Text information processing method, apparatus, and computer-readable storage medium

Abstract

The present disclosure relates to a text information processing method, an apparatus and a computer-readable storage medium, and to the technical field of information processing. The text information processing method comprises: using a machine learning model to determine the text type of text information and at least one piece of feature information of each segment among a plurality of segments of the text information; and, on the basis of the text type and the at least one piece of feature information of each segment, determining among the plurality of segments a key segment of the text information.

Description

Text information processing method, device and computer readable storage medium TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of information processing, and in particular, to a text information processing method, a text information processing device, a computer readable storage medium and a computer program product. BACKGROUND

[0002] From paper reading to electronic reading, the change is the medium of information consumption, and the habit of user information consumption remains unchanged. In the fragmented information consumption scene such as information flow, the patience of users is limited. For a long article, on the one hand, the content carries a large amount of information, and on the other hand, the cost of user identification and understanding is high.

[0003] In the related art, the key information in the text information is marked by artificial means.

[0004] SUMMARY

[0005] According to some embodiments of the present disclosure, a text information processing method is provided, comprising: determining a text type to which text information belongs and at least one feature information of each of a plurality of segments of the text information by using a machine learning model; and determining a key segment of the text information in the plurality of segments according to the text type and the at least one feature information of each of the plurality of segments.

[0006] In some embodiments, the at least one feature information comprises a plurality of feature information; a priority of each of the plurality of feature information is determined according to the text type; and the key segment is determined according to the priority.

[0007] In some embodiments, the plurality of feature information comprises first feature information and second feature information; the key segment is determined according to the first feature information of each of the plurality of segments in response to the priority of the first feature information being higher than the priority of the second feature information; and the key segment is determined according to the second feature information of each of the plurality of segments in response to the key segment being unable to be determined according to the first feature information of each of the plurality of segments.

[0008] In some embodiments, the key segment is interacted with the user according to the key segment in response to the user initiating an interaction request for the key segment.

[0009] In some embodiments, an interaction interface of an intelligent agent is displayed in response to the user initiating a questioning request for the key segment; and the user's questioning is replied to by the intelligent agent according to the key segment in the interaction interface.

[0010] In some embodiments, candidate question information is generated according to the key segment by using a machine learning model; in the interactive interface, the candidate question information is displayed for the user to select; and reply information is generated according to the user-selected candidate question information and the key segment by using the machine learning model.

[0011] In some embodiments, in response to the user's triggering operation on the key segment, a plurality of candidate interaction requests are displayed for the user to select, the plurality of candidate interaction requests including multiple items of a copy request, a share request, a like request, a collection request, and a question request.

[0012] In some embodiments, in response to the user's marking operation on the segment in the detail page of the text information, the segment is determined as a key segment of the text information.

[0013] In some embodiments, in the detail page of the text information, the number of users marking the segment as a key segment is displayed.

[0014] In some embodiments, in the detail page of the text information, it is displayed whether the key segment is determined according to a machine learning model or a marking operation.

[0015] In some embodiments, in response to the number of times the segment is determined as a key segment of the text information being greater than a first threshold value, and / or the number of times the segment is displayed being greater than a second threshold value, the user is fed back relevant information of the segment, the relevant information including the number of times the segment is determined as a key segment of the text information, and / or the number of times the segment is displayed.

[0016] In some embodiments, in response to the user entering the detail page of the text information, first prompt information is displayed for prompting the user to mark a key segment.

[0017] In some embodiments, in response to the marking operation, second prompt information is displayed for prompting the user to mark a viewing path of the key segment.

[0018] In some embodiments, in response to determining a key segment of the text information, the key segment is displayed in a push page of the text information.

[0019] In some embodiments, in response to the user requesting an intelligent agent to obtain a key segment of the text information, the intelligent agent pushes the key segment to the user.

[0020] In some embodiments, in the detail page of the text information, the key segment of the text information is displayed in a specified style, which is different from the display style of other segments in the text information.

[0021] According to some embodiments of the present disclosure, a text information processing apparatus is provided, comprising: a first determining unit configured to determine, by using a machine learning model, a text type to which text information belongs and at least one feature information of each of a plurality of segments of the text information; and a second determining unit configured to determine, according to the text type and the at least one feature information of each of the plurality of segments, a key segment of the text information.

[0022] In some embodiments, the at least one feature information comprises a plurality of feature information; the second determining unit is configured to determine, according to the text type, a priority of each of the plurality of feature information, and determine the key segment according to the priority.

[0023] In some embodiments, the plurality of feature information comprises first feature information and second feature information; the second determining unit is configured to determine the key segment according to the first feature information of each of the plurality of segments in response to the priority of the first feature information being higher than the priority of the second feature information, and determine the key segment according to the second feature information of each of the plurality of segments in response to the key segment being unable to be determined according to the first feature information of each of the plurality of segments.

[0024] In some embodiments, the text information processing apparatus further comprises an interaction unit configured to interact with the user according to the key segment in response to the user initiating an interaction request for the key segment.

[0025] In some embodiments, the interaction unit is configured to display an interaction interface of an intelligent agent in response to the user initiating a questioning request for the key segment, and the intelligent agent answers the questioning of the user according to the key segment in the interaction interface.

[0026] In some embodiments, the interaction unit is configured to generate candidate questioning information according to the key segment by using a machine learning model, display the candidate questioning information for the user to select in the interaction interface, and generate answer information according to the candidate questioning information selected by the user and the key segment by using the machine learning model.

[0027] In some embodiments, the interaction unit is configured to display a plurality of candidate interaction requests for the user to select in response to a triggering operation of the user on the key segment, and the plurality of candidate interaction requests comprise a plurality of items in a copy request, a share request, a like request, a collection request, and a questioning request.

[0028] In some embodiments, the second determining unit is configured to determine the segment as the key segment of the text information in response to a marking operation of the user on the segment in a detail page of the text information.

[0029] In some embodiments, the interaction unit is configured to display, in the detail page of the text information, a number of users who mark the segment as the key segment.

[0030] In some embodiments, in the detail page of the text information, it is displayed whether the key segment is determined according to the machine learning model or according to the marking operation.

[0031] In some embodiments, the interaction unit feeds back the relevant information of the segment to the user in response to that the number of times the segment is determined as the key segment of the text information is greater than a first threshold value, and / or the number of times the segment is displayed is greater than a second threshold value, and the relevant information includes the number of times the segment is determined as the key segment of the text information, and / or the number of times the segment is displayed.

[0032] In some embodiments, the interaction unit displays first prompt information in response to that the user enters the detail page of the text information, for prompting the user to mark the key segment.

[0033] In some embodiments, the interaction unit displays second prompt information in response to the marking operation, for prompting the user to mark the viewing path of the key segment.

[0034] In some embodiments, the interaction unit displays the key segment in the push page of the text information in response to that the key segment of the text information is determined.

[0035] In some embodiments, the interaction unit pushes the key segment to the user through the intelligent agent in response to that the user requests the intelligent agent to obtain the key segment of the text information.

[0036] In some embodiments, the interaction unit displays the key segment of the text information in a specified style in the detail page of the text information, and the specified style is different from the display style of other segments in the text information.

[0037] According to still some embodiments of the present disclosure, there is provided a computer readable storage medium having stored thereon a computer program, which, when executed by a processor, implements the text information processing method according to any of the above embodiments.

[0038] According to some embodiments of the present disclosure, there is also provided a computer program product comprising instructions which, when executed by a processor, cause the processor to perform the text information processing method according to any of the above embodiments.

[0039] Other features and advantages of the present disclosure will be apparent from the following detailed description of exemplary embodiments thereof, taken in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0040] The accompanying drawings, which are included to provide a further understanding of the present disclosure and constitute a part of this application, illustrate certain exemplary embodiments of the present disclosure and together with the description serve to explain the present disclosure. In the drawings:

[0041] FIG. 1 shows a flow chart of a text information processing method according to some embodiments of the present disclosure;

[0042] FIGS. 2a-2c show schematic diagrams of a key segment presentation method according to some embodiments of the present disclosure;

[0043] FIGS. 3a-3d show schematic diagrams of interaction based on key segments according to some embodiments of the present disclosure;

[0044] FIGS. 4a, 4b show schematic diagrams of prompt information presentation according to some embodiments of the present disclosure;

[0045] FIGS. 5a, 5b show flow charts of a text information processing method according to some other embodiments of the present disclosure;

[0046] FIG. 6 shows a block diagram of a text information processing apparatus according to some embodiments of the present disclosure;

[0047] FIG. 7 shows a block diagram of a text information processing apparatus according to some other embodiments of the present disclosure;

[0048] FIG. 8 shows a block diagram of a text information processing apparatus according to yet some other embodiments of the present disclosure. DETAILED DESCRIPTION

[0049] The technical solutions in the embodiments of the present disclosure will be described clearly and completely below with reference to the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, rather than all the embodiments. The following description of at least one exemplary embodiment is merely illustrative in nature and not intended to limit the present disclosure and its applications or uses in any way. Based on the embodiments in the present disclosure, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of the present disclosure.

[0050] Unless otherwise specifically noted, the relative arrangement of parts and steps, numerical expressions, and numerical values set forth in the various examples herein are not intended to limit the scope of the present disclosure. It is to be understood that the various graphical illustrations shown in the drawings are not necessarily drawn to scale. Techniques, methods, and devices known to those of ordinary skill in the relevant art can not be discussed in detail, but should be considered within the scope of the present disclosure. In all examples shown and discussed herein, any specific values should be interpreted as merely illustrative and not as a limitation on the scope of the exemplary embodiments. Thus, other example embodiments of the exemplary embodiments can have different values. It is noted that like references and letters designate like items in the following drawings, and thus once an item is defined in one drawing, it need not be further discussed in subsequent drawings.

[0051] As described above, for a long text information, a user cannot obtain key information in the first time. From the dimension of consumption value, such key information includes not only the content theme of the text information, but also some segments with incremental information and emotional value. By manually marking the key information in the text information, the efficiency and accuracy are both low, resulting in a poor effect of information pushing

[0052] The inventors of the present disclosure find that the above related technologies have the following problems: poor effect of information pushing. In view of this, the present disclosure provides a technical solution of text information processing.

[0053] In view of the above technical problems, the technical solution of the present disclosure combines the text type to which the text information belongs and the feature information of the segments in the text information, and determines the key segments of the text information by using a machine learning model, thereby improving the efficiency and accuracy of determining the key information and improving the effect of information pushing.

[0054] For example, the technical solution of the present disclosure can be implemented by the following embodiments.

[0055] FIG. 1 shows a flowchart of a text information processing method according to some embodiments of the present disclosure.

[0056] As shown in FIG. 1, in step 110, a machine learning model is used to determine the text type to which the text information belongs and at least one feature information of each segment of the multiple segments of the text information.

[0057] In step 120, according to the text type and the at least one feature information of each segment, the key segments of the text information are determined from the multiple segments.

[0058] In the above embodiments, the text type to which the text information belongs and the feature information of the segments in the text information are combined, and the key segments of the text information are determined by using a machine learning model, thereby improving the efficiency and accuracy of determining the key information and improving the effect of information pushing.

[0059] The following embodiments are used to illustrate how to show the key segments of the text information to a user.

[0060] In some embodiments, the above determined key segments can be shown on a pushing page of the text information or on an interaction interface of an intelligent body. For example, in response to determining the key segments of the text information, the key segments are displayed in the pushing page of the text information, or in response to a user requesting the intelligent body to obtain the key segments of the text information, the intelligent body pushes the key segments to the user.

[0061] The following embodiments in FIGS. 2a-2c are used to illustrate how to show the key segments of the text information to a user.

[0062] FIGS. 2a-2c show schematic diagrams of a key segment presentation method according to some embodiments of the present disclosure.

[0063] As shown in FIGS. 2a-2b, the determined key segment can be pushed to the user through push information on a push page of the text information. For example, the push information can be presented in a push information card; the text information pushed to the user can be determined through a machine learning model.

[0064] For example, the push information card can include a text presentation area, an image presentation area, a label presentation area, an information presentation area, etc.

[0065] In some embodiments, as shown in FIG. 2a, the push page of the text information presents a push information card, which includes a first text presentation area 21a, a second text presentation area 22a, an image presentation area 23, and a label presentation area 24. For example, the first text presentation area 21a is used to present the key segment of the text information; the second text presentation area 22a is used to present the title information of the text information; the image presentation area 23 is used to present the related image of the text information; the label presentation area 24 is used to present the related attribute of the key segment; and the information presentation area 25a is used to present the publishing information of the text information.

[0066] For example, the first text presentation area 21a can be pre-configured with a line limit, such as 2 lines, etc.; in response to the length of the key segment exceeding the line limit, a certain number of specified symbols (such as ellipsis, etc.) can be used to replace the last few characters in the first text presentation area 21a. In this way, the content in the first text presentation area 21a can be ensured not to exceed the line limit, and it can be indicated that the content in the first text presentation area 21a is not the complete key segment.

[0067] For example, in the first text presentation area 21a, specified symbols can be used to indicate the key segment, such as quotation marks, underlines, etc., to enhance the display of the key segment, thereby improving the effect of information pushing.

[0068] For example, the second text presentation area 22a and the image presentation area 23 can be arranged adjacent to each other; the title information presented in the second text presentation area 22a can be generated according to the text information using a machine learning model; and the related image presented in the image presentation area 23 can be obtained from the source content (such as news, blog, etc.) of the text information.

[0069] For example, the information presentation area 25a can present the publishing information of the text information, such as source information (published media information, author information, author account authentication information, etc.), publishing time; and the information presentation area 25a can be arranged adjacent to the second text presentation area 22a.

[0070] For example, the label display area 24 can display the relevant attributes of the key fragment, such as determining whether the source of the key fragment is a machine learning model or a user, how many users label the key fragment, etc.

[0071] In some embodiments, as shown in FIG. 2b, the push page of the text information is displayed with an information push card, which includes a first text display area 21b, a second text display area 22b, an image display area 23, a label display area 24, and an information display area 25b. For example, the first text display area 21b is used to display the title information of the text information; the second text display area 22b is used to display the key fragment of the text information; the image display area 23 is used to display the relevant images of the text information; the label display area 24 is used to display the relevant attributes of the key fragment; and the information display area 25b is used to display the publishing information of the text information.

[0072] For example, the title information displayed in the first text display area 21b can be generated by using a machine learning model according to the text information. The first text display area 21b can be pre-configured with a line limit, such as 2 lines, etc. In response to the length of the title information exceeding the line limit, a certain number of specified symbols (such as ellipsis, etc.) can be used to replace the last few characters in the first text display area 21b. In this way, the content in the first text display area 21b can be ensured not to exceed the line limit, and it is indicated that the content in the first text display area 21b is not the complete title information.

[0073] For example, the second text display area 22b can be pre-configured with a line limit, such as 2 lines, etc. In response to the length of the key fragment exceeding the line limit, a certain number of specified symbols (such as ellipsis, etc.) can be used to replace the last few characters in the second text display area 21b. In this way, the content in the second text display area 21b can be ensured not to exceed the line limit, and it is indicated that the content in the second text display area 21b is not the complete key fragment.

[0074] For example, in the second text display area 22b, the key fragment can be indicated by using specified symbols, such as quotation marks, underlines, etc., to enhance the display of the key fragment, thereby improving the effect of information push.

[0075] For example, the second text display area 22b and the image display area 23 can be arranged adjacent to each other; the relevant images displayed in the image display area 23 can be obtained from the source content (such as news, blog, etc.) of the text information.

[0076] For example, the label display area 24 can be arranged adjacent to the information display area 25b. The label display area 24 can display the relevant attributes of the key segment, such as determining whether the source of the key segment is a machine learning model or a user, how many users have labeled the key segment, etc.; the information display area 25b can display the publishing information of the text information, such as source information (published media information, author information, etc.), publishing time.

[0077] In some embodiments, the user can determine whether to be interested in the pushed text information or enter the detail page of the text information by interacting with the information push card in FIGS. 2a and 2b. For example, the user can jump to the detail page of the text information by interacting with the information push card; the user can also trigger an interaction panel by interacting, and select the “not interested” control or the “shield” control in the interaction panel.

[0078] For example, in response to the user selecting the “not interested” control, the text information is displayed again after a first time length (such as 1 day); in response to the user selecting the “not interested” control multiple times (such as 3 times), the text information is no longer displayed.

[0079] For example, in response to the user selecting the “shield” control, the text information is displayed again after a second time length (such as 7 days); in response to the user selecting the “shield” control multiple times (such as 3 times), the text information is no longer displayed. The first time length is shorter than the second time length.

[0080] The above illustrates that the determined key segment is displayed on the push page of the text information in combination with FIGS. 2a and 2b. The following illustrates that the determined key segment can also be pushed to the user by the intelligent agent on the interaction interface of the intelligent agent.

[0081] As shown in FIG. 2c, the interaction interface of the user and the intelligent agent can display the avatar, name and other intelligent agent related information of the current intelligent agent. In response to the user requesting the current intelligent agent to obtain the key segment of the text information, the relevant information of the text information can be displayed in the user dialogue box 21c, such as the source information (such as a website, etc.) of the text information, title information, etc. The user initiated request information can also be displayed in the user dialogue box 21c, such as “draw the most valuable sentence of this article”. In response to the user requesting the current intelligent agent to obtain the key segment of the text information, the determined key segment can be displayed in the intelligent agent dialogue box 22c, such as “the most valuable sentence: xxxxxxxx”.

[0082] As shown in FIG. 2c, the interactive control can also be displayed in the agent dialog box 22c. For example, the interactive control can include a like control, a dislike control, a play control, and the like. In response to the user selecting the like control, it can be marked that the user likes the key segment; in response to the user selecting the dislike control, it can be marked that the user is not interested in the key segment; and in response to the user selecting the play control, the content in the agent dialog box 22c can be played.

[0083] In this way, more fine-grained interaction can be performed on the key segment such as a sentence or a paragraph in the text information, so as to improve the effect of information pushing.

[0084] The above describes how to show the user the key segment of the text information by taking FIGS. 2a-2c as examples. In the following, some embodiments are described to illustrate how to determine the key segment.

[0085] In some embodiments, a machine learning model is used to determine a text type to which the text information belongs and a plurality of feature information of each of a plurality of segments of the text information; according to the text type, a priority of each of the plurality of feature information is determined; and according to the priority, the key segment is determined.

[0086] For example, the text type can include a news type, a knowledge type (such as having a high information value), and an entertainment type (such as having a high emotional type), and the like. The text information of the news type can include international news information, social news information, and the like, the text information of the knowledge type can include financial knowledge information, legal knowledge information, historical knowledge information, and the like, and the text information of the entertainment type can include game information, film and television information, food information, and the like.

[0087] For example, the feature information can include an importance feature, a typicality (or saliency) feature, and an interestingness feature, and the like. The strength of the feature information can be determined according to the content and the object (such as a person, a group, a place, and the like) involved in the text information. For example, the greater the influence of the content, the stronger the importance feature; the higher the popularity of the object involved, the stronger the typicality feature; and the more the objects involved (i.e., the higher the diversity and the greater the number of levels), the stronger the interestingness feature.

[0088] In the above embodiments, the key segment of the text information is determined by using a machine learning model in combination with the text type to which the text information belongs and the feature information of the segment in the text information, so as to improve the efficiency and accuracy of determining the key information and improve the effect of information pushing.

[0089] In some embodiments, on the basis of the above text type and the above feature information, different priorities can be configured for the feature information for different text types. The priority is used to determine the importance of the feature information in the process of determining the key segment.

[0090] For example, for the news type of text information, the priority from high to low can be set as the importance feature, the typicality feature, and the interestingness feature; for the knowledge type of text information, the priority from high to low can be set as the typicality feature, the interestingness feature, and the importance feature; for the entertainment type of text information, the priority from high to low can be set as the interestingness feature, the typicality feature, and the importance feature.

[0091] In the above embodiment, different feature information priorities are determined for different text types, which can adaptively adjust the basis for determining the key information according to the text type, thereby improving the efficiency and accuracy of determining the key information and improving the effect of information pushing.

[0092] In the following some embodiments, examples are given to determine the key segment by using different feature information priorities for different text types.

[0093] In some embodiments, the plurality of feature information includes first feature information and second feature information; in response to the priority of the first feature information being higher than the priority of the second feature information, the key segment is determined according to the first feature information of each segment.

[0094] For example, the current text information belongs to the news type, the importance feature of the first segment is higher than that of the second segment, and the typicality feature of the first segment is lower than that of the second segment; the priority corresponding to the news type is that the importance feature is higher than the typicality feature, and the first segment can be determined as the key segment.

[0095] For example, the current text information belongs to the knowledge type, the importance feature of the first segment is higher than that of the second segment, and the typicality feature of the first segment is lower than that of the second segment; the priority corresponding to the knowledge type is that the typicality feature is higher than the importance feature, and the second segment can be determined as the key segment.

[0096] In some embodiments, in response to being unable to determine the key segment according to the first feature information of each segment, the key segment is determined according to the second feature information of each segment. For example, the current text information belongs to the news type, the importance feature of the first segment is equal to that of the second segment, and the typicality feature of the first segment is lower than that of the second segment; the priority corresponding to the news type is that the importance feature is higher than the typicality feature, but the importance features of the two segments are the same, which leads to being unable to determine the key segment according to the best importance feature; in this case, the key segment can be determined according to the second highest typicality feature, i.e., the second segment is determined as the key segment.

[0097] In the above embodiments, different feature information priorities are determined for different text types, which can adaptively adjust the basis for determining key information according to the text type, thereby improving the efficiency and accuracy of determining key segments and improving the effect of information pushing.

[0098] In some embodiments, the key segment can also be determined according to the degree of coincidence of the segment with the title. For example, in response to the degree of coincidence of the segment with the title of the text information being higher than or equal to a threshold, it is determined that the segment is not a key segment; in response to the degree of coincidence of the segment with the title of the text information being lower than the threshold, it is determined whether the segment is a key segment according to the text type and the feature information of the segment.

[0099] In this way, it can be avoided to determine a segment similar to the title information as a key segment, and the information amount of the key segment relative to the title information is improved, thereby improving the efficiency of information pushing.

[0100] On the basis of determining and displaying the key segments of the text information by any of the above embodiments, the key segments can also be interacted with the user. The following embodiments will be described by way of example to illustrate the interaction between the displayed key segments and the user.

[0101] In some embodiments, the key segments of the text information are displayed in a specified style in the detail page of the text information, and the specified style is different from the display style of other segments in the text information. For example, the specified style can be underlined, bold, specified font, etc. which can remind the user to pay attention.

[0102] In some embodiments, in response to the user initiating an interaction request for the key segment, the key segment is interacted with the user. For example, the interaction can include asking questions, copying, sharing, liking, collecting, etc. for the key segment.

[0103] In this way, more granular interaction can be performed on the key segment such as a sentence or a paragraph in the text information, so as to improve the effect of information pushing.

[0104] The following embodiments in FIGS. 3a-3d will be described by way of example to illustrate the interaction between the displayed key segments and the user.

[0105] FIGS. 3a-3d show a schematic diagram of interaction based on key segments according to some embodiments of the present disclosure.

[0106] As shown in FIG. 3a, in response to the user entering the detail page of the text information, the detailed content of the text information can be displayed in the text display area 31; the previously determined key segment can be highlighted in the text display area 31 using a display style different from other segments, such as underlining.

[0107] In some embodiments, in response to the user's triggering operation on the key segment, a plurality of candidate interaction requests are displayed for the user to select, the plurality of candidate interaction requests including multiple items of a copy request, a share request, a like request, a collection request, and a question request.

[0108] For example, as shown in FIG. 3b, in response to the user's selection operation on the underlined key segment in the text display area 31, the interaction panel 32 is displayed. For example, a plurality of interaction controls, such as a copy control, a share control, and a collection control, can be displayed on the interaction panel 32. The copy control can be used to copy the key segment, the share control can be used to share the key segment, and the collection control can be used to collect the key segment.

[0109] In this way, more granular interactions can be performed on a key segment such as a sentence or a paragraph in the text information, so as to improve the effect of information pushing.

[0110] In some embodiments, in response to the user's marking operation on the segment in the detail page of the text information, the segment is determined as a key segment of the text information. For example, as shown in FIG. 3b, the underlined control can be displayed on the interaction panel 32 to determine the segment selected by the user as a key segment marked by the user.

[0111] In some embodiments, in the detail page of the text information, the number of users who mark the segment as a key segment is displayed. In the detail page of the text information, it can also be displayed whether the key segment is determined according to a machine learning model or a marking operation.

[0112] For example, the subject marking the key segment, such as "AI (artificial intelligence) marking" or "user marking", can be displayed at the key segment; and the number of users marking the key segment can also be displayed.

[0113] In some embodiments, in response to the number of times the segment is determined as a key segment of the text information being greater than a first threshold value, and / or the number of times the segment is displayed being greater than a second threshold value, the user is fed back the related information of the segment, the related information including the number of times the segment is determined as a key segment of the text information, and / or the number of times the segment is displayed.

[0114] For example, in response to the number of times a user adds AI to determine a segment as a key segment being greater than 3 times, and the number of times the segment is displayed to the user being greater than 6 times, the user who marks the segment as a key segment can be fed back "Congratulations, the content you underlined is selected as the highlight underline and will be distributed on the home page, XX users have underlined the same content as you, N users have viewed the content you underlined, and more wonderful content is expected to be underlined by you".

[0115] For example, in response to the user triggering the operation on the relevant information of the feedback, a detail page of the text information can be entered, and the detail page is positioned to the location where the key segment is located.

[0116] In this way, the willingness of the user to participate in the interaction can be stimulated, thereby improving the efficiency of information pushing.

[0117] In some embodiments, in response to the user initiating a questioning request on the key segment, an interaction interface of the intelligent agent is displayed; in the interaction interface, the intelligent agent answers the question of the user according to the key segment. For example, as shown in FIG. 3b, a question and answer control can be displayed on the interaction panel 32 to trigger the display of the interaction interface of the intelligent agent. Information prompting the user to interact with the key segment by asking questions can also be displayed in the interaction panel 32, such as “Content from AI highlight, more questions can be clicked to ask and answer”.

[0118] For example, as shown in FIG. 3c, in response to the user triggering the question and answer control on the interaction panel 32, the interaction interface 33 of the intelligent agent is displayed. The name of the intelligent agent can be displayed on the interaction interface 33 of the intelligent agent. The user can input the questioning information of the user for the current key segment on the interaction interface 33, and the questioning information can be displayed in the user dialogue box 331; based on the questioning information, the intelligent agent can generate reply information and display the reply information in the intelligent agent dialogue box 332.

[0119] From the perspective of the user, there is a situation that the user does not agree with the main idea of the whole content, but agrees with a certain segment of the content, that is, there is a demand of the user to express the viewpoint and attitude of the user for a certain segment of the article. In response to the above technical problem, in the above embodiments, the user can not only interact with the content of the whole text information, but also interact with a certain sentence or a certain paragraph of the text information in a finer granularity, such as liking, collecting, commenting, forwarding, and asking questions. In this way, the demand of the user for fine-grained interaction with the text information can be met, thereby improving the effect of information pushing.

[0120] In some embodiments, a machine learning model can also be used to generate candidate questioning information according to the key segment; in the interaction interface, the candidate questioning information is displayed for the user to select; and the machine learning model is used to generate reply information according to the candidate questioning information selected by the user and the key segment. For example, the candidate questioning information for the user to select can be displayed in the detail page of the text information or the interaction interface of the intelligent agent.

[0121] For example, as shown in FIG. 3d, in response to the user initiating a request for candidate questions for the current key segment in the smart body's interactive interface, the user dialog box 31d, the generated candidate questions "Related Question 1" and "Related Question 2" are displayed in the smart body dialog box 32d; in response to the user selecting "Related Question 1" or "Related Question 2", the corresponding reply information is displayed in the interactive interface. For example, the candidate questions for the current key segment can also be directly displayed in the interactive interface for the user to select, without the user triggering by request. In this way, the user's interaction efficiency for the key segment can be improved.

[0122] The following illustrates how to guide the user to mark the key segment in the text information through some embodiments.

[0123] In some embodiments, in response to the user entering the detail page of the text information, a first prompt information is displayed for prompting the user to mark the key segment. For example, in response to the user having been shown the key segment in the text information in the detail page of the text information before, and the user entering the detail page of the text information again, the first prompt information is displayed.

[0124] For example, the user can be prompted to mark the key segment through the embodiment in FIG. 4a.

[0125] FIG. 4a shows a schematic diagram of prompt information display according to some embodiments of the present disclosure.

[0126] As shown in FIG. 4a, in response to the user entering the detail page of the text information, a first prompt information is displayed in the guide panel 41. For example, the guide panel 41 can pop up from the bottom of the detail page. The guide panel 41 can display the first prompt information "Article Highlighting Function Online", "You can long press to select the content you are interested in to highlight, and users who read this content can see the highlight marks", and other content for prompting the user to mark the key segment. For example, the guide panel 41 can also display a close control 411 for closing the guide panel 41.

[0127] For example, the display frequency and number of times of the first prompt information can be preset for a user. For example, the first prompt information can be displayed at most once a day for a user device identifier; and / or the first prompt information can be displayed at most three times in a life cycle content. For example, the first prompt information can also not be displayed in response to the user completing the marking of the key segment.

[0128] In some embodiments, in response to the marking operation, a second prompt information is displayed for prompting the user to mark the viewing path of the key segment. For example, the user can be prompted about the viewing path of the key segment through the embodiment in FIG. 4b.

[0129] FIG. 4b shows a schematic diagram of a prompt information display according to some embodiments of the present disclosure.

[0130] As shown in FIG. 4b, in response to the user marking the key segment in the detail page of the text information, a prompt pop-up window 42 is displayed in the detail page for displaying the second prompt information. For example, the second prompt information "view path", "viewable at'my-all-my content-strikethrough'" and the like for prompting the user to view the marked key segment can be displayed on the prompt pop-up window 42. For example, a close control 421 can also be displayed in the prompt pop-up window 42 for closing the prompt pop-up window 42.

[0131] For example, the second prompt information can be displayed in response to the user marking the key segment for the first time, and the second prompt information is no longer displayed when the user marks the key segment again.

[0132] The text information processing method of the present disclosure is illustrated below by the embodiments in FIGS. 5a and 5b.

[0133] FIGS. 5a-5b show flowcharts of a text information processing method according to some embodiments of the present disclosure.

[0134] As shown in FIG. 5a, in step 510a, a machine learning model is input with text information to be processed. For example, the text information to be processed needs to satisfy a word number greater than or equal to a word number threshold (e.g., 1500 words) and / or a browsing number (e.g., an average daily browsing number) greater than a number threshold (e.g., 50 times).

[0135] In step 520a, the text type to which the text information belongs is identified by using the machine learning model. For example, the text type can include a news type, a knowledge type (e.g., with a high information value), and an entertainment type (e.g., with a high emotional type), etc. The text information of the news type can include international news information, social news information, etc., the text information of the knowledge type can include financial knowledge information, legal knowledge information, historical knowledge information, etc., and the text information of the entertainment type can include game information, film and television information, food information, etc.

[0136] In step 530a, the feature information of each segment in the text information is determined by using the machine learning model. For example, the feature information can include importance features, typicality (or saliency) features, and interestingness features, etc. The strength of the feature information can be determined according to the content and objects (e.g., entities such as persons, groups, places, etc.) involved in the text information. For example, the greater the influence of the content, the stronger the importance features, the higher the notability of the objects involved, the stronger the typicality features, and the more objects involved (i.e., the higher the diversity and the greater the number of levels), the stronger the interestingness features.

[0137] In step 540, according to the text type and the characteristic information of each segment, the key segment of the text information is output. For example, for the text information under a content identifier, at most one key segment is output.

[0138] As shown in FIG. 5b, in the supply and distribution stage of the key segment, steps 510b and 520b are included. In step 510b, the key segment determined through the embodiment in FIG. 5a is pushed to the user through the information push card on the push page of the text information. For example, the information push card can be distributed in the information stream.

[0139] In step 520b, in response to the interactive operation of the user on the information push card, the detail page of the text information is displayed. For example, the detail page of the information can also be directly displayed to the user without the information push card.

[0140] In the consumption stage of the key segment, steps 530b and 540b are included. In step 530b, the user consumes the key segment in the detail page. For example, the key segment can be highlighted in the detail page by using a specified display style (such as marked with a dashed line, etc.); and the number of times the key segment is marked can also be displayed at the end of the key segment. For example, in response to the key segment being marked by the machine learning model, 1 person is displayed; and in response to the key segment being marked not only by the machine learning model but also by M users, M+1 people are displayed.

[0141] In the above embodiment, the key segment with importance, typicality, and interestingness in the text information is determined by the machine learning model, and is highlighted and displayed in a specified display style; and the user can first notice and consume the key information in the text information during browsing the longer text information.

[0142] In step 530b, the user interacts with the key segment. For example, the user clicks the key segment to trigger an interaction panel (such as the interaction panel 32 in FIG. 3b). For example, the interaction includes copying, sharing, liking, collecting, asking questions, etc.

[0143] In the above embodiment, the key segment is interacted with the artificial intelligence to ask questions, which not only improves the consumption effect of the text information, but also improves the incremental information of the text information, thereby improving the information push effect.

[0144] In the stage of guiding the marking of the key segment, step 550b is included. In step 550b, the user is guided to mark the key segment in the text information. For example, in response to the key segment of other text information having been displayed to the user and the user entering the detail page of the current text information, a guide panel is displayed to display first prompt information to prompt the user to mark the key segment.

[0145] In the stage of using the key segments, step 560b is included. In step 560b, the user is guided to imitate the marking of the machine learning model by showing the user the key segments marked by the machine learning model. For example, the complete marking process can be shown through animation to guide the user to complete the marking.

[0146] In the above embodiment, by showing the user the key segments determined by the machine learning model as an example, on the one hand, the technical problem that the user does not know how to mark the key segments is solved; on the other hand, the technical problem that the supply of the key segment marking function is less in the cold start stage is solved, and the user can directly interact with the key segments determined by the machine learning model in a more detailed granularity.

[0147] The stage of consuming the key segments can also include step 570b. In step 570b, the user is fed back the selected key segments. For example, in response to the number of times the segment is determined to be a key segment of text information being greater than a first threshold value, and / or the number of times the segment is displayed being greater than a second threshold value, the user is fed back the related information of the segment, and the related information includes the number of times the segment is determined to be a key segment of text information, and / or the number of times the segment is displayed.

[0148] In the above embodiment, the key segments are automatically determined by the machine learning model, and the user can complete the labeling of the key segments through browsing, clicking, and the like, thereby reducing the cost of screening information, stimulating the user's willingness to mark the key segments and repeatedly access, and thereby improving the effect of information pushing.

[0149] In this way, through the labeling of the key segments in the text information, the repeated access demand of the user can be met. That is, on the one hand, the demand of the user for reading traces can be met, and on the other hand, the demand of the user for repeated access for subsequent searching can be met.

[0150] FIG. 6 shows a block diagram of a text information processing apparatus according to some embodiments of the present disclosure.

[0151] As shown in FIG. 6, the text information processing apparatus 6 includes a first determination unit 61 configured to determine, by using a machine learning model, a text type to which text information belongs and at least one feature information of each of a plurality of segments of the text information, and a second determination unit 62 configured to determine, according to the text type and the at least one feature information of each of the segments, a key segment of the text information from the plurality of segments.

[0152] In some embodiments, the at least one feature information includes a plurality of feature information, and the second determination unit 62 determines a priority of each of the plurality of feature information according to the text type, and determines the key segment according to the priority.

[0153] In some embodiments, the plurality of feature information includes first feature information and second feature information; the second determining unit 62 determines the key segment according to the first feature information of each segment in response to the priority of the first feature information being higher than the priority of the second feature information, and determines the key segment according to the second feature information of each segment in response to the key segment being unable to be determined according to the first feature information of each segment.

[0154] In some embodiments, the text information processing apparatus 6 further includes an interaction unit 63 configured to interact with the user according to the key segment in response to the user initiating an interaction request for the key segment.

[0155] In some embodiments, the interaction unit 63 displays an interaction interface of the intelligent agent in response to the user initiating a questioning request for the key segment, and the intelligent agent answers the questioning of the user according to the key segment in the interaction interface.

[0156] In some embodiments, the interaction unit 63 generates candidate questioning information according to the key segment by using a machine learning model, displays the candidate questioning information for the user to select in the interaction interface, and generates answer information according to the candidate questioning information selected by the user and the key segment by using the machine learning model.

[0157] In some embodiments, the interaction unit 63 displays a plurality of candidate interaction requests for the user to select in response to the user triggering the key segment, and the plurality of candidate interaction requests include multiple items in the copying request, the sharing request, the like request, the collection request, and the questioning request.

[0158] In some embodiments, the second determining unit 62 determines the segment as the key segment of the text information in response to the user marking the segment in the detail page of the text information.

[0159] In some embodiments, the interaction unit 63 displays the number of users marking the segment as the key segment in the detail page of the text information.

[0160] In some embodiments, it is displayed in the detail page of the text information whether the key segment is determined according to the machine learning model or the marking operation.

[0161] In some embodiments, the interaction unit 63 feeds back the related information of the segment to the user in response to the number of times the segment being determined as the key segment of the text information being greater than a first threshold value, and / or the number of times the segment being displayed being greater than a second threshold value, and the related information includes the number of times the segment being determined as the key segment of the text information, and / or the number of times the segment being displayed.

[0162] In some embodiments, the interaction unit 63 displays first prompt information for prompting the user to mark the key segment in response to the user entering the detail page of the text information.

[0163] In some embodiments, the interaction unit 63 displays second prompt information for prompting the user to view the marked key segment in response to the marking operation.

[0164] In some embodiments, the interaction unit 63 displays the key segment in the push page of the text information in response to determining the key segment of the text information.

[0165] In some embodiments, the interaction unit 63 pushes the key segment to the user through the intelligent agent in response to the user requesting the intelligent agent to obtain the key segment of the text information.

[0166] In some embodiments, the interaction unit 63 displays the key segment of the text information in a specified style in the detail page of the text information, and the specified style is different from the display style of other segments in the text information.

[0167] FIG. 7 shows a block diagram of a text information processing apparatus according to some other embodiments of the present disclosure.

[0168] As shown in FIG. 7, the text information processing apparatus 7 of this embodiment includes a memory 71 and a processor 72 coupled to the memory 71, and the processor 72 is configured to execute the text information processing method in any one of the embodiments of the present disclosure based on instructions stored in the memory 71.

[0169] The memory 71 may, for example, include a system memory, a fixed non-volatile storage medium, etc. The system memory may, for example, store an operating system, an application program, a Boot Loader, a database, and other programs, etc.

[0170] FIG. 8 shows a block diagram of a text information processing apparatus according to yet some other embodiments of the present disclosure.

[0171] As shown in FIG. 8, the text information processing apparatus 8 of this embodiment includes a memory 810 and a processor 820 coupled to the memory 810, and the processor 820 is configured to execute the method in any one of the foregoing embodiments based on instructions stored in the memory 810.

[0172] The memory 810 may, for example, include a system memory, a fixed non-volatile storage medium, etc. The system memory may, for example, store an operating system, an application program, a Boot Loader, and other programs, etc.

[0173] The text information processing apparatus 8 can further include an input / output interface 830, a network interface 840, a storage interface 850, and the like. These interfaces 830, 840, 850, and the memory 810 and the processor 820 can be connected, for example, through a bus 860. Among them, the input / output interface 630 provides a connection interface for display, mouse, keyboard, touch screen, microphone, speaker, and the like input / output devices. The network interface 640 provides a connection interface for various networking devices. The storage interface 850 provides a connection interface for external storage devices such as SD card and U disk.

[0174] Those skilled in the art will appreciate that embodiments of the present disclosure can be provided as methods, systems, or computer program products. Accordingly, the present disclosure can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present disclosure can take the form of a computer program product implemented on one or more computer-usable non-transitory storage media (including, but not limited to, disk storage, CD-ROMs, optical storage media, and the like) having computer-usable program code embodied therein.

[0175] So far, the text information processing method, the text information processing apparatus, the computer readable storage medium, and the computer program product according to the present disclosure have been described in detail. In order to avoid obscuring the concept of the present disclosure, some details known in the art are not described. Those skilled in the art can fully understand how to implement the technical solutions disclosed herein according to the above description.

[0176] The methods and systems of the present disclosure can be implemented in a number of ways. For example, the methods and systems of the present disclosure can be implemented using software, hardware, firmware, or any combination of software, hardware, and firmware. The above described order of steps for the methods is merely for illustration, and the steps of the methods of the present disclosure are not limited to the above specifically described order, unless otherwise specifically stated. Furthermore, in some embodiments, the present disclosure can also be implemented as programs recorded in recording media, which include machine-readable instructions for implementing the methods according to the present disclosure. Thus, the present disclosure also covers recording media storing programs for executing the methods according to the present disclosure.

[0177] Although some specific embodiments of the present disclosure have been described in detail through examples, those skilled in the art should understand that the above examples are merely for illustration, and are not intended to limit the scope of the present disclosure. Those skilled in the art should understand that the above embodiments can be modified without departing from the scope and spirit of the present disclosure. The scope of the present disclosure is defined by the appended claims.

Claims

1. A method for processing text information, comprising: determining, by using a machine learning model, a text type to which the text information belongs and at least one feature information of each of a plurality of segments of the text information; determining, according to the text type and the at least one feature information of each of the segments, a key segment of the text information from the plurality of segments.

2. The text information processing method according to claim 1, wherein the at least one feature information comprises a plurality of feature information, the determining, according to the text type and the at least one feature information of each of the segments, the key segment of the text information from the plurality of segments comprises: determining, according to the text type, a priority of each of the plurality of feature information; determining, according to the priority, the key segment.

3. The text information processing method according to claim 2, wherein the plurality of feature information comprises a first feature information and a second feature information, the determining, according to the priority, the key segment comprises: determining, according to the first feature information of each of the segments, the key segment in response to a priority of the first feature information being higher than a priority of the second feature information; determining, according to the second feature information of each of the segments, the key segment in response to the key segment being unable to be determined according to the first feature information of each of the segments. 4.The method of claim 1, further comprising: interacting, according to the key segment, with a user in response to the user initiating an interaction request for the key segment.

5. The text information processing method according to claim 4, wherein the interacting, according to the key segment, with the user in response to the user initiating the interaction request for the key segment comprises: displaying, in response to the user initiating a question request for the key segment, an interaction interface of an intelligent agent; answering, by the intelligent agent, a question of the user in the interaction interface according to the key segment.

6. The text information processing method according to claim 5, wherein the answering, by the intelligent agent, the question of the user in the interaction interface according to the key segment comprises: generating, by using a machine learning model, candidate question information according to the key segment; displaying, in the interaction interface, the candidate question information for selection by the user; generating, by using the machine learning model, answer information according to the candidate question information selected by the user and the key segment.

7. The text information processing method according to claim 4, wherein the interacting, according to the key segment, with the user in response to the user initiating the interaction request for the key segment comprises: displaying, in response to a trigger operation of the user on the key segment, a plurality of candidate interaction requests for selection by the user, the plurality of candidate interaction requests comprising a plurality of items in a copy request, a share request, a like request, a collection request, and a question request. 8.The method of any one of claims 1-7, further comprising: determining, in response to a marking operation of a user on a segment in a detail page of the text information, the segment as a key segment of the text information. 9.The method of claim 8, further comprising: displaying, in the detail page of the text information, a number of users marking the segment as the key segment; and / or displaying, in the detail page of the text information, whether the key segment is determined according to the machine learning model or according to the marking operation. ​ 10. The text information processing method according to claim 8, wherein The determining the segment as the key segment of the text information comprises: in response to the number of times that the segment is determined as the key segment of the text information being greater than a first threshold value, and / or the number of times that the segment is displayed being greater than a second threshold value, feeding back relevant information of the segment to the user, the relevant information comprising the number of times that the segment is determined as the key segment of the text information, and / or the number of times that the segment is displayed.

11. The text information processing method according to claim 8, further comprising: in response to the user entering a detail page of the text information, displaying first prompt information for prompting the user to mark a key segment; and / or in response to the marking operation, displaying second prompt information for prompting a viewing path of the key segment marked by the user.

12. The text information processing method according to any one of claims 1-11, further comprising: in response to determining a key segment of the text information, displaying the key segment in a push page of the text information; and / or in response to a user requesting an intelligent agent to obtain a key segment of the text information, pushing the key segment to the user by the intelligent agent.

13. The text information processing method according to any one of claims 1-12, further comprising: in a detail page of the text information, displaying a key segment of the text information in a specified style, the specified style being different from a display style of other segments in the text information.

14. A text information processing apparatus, comprising: a first determining unit configured to determine, by using a machine learning model, a text type to which a text information belongs and at least one feature information of each of a plurality of segments of the text information; a second determining unit configured to determine, according to the text type and the at least one feature information of each of the segments, a key segment of the text information from the plurality of segments.

15. A text information processing apparatus, comprising: a memory; and a processor coupled to the memory, the processor being configured to perform the text information processing method according to any one of claims 1-13 based on instructions stored in the memory.

16. A computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the text information processing method according to any one of claims 1-13.

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

Images