the t time step.

comparing the maximum output probability in the probability calculation result with a probability threshold value;

If the maximum output probability is greater than or equal to the probability threshold, triggering and executing the selection of the high-frequency word from the word list according to the probability calculation result, and taking the high-frequency word as a target word output by the decoder at the t-th time step;

and if the maximum output probability is smaller than the probability threshold, selecting an input word corresponding to the hidden state with the maximum attention weight from the M input words according to the attention weight of the hidden state of each input word at the t-th time step, and taking the input word as a target word output by the decoder at the t-th time step.

According to another embodiment of the present application, each unit in the information processing apparatus shown in fig. 6 may be individually or collectively combined into one or several other units, or some unit(s) thereof may be further split into a plurality of units having smaller functions, which can achieve the same operation without affecting the achievement of the technical effects of the embodiments of the present application. The above units are divided based on logic functions, and in practical applications, the functions of one unit may be implemented by a plurality of units, or the functions of a plurality of units may be implemented by one unit. In other embodiments of the present application, the information-based processing apparatus may also include other units, and in practical applications, these functions may also be realized with assistance of other units, and may be realized by cooperation of a plurality of units.

According to another embodiment of the present application, an information processing apparatus device as shown in fig. 6 may be constructed by running a computer program (including program code) capable of executing the steps involved in the respective methods as shown in fig. 2 or fig. 4 on a general-purpose computing device such as a computer including a processing element such as a Central Processing Unit (CPU), a random access storage medium (RAM), a read only storage medium (ROM), and the like, and a storage element, and an information processing method of an embodiment of the present application is implemented. The computer program may be recorded on, for example, a computer-readable recording medium, and loaded into and executed by the above-described computing device via the computer-readable recording medium.

Based on the description of the method embodiment and the device embodiment, the embodiment of the application also provides a terminal. Referring to fig. 7, the terminal includes at least a processor 701, an input interface 702, an output interface 703, and a computer storage medium 704. Wherein the processor 701, input interface 702, output interface 703, and computer storage medium 704 within the terminal may be connected by a bus or other means. A computer storage medium 704 may be stored in a memory of the terminal, the computer storage medium 704 being adapted to store a computer program comprising program instructions, the processor 701 being adapted to execute the program instructions stored by the computer storage medium 704. The processor 701, or CPU (Central Processing Unit ), is a computing core as well as a control core of the terminal, which is adapted to implement one or more instructions, in particular to load and execute one or more instructions to implement a corresponding method flow or a corresponding function.

In one embodiment, the processor 701 according to the embodiment of the present application may be configured to perform a series of processes shown in fig. 2 or fig. 4, for example, may perform the following operations: displaying a first feedback page of query information, wherein the first feedback page comprises: the information content of the feedback information corresponding to the query information in the first browsing mode; if the mode switching operation is detected, displaying a second feedback page of the query information; the second feedback page comprises: the information content of the feedback information in the second browsing mode; wherein, in the first feedback page and the second feedback page, there is a case that the information content in one feedback page is the whole information content of the feedback information, the information content in the other feedback page is the reduced information content obtained based on the whole information content, and so on.

Further, an embodiment of the present application also provides a computer storage medium (Memory), which is a Memory device in a terminal, for storing programs and data. It will be appreciated that the computer storage medium herein may include both a built-in storage medium in the terminal and an extended storage medium supported by the terminal. The computer storage medium provides a storage space that stores an operating system of the terminal. Also stored in the memory space are one or more instructions, which may be one or more computer programs (including program code), adapted to be loaded and executed by the processor. The computer storage medium herein may be a high-speed RAM memory or a non-volatile memory (non-volatile memory), such as at least one magnetic disk memory; alternatively, it may be at least one computer storage medium located remotely from the aforementioned processor. In particular implementations, one or more instructions stored in a computer-storage medium may be loaded by a processor and executed to perform the various method steps described above with respect to fig. 2 or 4.

It should be noted that according to an aspect of the present application, there is also provided a computer program product or a computer program comprising computer instructions stored in a computer readable storage medium. The processor of the terminal reads the computer instructions from the computer readable storage medium and executes the computer instructions to cause the terminal to perform the methods provided in the various alternatives of the method embodiment aspects shown in fig. 2 or fig. 4 described above.

It is also to be understood that the foregoing is merely illustrative of the present application and is not to be construed as limiting the scope of the application, which is defined by the appended claims.

Claims

1. An information processing method, characterized by comprising:

2. The method of claim 1, wherein any feedback page comprises: a first mode component corresponding to the first browsing mode and a second mode component corresponding to the second browsing mode;

3. The method of claim 1 or 2, wherein among the first feedback page and the second feedback page, a feedback page displaying the reduced information content is referred to as a reduced feedback page; the reduced feedback page further includes: the sharing component is used for sharing the simplified information content;

the method further comprises the steps of:

4. The method of claim 1 or 2, wherein among the first feedback page and the second feedback page, a feedback page displaying the reduced information content is referred to as a reduced feedback page; the simplified feedback page further comprises a browsing identifier, wherein the browsing identifier is used for indicating at least one of the following: the historical browsing amount of the simplified information content and the object identification of all or part of the objects of the simplified information content in a historical browsing manner;

The method further comprises the steps of:

5. The method of claim 1 or 2, wherein among the first feedback page and the second feedback page, a feedback page displaying the reduced information content is referred to as a reduced feedback page; the reduced feedback page further includes: the history praise amount of the simplified information content;

the method further comprises the steps of:

6. The method according to claim 1 or 2, wherein, among the first feedback page and the second feedback page, a feedback page displaying the entire information content is referred to as a reference feedback page; the total information content comprises the following components: a video identification of one or more videos related to the query information;

The method further comprises the steps of:

7. The method of claim 6, wherein the method further comprises:

8. The method of claim 1 or 2, wherein the first feedback page includes the entire information content, the method further comprising:

9. The method of claim 1, wherein the reduced information content comprises: the method comprises the steps of extracting and processing all information contents of feedback information to obtain target abstract text, wherein the target abstract text is obtained in the following way:

10. The method of claim 9, wherein for a kth sentence of the K sentences, the kth sentence is segmented into M input words, K e [1, K ], M being a positive integer; the initial abstract text corresponding to the kth sentence is determined according to target words output by the decoder in N time steps;

11. The method of claim 10, wherein said determining the attention weight of the hidden state of each input word at the t-th time step based on the attention degree of each input word by the target word generated by the decoder at the t-th time step comprises:

12. The method of claim 10, wherein the selecting a high frequency word from the vocabulary according to the probability calculation result as the target word output by the decoder at the t-th time step comprises:

13. The method of any one of claims 10-12, wherein the method further comprises:

14. An information processing apparatus, characterized by comprising:

15. A computer device comprising an input interface and an output interface, further comprising:

computer storage medium storing one or more instructions adapted to be loaded by the processor and to perform the information processing method according to any one of claims 1-13.

16. A computer storage medium storing one or more instructions adapted to be loaded by a processor and to perform the information processing method according to any one of claims 1-13.

17. A computer program product comprising a computer program which, when executed by a processor, implements the information processing method according to any one of claims 1-13.