Creative text intelligent processing method and device based on artificial intelligence

Abstract

The invention relates to the technical field of information processing, and discloses a creative text intelligent processing method and device based on artificial intelligence, and the method comprises the steps: responding to a template summary viewing operation, and displaying a summary page corresponding to the template summary viewing operation; displaying template information corresponding to the at least one summarized text template on a summarized page; each summarized text template meets an arrangement condition corresponding to the template summarized viewing operation and also meets a text interception condition; the template information comprises template content and further comprises theme types and / or template associated information; the template associated information comprises at least one of associated text information, first associated creative information, first associated product information and first popularity information. It can be seen that the creative text analysis result can be efficiently displayed while the processing efficiency of the creative text content is improved by efficiently summarizing the creative text templates, and the decision-making efficiency and the decision-making accuracy of the user for creative text content selection can be improved.

Description

Technical Field

[0001] This invention relates to the field of information processing technology, and in particular to a method and apparatus for intelligent processing of creative text based on artificial intelligence. Background Technology

[0002] With the rapid development of social media e-commerce, short videos have become a core channel for brand promotion, product sales, and live streaming traffic generation. On social media platforms, the dialogue in the first 3 seconds of a short video is crucial for attracting viewers' attention, conveying brand information, and increasing video completion rates. However, faced with a massive amount of short video creative content, users (such as advertisers and operations personnel) struggle to efficiently identify current trends, leading to inefficient decision-making during the creative process.

[0003] To address these challenges, current technologies still focus on analyzing single or small amounts of dialogue from short videos, manually selecting popular videos, and then using techniques such as manual transcription and subtitle extraction to extract the dialogue text from the video content. However, this approach is time-consuming and labor-intensive, unable to handle large-scale data processing, and struggles to efficiently identify and categorize the text content of short video content. Consequently, it is difficult to efficiently present the analysis results to users and provide effective assistance for their decision-making.

[0004] Therefore, it is particularly important to propose a technical solution that can improve the processing efficiency of creative text content while efficiently displaying the results of creative text analysis, thereby improving the efficiency and accuracy of users' decision-making in selecting creative text content during the creative creation process. Summary of the Invention

[0005] This invention provides an intelligent processing method and apparatus for creative text based on artificial intelligence, which can improve the processing efficiency of creative text content and efficiently display the analysis results of creative text, thereby helping to improve the efficiency and accuracy of users' decision-making in selecting creative text content during the creative creation process.

[0006] To address the aforementioned technical problems, the first aspect of this invention discloses an intelligent processing method for creative text based on artificial intelligence, the method comprising: In response to the template summary view operation, the summary page corresponding to the template summary view operation is displayed; The summary page displays template information corresponding to at least one summary text template. Each of the summarized text templates satisfies the sorting conditions corresponding to the template summary viewing operation, and each of the summarized text templates also satisfies the pre-set text extraction conditions; the template information includes template content, and also includes theme type and / or template association information; wherein, the template association information includes at least one of associated text information, first associated creative information, first associated product information, and first popularity information.

[0007] As an optional implementation, in the first aspect of the present invention, the method further includes: In response to a text template viewing operation on the summary page, the target text template corresponding to the text template viewing operation is determined from all the summary text templates, and a preview page corresponding to the target text template is displayed based on preset preview display parameters; The preview page displays the associated preview information corresponding to the target text template; The preset preview display parameters include at least the display method of the preview page; the associated preview information includes at least one of core associated data, text preview information of at least one associated text, and creative preview information of at least one associated creative; wherein each associated text and each associated creative is associated with the target text template.

[0008] As an optional implementation, in the first aspect of the present invention, the method further includes: In response to a details viewing operation on the preview page, the target details page corresponding to the target text template is displayed; Based on preset detail display parameters, the data detail information and / or the text detail information corresponding to the target text template are displayed on the target detail page; The data details include the core related data, creative placement trends, and category distribution; wherein, the creative placement trend indicates the trend of changes in the number of placements of related creatives corresponding to the target text template; and the category distribution indicates the product category distribution of related products corresponding to the target text template. The text details information includes at least one of the associated texts; wherein the text details information includes at least one of the following: text content, second associated creative information, second associated product information, and second popularity information.

[0009] As an optional implementation, in the first aspect of the present invention, each of the summary text templates is a text template, and the text template is generated in the following manner: Perform a text recognition operation on each newly added platform creative to obtain a text recognition result corresponding to each platform creative; the text recognition result includes the recognized text and the timestamp information corresponding to the recognized text. For each text recognition result, a target text segment satisfying the text extraction conditions is extracted from the recognized text of the text recognition result; wherein, the text extraction conditions include start and end time constraints and text integrity conditions; For all the target text segments, a text merging operation is performed to obtain at least one text set; each text set includes at least two similar texts, and all the similar texts included in each text set satisfy a preset similarity condition; For each text set, a text template corresponding to that text set is determined based on preset template conditions.

[0010] As an optional implementation, in the first aspect of the invention, before performing a text merging operation on all the target fragment texts to obtain at least one text set, the method further includes: Based on a pre-trained topic recognition model, the topic type corresponding to each target text segment is identified; Obtain the original additional associated data corresponding to each of the platform's creative ideas; For each platform creative, the topic type corresponding to the platform creative and the original additional related data corresponding to the platform creative are synchronized to the determined search analysis engine.

[0011] As an optional implementation, in the first aspect of the invention, performing a text merging operation on all the target fragment texts to obtain at least one text set includes: From all the target fragment texts, determine the current identified text to be analyzed; The similarity query algorithm of the search analysis engine is invoked to filter out at least one candidate similar text that matches the currently identified text from all historical synchronized texts corresponding to the search analysis engine; wherein the similarity between each candidate similar text and the currently identified text is greater than or equal to a preset similarity threshold. Calculate the average relative edit distance between each candidate similar text and the currently identified text; From all the candidate similar texts, target similar texts with an average relative edit distance less than or equal to a preset distance threshold are selected, and the currently identified text and all the target similar texts corresponding to the currently identified text are determined as a text set; wherein, each similar text included in the text set is either the currently identified text or the target similar text; The operation of determining the current identified text to be analyzed from all the target fragment texts is repeated until each target fragment text has a corresponding text set.

[0012] As an optional implementation, in the first aspect of the present invention, for each text set, determining the text template corresponding to that text set according to preset template conditions includes: Determine the recognition time corresponding to each of the similar texts; The text with the earliest recognition time among all the similar texts in the text set is determined as the text template corresponding to the text set. In this text set, all similar texts are associated texts of the text template.

[0013] A second aspect of this invention discloses an intelligent creative text processing device based on artificial intelligence, the device comprising: The display module is used to respond to the template summary viewing operation and display the summary page corresponding to the template summary viewing operation. The display module is also used to display template information corresponding to at least one summary text template on the summary page; Each of the summarized text templates satisfies the sorting conditions corresponding to the template summary viewing operation, and each of the summarized text templates also satisfies the pre-set text extraction conditions; the template information includes template content, and also includes theme type and / or template association information; wherein, the template association information includes at least one of associated text information, first associated creative information, first associated product information, and first popularity information.

[0014] As an optional implementation, in a second aspect of the invention, the apparatus further includes: The determination module is configured to, in response to a text template viewing operation on the summary page, determine the target text template corresponding to the text template viewing operation from all the summary text templates; The display module is also used to display a preview page corresponding to the target text template based on preset preview display parameters; The display module is also used to display the associated preview information corresponding to the target text template on the preview page; The preset preview display parameters include at least the display method of the preview page; the associated preview information includes at least one of core associated data, text preview information of at least one associated text, and creative preview information of at least one associated creative; wherein each associated text and each associated creative is associated with the target text template.

[0015] As an optional implementation, in a second aspect of the invention, the display module is further configured to display a target details page corresponding to the target text template in response to a details viewing operation on the preview page; The display module is also used to display data details information and / or text details information corresponding to the target text template on the target details page based on preset details display parameters; The data details include the core related data, creative placement trends, and category distribution; wherein, the creative placement trend indicates the trend of changes in the number of placements of related creatives corresponding to the target text template; and the category distribution indicates the product category distribution of related products corresponding to the target text template. The text details information includes at least one of the associated texts; wherein the text details information includes at least one of the following: text content, second associated creative information, second associated product information, and second popularity information.

[0016] As an optional implementation, in the second aspect of the invention, each of the summary text templates is a text template, and the text template is generated in the following manner: Perform a text recognition operation on each newly added platform creative to obtain a text recognition result corresponding to each platform creative; the text recognition result includes the recognized text and the timestamp information corresponding to the recognized text. For each text recognition result, a target text segment satisfying the text extraction conditions is extracted from the recognized text of the text recognition result; wherein, the text extraction conditions include start and end time constraints and text integrity conditions; For all the target text segments, a text merging operation is performed to obtain at least one text set; each text set includes at least two similar texts, and all the similar texts included in each text set satisfy a preset similarity condition; For each text set, a text template corresponding to that text set is determined based on preset template conditions.

[0017] As an optional implementation, in a second aspect of the present invention, the method for generating the text template further includes: Before performing a text merging operation on all the target text fragments to obtain at least one text set, the topic type corresponding to each target text fragment is identified based on a pre-trained topic recognition model. Obtain the original additional associated data corresponding to each of the platform's creative ideas; For each platform creative, the topic type corresponding to the platform creative and the original additional related data corresponding to the platform creative are synchronized to the determined search analysis engine.

[0018] As an optional implementation, in the second aspect of the invention, the specific method of performing a text merging operation on all the target text fragments to obtain at least one text set includes: From all the target fragment texts, determine the current identified text to be analyzed; The similarity query algorithm of the search analysis engine is invoked to filter out at least one candidate similar text that matches the currently identified text from all historical synchronized texts corresponding to the search analysis engine; wherein the similarity between each candidate similar text and the currently identified text is greater than or equal to a preset similarity threshold. Calculate the average relative edit distance between each candidate similar text and the currently identified text; From all the candidate similar texts, target similar texts with an average relative edit distance less than or equal to a preset distance threshold are selected, and the currently identified text and all the target similar texts corresponding to the currently identified text are determined as a text set; wherein, each similar text included in the text set is either the currently identified text or the target similar text; The operation of determining the current identified text to be analyzed from all the target fragment texts is repeated until each target fragment text has a corresponding text set.

[0019] As an optional implementation, in a second aspect of the present invention, for each text set, the specific method of determining the text template corresponding to that text set according to preset template conditions includes: Determine the recognition time corresponding to each of the similar texts; The text with the earliest recognition time among all the similar texts in the text set is determined as the text template corresponding to the text set. In this text set, all similar texts are associated texts of the text template.

[0020] A third aspect of this invention discloses another intelligent creative text processing device based on artificial intelligence, the device comprising: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute some or all of the steps of the AI-based intelligent creative text processing method disclosed in the first aspect of the present invention.

[0021] The fourth aspect of the present invention discloses a computer storage medium storing computer instructions, which, when invoked, are used to execute some or all of the steps of the AI-based intelligent creative text processing method disclosed in the first aspect of the present invention.

[0022] Compared with the prior art, the present invention has the following beneficial effects: In this invention, in response to a template summary viewing operation, a summary page corresponding to the template summary viewing operation is displayed; the summary page displays template information corresponding to at least one summary text template; wherein each summary text template satisfies the arrangement conditions corresponding to the template summary viewing operation, and each summary text template also satisfies pre-set text extraction conditions; the template information includes template content, and also includes theme type and / or template association information; wherein the template association information includes at least one of associated text information, first associated creative information, first associated product information, and first popularity information. Therefore, implementing this invention can respond to a template summary viewing operation, display a summary page corresponding to the template summary viewing operation, and display template information corresponding to at least one summary text template on the summary page. It can efficiently filter and aggregate various text templates used in a large number of creative works based on arrangement conditions, thereby improving the processing efficiency of creative text content while efficiently displaying creative text analysis results. This helps users efficiently browse the template content of various creative text templates, as well as related popularity information and associated data, thus improving the efficiency and accuracy of users' decision-making in selecting creative text content during the creative creation process. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart illustrating an intelligent creative text processing method based on artificial intelligence disclosed in an embodiment of the present invention. Figure 2 This is a flowchart illustrating another method for intelligent processing of creative text based on artificial intelligence, as disclosed in an embodiment of the present invention. Figure 3 This is a schematic diagram of an interface for an intelligent creative text processing method based on artificial intelligence disclosed in an embodiment of the present invention; Figure 4This is a schematic diagram of an interface for another AI-based intelligent text processing method disclosed in an embodiment of the present invention. Figure 5 This is a schematic diagram of the interface of another intelligent creative text processing method based on artificial intelligence disclosed in an embodiment of the present invention; Figure 6 This is a schematic diagram of the interface of another intelligent creative text processing method based on artificial intelligence disclosed in an embodiment of the present invention; Figure 7 This is a schematic diagram of the structure of an intelligent creative text processing device based on artificial intelligence disclosed in an embodiment of the present invention; Figure 8 This is a schematic diagram of another intelligent creative text processing device based on artificial intelligence disclosed in an embodiment of the present invention; Figure 9 This is a schematic diagram of the structure of another intelligent creative text processing device based on artificial intelligence disclosed in an embodiment of the present invention. Detailed Implementation

[0025] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this invention are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, apparatus, product, or end that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or ends.

[0027] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0028] This invention discloses an intelligent creative text processing method and apparatus based on artificial intelligence. Responding to a template summary viewing operation, it displays a summary page corresponding to the operation, showing template information for at least one summarized text template. Based on sorting conditions, it can efficiently filter and aggregate various text templates used in a large number of creative works. This improves the processing efficiency of creative text content while efficiently displaying the creative text analysis results. It helps users efficiently browse the template content of various creative text templates, as well as related popularity information and associated data, thereby improving the efficiency and accuracy of users' decision-making in selecting creative text content during the creative creation process. Detailed descriptions follow.

[0029] Example 1 Please see Figure 1 , Figure 1 This is a flowchart illustrating an intelligent creative text processing method based on artificial intelligence, as disclosed in an embodiment of the present invention. Figure 1 The described AI-based intelligent creative text processing method can be applied to an AI-based intelligent creative text processing device. This device may include one of a smart terminal, a smart device, a smart system, and a server. The server may be a local server or a cloud server; this embodiment of the invention does not limit the specific application. Figure 1 As shown, this AI-based intelligent processing method for creative text may include the following operations: 101. In response to the template summary view operation, display the summary page corresponding to the template summary view operation.

[0030] In this embodiment of the invention, optionally, the template summary viewing operation may include a summary page jump operation and / or a template summary filtering operation, which is not limited in this embodiment of the invention; further optionally, the summary page jump operation may be an operation of jumping from other pages to the summary page, more specifically, it may be an operation of jumping to the summary page by entering a page address, which is not limited in this embodiment of the invention; further optionally, the template summary filtering operation may be an operation of filtering at least one summary text template that meets the determined filtering conditions from all platform text templates.

[0031] 102. Display template information corresponding to at least one summary text template on the summary page.

[0032] In this embodiment of the invention, each summary text template is a text template; optionally, the text template may include a dialogue template and / or a copywriting template, which is not limited in this embodiment of the invention; further optionally, different text templates may be summarized in different creative text processing scenarios; for example, when the creative text processing scenario is a processing scenario for the dialogue mentioned in the creative, the summary text template may be a dialogue template (e.g., dialogue routines), which is not limited in this embodiment of the invention.

[0033] In this embodiment of the invention, optionally, when all the summary text templates are dialogue templates, the summary page can be a dialogue routine ranking page, but this embodiment of the invention does not impose any limitations.

[0034] In this embodiment of the invention, each summary text template satisfies the arrangement conditions corresponding to the template summary viewing operation, and each summary text template also satisfies pre-set text extraction conditions. Optionally, the arrangement conditions may include sorting conditions and / or filtering conditions; optionally, the filtering conditions corresponding to the template summary viewing operation may include filtering rules based on at least one of the following: product category, advertising platform, creative type, theme type, and time range. This embodiment of the invention does not impose any limitations. For example, product categories may include, but are not limited to, one or more of the following: apparel and footwear, cosmetics and skincare, personal care and body care, daily necessities, home furnishings and decoration, food and beverages, cultural and entertainment products, digital home appliances, sports and outdoor products, automobiles, maternal and infant products, flowers and gardening, medical and health products, and watches and jewelry. This embodiment of the invention does not impose any limitations. Optionally, the advertising platform may be a traffic platform used for advertising creatives. Optionally, the creative type may specifically include creative promotion types. For example, creative promotion types may include, but are not limited to, product advertising creative types and / or live streaming traffic-driving advertising creative types. This embodiment of the invention does not impose any limitations. The theme type summarizes the theme of the content described by the text template or the role of the content described by the text template in a creative (material). For example, if the text template is a dialogue template, the theme type can be a dialogue pattern type, and the dialogue pattern type can include one or more of the following: price benefit, scarcity, target audience, suspense resolution, direct statement of pain points / effects, product display, and emotional shaping. This embodiment of the invention does not limit this. The time range can, for example, include but is not limited to one of the following: the last 1 day, the last 2 days, the last 7 days, the last 30 days, the last 60 days, and "newly listed," and can also include other custom time ranges. This embodiment of the invention does not limit this. In this embodiment of the invention, optionally, the filtering conditions can be determined by the selection status of at least one page component corresponding to any of the above filtering criteria. For example, the page component may include, but is not limited to, at least one button component. That is, whether the user selects the filtering criterion can be determined by whether the button component is selected. This embodiment of the invention does not limit this.

[0035] In this embodiment of the invention, optionally, the template information includes template content, and further includes topic type and / or template association information; wherein, the template association information includes at least one of associated text information, first associated creative information, first associated product information, and first popularity information. The template content is the specific text of the summary text template. Further optionally, the associated text information may include the number of associated texts and / or the text content of each associated text, which is not limited in this embodiment of the invention; further, each associated text corresponding to the text template is similar to the text template. Further optionally, the first associated creative information may include the number of first associated creatives, specifically, the number of first associated creatives of at least one creative type associated with the text template. Further optionally, the first associated product information may include the number of first associated products, specifically, the number of first associated products associated with the text template, wherein the first associated products may be products introduced through associated creatives associated with the text template. Optionally, the first popularity information may include one or more of the following: the estimated exposure value, the estimated growth rate of exposure, and the number of days of placement for the text template. This embodiment of the invention does not limit this information. The number of days of placement may be used to represent the number of days of placement of the associated creative corresponding to the text template or the number of days from the date the text template first appeared to today. Furthermore, the method for calculating the exposure value of each text template may be preset by the platform or the user. This embodiment of the invention does not limit this method.

[0036] In this embodiment of the invention, it should be noted that each associated creative is one of the platform creatives collected / linked by the platform; furthermore, each platform creative may belong to the materials that can be viewed by users and / or used and / or referenced in content creation, and the data type of each platform creative may include at least one of text type, image type, audio type and video type, which is not limited in this embodiment of the invention.

[0037] In this embodiment of the invention, the exemplary interface display effect of the summary page can be referred to... Figure 3 , Figure 3 This is a schematic diagram of an interface for an artificial intelligence-based intelligent text processing method disclosed in an embodiment of the present invention; wherein, as shown... Figure 3As shown, the summary page can be a list of the most effective lines in the "golden 3 seconds" dialogue, and the summary page contains at least one line template (i.e., Figure 3 The template information corresponding to the "dialogue patterns" in the text can specifically include template content (i.e., Figure 3 "Dialogue patterns" and related text information (such as...) Figure 3 The "number of related lines", the specific content of each related text, and the first related creative information (such as...) Figure 3 The number of live streaming promotion materials), and the first related product information (such as...) Figure 3 The number of related products), and the top popularity information (such as...) Figure 3 The "exposure forecast", "exposure forecast growth rate", and "number of days" in the data are not limited in this embodiment of the invention; furthermore, the summary page may also include multiple button components for determining different filtering criteria, such as... Figure 3 As shown, the filtering criteria may include category filtering (i.e., the product categories mentioned above), traffic platform, promotion type (i.e., the creative promotion type mentioned above), strategy type (i.e., the theme type mentioned above), and advanced filtering (including the time range mentioned above). This embodiment of the invention does not limit these criteria. Furthermore, the summary page may also include the results of dialogue hot word analysis. This embodiment of the invention does not limit these criteria.

[0038] As can be seen, the method described in the embodiments of the present invention can respond to the template summary viewing operation, display the summary page corresponding to the template summary viewing operation, and display the template information corresponding to at least one summary text template on the summary page. It can efficiently filter and aggregate various text templates used in a large number of creative works based on the arrangement conditions, thereby improving the processing efficiency of creative text content and efficiently displaying the creative text analysis results. This helps users to efficiently browse the template content of various creative text templates as well as related popularity information and associated data, which in turn helps improve the efficiency and accuracy of users' decision-making in selecting creative text content during the creative creation process.

[0039] In an optional embodiment, the text template can be generated in the following way: Perform text recognition operation on each newly added platform creative to obtain the text recognition result corresponding to each platform creative; the text recognition result includes the recognized text and the timestamp information corresponding to the recognized text. For each text recognition result, based on that text recognition result, extract the target text segment that meets the text extraction conditions from the recognized text of that text recognition result; wherein, the text extraction conditions include start and end time constraints and text integrity conditions; For all target text segments, perform a text merging operation to obtain at least one text set; each text set includes at least two similar texts, and all similar texts included in each text set satisfy a preset similarity condition; For each text set, determine the corresponding text template based on preset template conditions.

[0040] The identified text includes at least one character, and the timestamp information includes the timestamp corresponding to each character.

[0041] Optionally, the start and end time restrictions may include at least one of the start restriction time, end restriction time, and restriction duration, which are not limited in this embodiment of the invention.

[0042] As can be seen, this optional embodiment can first perform text recognition on the newly added platform creative to obtain the text recognition result, and then extract the target text fragments that meet the text extraction conditions from the text recognition result. Then, a text merging operation is performed on all target text fragments to obtain at least one text set. Finally, the text template corresponding to each text set is determined. This can achieve efficient analysis and extraction of key text content in platform creatives, thereby improving the completeness of the restoration of platform creative content, improving the accuracy of the extraction, aggregation, and filtering of creative text content, and thus helping to improve the accuracy of text template generation.

[0043] In this optional embodiment, optionally, performing a text recognition operation on each newly added platform creative to obtain the text recognition result corresponding to each platform creative may include the following operations: For each newly added platform creative, when the creative type of the platform creative is video, the audio truncated duration of the text to be recognized is determined based on the obtained text recognition requirements and / or development cost requirements; wherein, the audio truncated duration can be the full duration of the platform creative or a preset duration, and the preset duration is longer than the limited duration corresponding to the above start and end time constraints. The audio extraction operation is performed on the segment of the creative content created on the platform to obtain the target audio data; the target audio data is the audio data of the speech signal. Perform speech recognition on the target audio data to obtain the text recognition result corresponding to the creative content on the platform.

[0044] The target audio data is audio data that is distinct from non-speech signals.

[0045] Optionally, a pre-defined speech recognition algorithm can be used to perform speech recognition on the target audio data to obtain the text recognition result corresponding to the platform's creative content. Optionally, the speech recognition algorithm can be the Paraformer-large long audio model from FunASR, an open-source Automatic Speech Recognition (ASR) system that supports Chinese character recognition. The speech recognition algorithm can also be other algorithms capable of recognizing text content in audio; this embodiment of the invention is not limited in this regard. Optionally, the recognized text included in the text recognition result can be text with punctuation marks; this embodiment of the invention is not limited in this regard.

[0046] Optionally, the timestamp corresponding to each character may include the start and end timestamps of the character's appearance; however, this embodiment of the invention does not impose such limitations.

[0047] As can be seen, this optional embodiment can also extract the target audio data of the platform creative based on the determined audio extraction duration when the platform creative is a video creative, and then perform speech recognition on the target audio data to obtain the text recognition result. This can efficiently extract audio data while restoring the complete speech information required by the video creative, thereby improving the recognition efficiency and accuracy of the text content of the video creative, and thus helping to improve the subsequent analysis efficiency and accuracy of the text recognition results.

[0048] In this optional embodiment, optionally, for each newly added platform creative, performing an audio extraction operation on a segment of the preceding audio length of the platform creative to obtain target audio data may include the following operations: Extract the audio data of the content created on this platform; Based on a pre-defined audio analysis algorithm, audio features corresponding to the audio data are extracted; Based on audio characteristics, determine whether the audio data is a song type; When the audio data is determined to be a song, no further processing is performed on the audio data. When it is determined that the audio data is not a song, Voice Activity Detection (VAD) is performed on the audio data to filter out non-speech signal audio segments, obtain audio segments of speech signals, and identify the audio segments of speech signals as the target audio data.

[0049] Optionally, extracting audio data from the platform's content can specifically involve converting the video file created on the platform into an audio file as audio data, or extracting audio data from the video file created on the platform. This embodiment of the invention does not impose any limitations on this method. For example, the audio data format can be 16-bit mono with a 16kHz sampling rate. This embodiment of the invention does not impose any limitations on this method.

[0050] Optionally, the audio analysis algorithm can be the open-source audio processing framework AudioFlux, or other audio analysis algorithms capable of extracting audio features; this embodiment of the invention does not impose any limitations.

[0051] Optionally, the audio segments that are not speech signals may include silent segments and / or noise segments, and may also include other audio segments that are unrelated to the extracted speech content. This embodiment of the invention does not impose any limitations.

[0052] As can be seen, this optional embodiment can also extract audio features of the audio data based on audio analysis algorithms after extracting the audio data of the creative content from the platform. Based on the audio features, it can determine whether the audio data is a song. If it is a song, no further processing is performed on the audio data. If it is not a song, invalid audio segments are filtered out to obtain valid audio segments of the speech signal as target audio data. This can achieve efficient filtering of invalid audio data, thereby improving the processing efficiency and accuracy of audio data of video creative content, which in turn helps to further improve the analysis efficiency and accuracy of subsequent text recognition results.

[0053] In this optional embodiment, optionally, for each text recognition result, extracting a target text segment that meets the text extraction conditions from the recognized text of the text recognition result may include the following operations: Based on the timestamp information and start and end time constraints in the text recognition result, extract the text segments that meet the start and end time constraints from the recognized text in the text recognition result; Determine whether the text content of the extracted fragment is a complete sentence; When it is determined that the text content of the extracted fragment is a complete sentence, the extracted fragment is determined to meet the text integrity condition, and the extracted fragment is determined as the target fragment text; When it is determined that the text content of the extracted fragment is not a complete sentence, it is determined that the extracted fragment does not meet the text integrity condition. Based on the identified text, the extracted fragment is padded to obtain the target fragment text that meets the text integrity condition.

[0054] For example, the start and end time restriction conditions include a limited duration, which is 3 seconds, but can also be other values; the start and end time restriction conditions may also include an end time restriction, which can be 00:03, but this embodiment of the invention does not limit it.

[0055] For example, in the audio data created on a certain platform, the complete dialogue for the first 3.5 seconds is "It's been a long time since I've seen such a rolled-up book; it was only published in April this year." However, the third second only goes up to the word "published," which is not a complete sentence. Therefore, the word "published" up to the third and fifth seconds can be added to the sentence to form a complete sentence, so that the extracted text fragment can be a complete sentence.

[0056] The method may further include the following operations: calculating the hash value of each target fragment text and using the hash value as the unique identifier of the target fragment text, so as to store the target fragment text and the hash value together in the database. This embodiment of the invention does not limit the scope of the method.

[0057] As can be seen, this optional embodiment can also extract text segments that meet the start and end time constraints from the recognized text of the text recognition result based on the timestamp information and start and end time constraints in the text recognition result. If it is determined that the text content of the extracted text segment is a complete sentence, then the extracted text segment is determined to meet the text integrity condition and is identified as the target text segment. Otherwise, it is determined that the extracted text segment does not meet the text integrity condition, and the extracted text segment is padded according to the recognized text to obtain the target text segment that meets the text integrity condition. This can improve the accuracy and flexibility of text extraction, as well as the content integrity of the extraction result, which is conducive to improving the integrity and accuracy of the target text segment, thereby improving the accuracy of subsequent text template generation.

[0058] In this optional embodiment, the text template generation method may also include: Before performing text merging operations on all target text fragments to obtain at least one text set, the topic type corresponding to each target text fragment is identified based on a pre-trained topic recognition model. Obtain the original additional related data corresponding to each platform's creative content; For each platform creative, the theme type corresponding to that platform creative and the original additional related data corresponding to that platform creative are synchronized to the determined search analysis engine.

[0059] Optionally, when the creative text processing scenario is a scenario targeting the lines mentioned in the creative, the topic type corresponding to each target text segment is identified based on a pre-trained topic recognition model. Specifically, this can be achieved by monitoring data changes in the database and calling a pre-trained line pattern recognition model to identify the topic type corresponding to each target text segment. This decouples the relevant recognition and data synchronization services by monitoring data changes, thereby enabling the distributed deployment of these services, improving the consumption speed of downstream services, and making it more conducive to processing massive amounts of data.

[0060] Optionally, the original additional associated data corresponding to each platform creative may include original data and additional data. The original data may include, but is not limited to, the video ID of the video produced using the platform creative and / or the product ID of the product associated with the video. The additional data may include, but is not limited to, the category ID of the product's category. This embodiment of the invention does not impose any limitations on this. Further, optionally, the video may be an advertising video, and the product may be the product to be promoted in the advertisement. This embodiment of the invention does not impose any limitations on this.

[0061] Optionally, the search analysis engine may include, but is not limited to, the ES (ElasticSearch) search engine; this embodiment of the invention does not impose any limitations. Further optionally, the search analysis engine may also be used to implement search functionality.

[0062] As can be seen, this optional embodiment can also identify the topic type corresponding to each target text segment based on the topic recognition model before merging the text set, and obtain the original additional related data corresponding to each platform creative. Then, for each platform creative, the topic type corresponding to the platform creative and the original additional related data are synchronized to the determined search analysis engine. This can classify the text according to topic type, making it easier for users to browse different creative text template styles. At the same time, it can improve the synchronization efficiency and comprehensiveness of the original additional related data of platform creatives, which is beneficial to the convenience and reliability of data acquisition in subsequent data analysis steps such as analyzing the similarity between texts.

[0063] In this optional embodiment, optionally, performing a text merging operation on all target fragment texts to obtain at least one text set may include the following operations: From all target text segments, determine the current identified text to be analyzed; The similarity query algorithm of the search analysis engine is invoked to filter out at least one candidate similar text that matches the currently identified text from all historical synchronized texts corresponding to the search analysis engine; wherein, the similarity between each candidate similar text and the currently identified text is greater than or equal to a preset similarity threshold. Calculate the average relative edit distance between each candidate similar text and the currently identified text; Target similar texts with an average relative edit distance less than or equal to a preset distance threshold are selected from all candidate similar texts, and the currently identified text and all target similar texts corresponding to the currently identified text are determined as a text set; wherein each similar text included in the text set is either the currently identified text or a target similar text; The operation of determining the currently identified text to be analyzed from all target fragment texts is repeated until every target fragment text has a corresponding text set.

[0064] Optionally, determining the current identified text to be analyzed from all target fragment texts can specifically involve: using an offline timed script to query the most recently added target fragment texts, and determining the target fragment text with the earliest addition time from all unanalyzed target fragment texts as the current identified text to be analyzed, based on the order of addition time. This embodiment of the invention does not limit this.

[0065] Among them, all historical synchronized text corresponding to the search analysis engine can be text that has been synchronized to the search analysis engine in the past.

[0066] Optionally, the similarity query algorithm of the search analysis engine can be the MLT (More Like This) query syntax of the ES search engine, or other algorithms that can analyze and filter texts with high similarity. This embodiment of the invention does not limit the specific algorithms.

[0067] As can be seen, this optional embodiment can also, during the text merging operation, first determine the currently identified text to be analyzed, then call the similarity query algorithm of the search analysis engine to filter out at least one candidate similar text that matches the currently identified text from all historical synchronized texts corresponding to the search analysis engine, then calculate the average relative edit distance between each candidate similar text and the currently identified text, and filter out target similar texts whose average relative edit distance is less than or equal to a preset distance threshold, then determine the currently identified text and all target similar texts corresponding to the currently identified text as a text set, and repeat the above process until all newly added target fragment texts have corresponding text sets. This can improve the accuracy of similarity calculation and average relative edit distance calculation for the currently identified text, and improve the accuracy of target similar text filtering, thereby improving the efficiency and accuracy of text set generation, and further improving the accuracy of subsequent text template determination.

[0068] In this optional embodiment, optionally, the similarity query algorithm of the search analysis engine is invoked to filter out at least one candidate similar text that matches the currently identified text from all historical synchronized texts corresponding to the search analysis engine. This may include the following operations: The relevance scoring algorithm of the search analysis engine is invoked to calculate the similarity between the currently identified text and each historical synchronized text; Filter out at least one target similarity that is greater than or equal to a preset similarity threshold from all similarities, and determine the historical synchronized text corresponding to each target similarity as candidate similar text.

[0069] Optionally, the relevance scoring algorithm can be as follows:

[0070] in, Q For querying, which is the currently identified text mentioned above; D This refers to a document, specifically any of the historical synchronized text mentioned above. q i As keywords, f ( q i ,D This is used to calculate the number of times the keyword appears; | D | is a document D The word count length; avgdl is the average document length of all documents (i.e., all historical synchronized text); k 1 and b These are free parameters; and, IDF( q i The inverse document weight of the query term is calculated using the following formula:

[0071] in, N It is the total number of documents (i.e., the total number of historical synchronized texts), n( q i ) is included q i The number of documents.

[0072] As can be seen, this optional embodiment can first calculate the similarity between the currently identified text and each historical synchronized text based on the relevance scoring algorithm of the search analysis engine, and then further filter out historical synchronized texts with high similarity as candidate similar texts. This can further improve the accuracy and efficiency of similarity calculation between texts, while improving the accuracy of candidate similar text selection, thereby helping to improve the accuracy of subsequent target similar text selection.

[0073] In this optional embodiment, calculating the average relative edit distance between each candidate similar text and the currently identified text may include the following operations: Based on the determined edit distance algorithm, the edit distance between each candidate similar text and the currently identified text is calculated; Based on the formula for calculating the average relative edit distance, the average relative edit distance between each candidate similar text and the currently identified text is calculated according to the edit distance mentioned above; wherein, the formula for calculating the average relative edit distance can be as follows: ; Edit Distance consists of two strings. S 1 and S Edit distance of 2, len( S 1) Represents a string S The length of 1, len( S 2) Represents a string S The length is 2; where, optionally, S 1 can be any candidate similar text. S 2 can be the currently recognized text; or, S 1 can be the currently recognized text. S 2 can be any candidate similar text, and the embodiments of the present invention are not limited thereto.

[0074] Optionally, the edit distance algorithm can be the Levenshtein Distance algorithm; further, the Levenshtein distance calculation method can be summarized as: the minimum number of editing operations (such as: addition, deletion, replacement) required to transform one string into another, for example: The Levenstein distance corresponding to the conversion from "kitten" to "sitting" is 3. The specific conversion process can be as follows: kitten→sitten(k→s) sitten→sittin(e→i) sittin → sitting (insert g); a total of 3 editing operations.

[0075] As can be seen, this optional embodiment can first calculate the edit distance between each candidate similar text and the currently identified text based on the edit distance algorithm, and then calculate the average relative edit distance between each candidate similar text and the currently identified text based on the average relative edit distance calculation formula. This can further improve the accuracy of the edit distance calculation and the accuracy of the average relative edit distance calculation, while balancing the sensitivity to text length differences and adapting to the characteristics of variable text length. This is conducive to further improving the accuracy of target similar text screening, and thus further improving the accuracy of text set generation.

[0076] In this optional embodiment, optionally, for each text set, determining the text template corresponding to that text set based on preset template conditions may include the following operations: Determine the recognition time for each similar text; The text with the earliest identification time among all similar texts in the text set is determined as the text template corresponding to the text set. In this text set, all similar texts are related texts of the text template.

[0077] As can be seen, this optional embodiment can also determine the earliest identified similar text in each text set as a text template, and merge other texts in the set under the text template to complete the aggregation of similar texts, which can improve the efficiency of determining text templates, thereby improving the aggregation efficiency of similar texts.

[0078] Example 2 Please see Figure 2 , Figure 2 This is a flowchart illustrating an intelligent creative text processing method based on artificial intelligence, as disclosed in an embodiment of the present invention. Figure 2 The described AI-based intelligent creative text processing method can be applied to an AI-based intelligent creative text processing device. This device may include one of a smart terminal, a smart device, a smart system, and a server. The server may be a local server or a cloud server; this embodiment of the invention does not limit the specific application. Figure 2 As shown, this AI-based intelligent processing method for creative text may include the following operations: 201. In response to the template summary view operation, display the summary page corresponding to the template summary view operation.

[0079] 202. Display template information corresponding to at least one summary text template on the summary page.

[0080] Each summary text template satisfies the sorting conditions corresponding to the template summary viewing operation, and each summary text template also satisfies the pre-set text extraction conditions; the template information includes the template content, as well as the theme type and / or template association information; wherein, the template association information includes at least one of the following: associated text information, first associated creative information, first associated product information, and first popularity information.

[0081] 203. In response to a text template viewing operation on the summary page, determine the target text template corresponding to the text template viewing operation from all summary text templates, and display the preview page corresponding to the target text template based on preset preview display parameters.

[0082] In this embodiment of the invention, optionally, the text template viewing operation for the summary page can be specifically as follows: performing a click operation on any summary text template in the summary page, where the target text template corresponding to the text template viewing operation is also the click object of the click operation. This embodiment of the invention does not impose any limitations on this.

[0083] In this embodiment of the invention, the preset preview display parameters include at least the display mode of the preview page; optionally, the display mode of the preview page can be an overlay display mode, that is, a display mode that overlays the summary page; when the display mode of the preview page is an overlay display mode, preferably, the display mode can be more specifically a side overlay display mode (such as: sidebar), optionally, it can also be one of a modal overlay display mode, a top overlay display mode, a bottom overlay display mode, or a full-screen overlay display mode, and this embodiment of the invention does not limit it.

[0084] 204. Display the associated preview information corresponding to the target text template on the preview page.

[0085] In this embodiment of the invention, the associated preview information includes at least one of core associated data, text preview information of at least one associated text, and creative preview information of at least one associated creative; wherein each associated text and each associated creative is associated with a target text template.

[0086] Further optionally, the core associated data may include the number of first associated creatives of at least one creative promotion type corresponding to the target text template, the number of first associated products corresponding to the target text template, and the number of associated reachers corresponding to the target text template. This embodiment of the invention does not impose any limitations.

[0087] Further optionally, the text preview information of each associated text may include the text content of the associated text and / or the number of second associated creatives of at least one creative promotion type corresponding to the associated text, which is not limited in this embodiment of the invention; further optionally, the associated preview information may include the text preview information of a first preset number of associated texts filtered based on preset sorting conditions, wherein, for example, the preset sorting conditions may be descending order of similarity, and the first preset number may be 3, which is not limited in this embodiment of the invention; further optionally, words and / or sentences in the text content of each associated text that are the same as the target text template may be displayed by a first preset marking method, and the remaining content of the text content of each associated text may be displayed by a second preset marking method, wherein the text display effects of the first preset marking method and the second preset marking method are different; further optionally, the first preset marking method may correspond to a first text color, and the second preset marking method may correspond to a second text color, and the first text color is significantly different from the second text color; for example, the first preset marking method may be a highlight marking method, and the first text color may be any highlight color, which is not limited in this embodiment of the invention.

[0088] Further optionally, the associated preview information may include creative preview information of a second preset number of associated creatives of at least one creative promotion type. For example, the second preset number may be 5, and the associated preview information may include 5 associated creatives corresponding to 3 creative promotion types respectively. This embodiment of the invention does not limit this. Furthermore, the creative preview information of each associated creative may include any screenshot of the associated creative. Furthermore, the screenshot may be a screenshot corresponding to the timestamp information of the associated text. This embodiment of the invention does not limit this.

[0089] In this embodiment of the invention, the interface display effect of the preview page corresponding to the target text template can be referred to as an example. Figure 4 , Figure 4 This is a schematic diagram of an interface for another AI-based intelligent text processing method disclosed in an embodiment of the present invention; wherein, as shown... Figure 4 As shown, the preview page displays related preview information, which includes core data (i.e., the aforementioned core related data) and text preview information for three related texts (such as...). Figure 4 The "Top 3 Related Lines List" shown includes the line content, the number of related ad creatives, the number of product promotion ad creatives, and the number of live stream traffic-driving creatives, as well as 3 types of creative promotion (e.g., Figure 4 The screenshots shown for the five related creatives corresponding to "Related Product Advertising Material", "Related Live Streaming Advertising Material" and "Related Live Streaming Advertising Material" are not limited in this embodiment of the invention.

[0090] For further detailed descriptions of steps 201-202 in this embodiment of the invention, please refer to the detailed description of steps 101-102 in Embodiment 1. These descriptions will not be repeated in this embodiment of the invention.

[0091] As can be seen, the method described in the embodiments of the present invention can respond to a template summary viewing operation, display the summary page corresponding to the template summary viewing operation, and display template information corresponding to at least one summary text template on the summary page. It can efficiently filter and aggregate various text templates used in a large number of creative works based on sorting conditions, thereby improving the processing efficiency of creative text content while efficiently displaying the creative text analysis results. This helps users efficiently browse the template content of various creative text templates, as well as related popularity information and associated data, thus improving the efficiency and accuracy of users' decision-making in selecting creative text content during the creative creation process. Furthermore, in response to a text template viewing operation on the summary page, it can determine the target text template corresponding to the text template viewing operation from all summary text templates, and display the preview page corresponding to the target text template based on preset preview display parameters. Then, it displays the associated preview information corresponding to the target text template on the preview page, enabling efficient display of key information of the target text template for users to browse. This further improves the efficiency of users' browsing of the general situation of the associated data of the text template, thereby further improving the efficiency and accuracy of users' decision-making in selecting creative text content during the creative creation process.

[0092] In an optional embodiment, the method may further include the following operations: In response to a view request for the preview page, display the target details page corresponding to the target text template; Based on preset details display parameters, display the data details information and / or the text details information corresponding to the target text template on the target details page; The data details include core related data, creative placement trends, and category distribution. The creative placement trend indicates the trend of changes in the number of related creatives placed for the target text template. The category distribution indicates the product category distribution of the related products for the target text template. The text details information includes at least one related text detail information; the text details information includes at least one of the following: text content, second related creative information, second related product information, and second popularity information.

[0093] Optionally, the core associated data may also include the number of associated texts.

[0094] More specifically, the creative placement trend is used to indicate the trend of the number of related creatives placed for each type of creative promotion and the target text template; the category distribution is used to indicate the product category distribution of related products for each type of creative promotion and the target text template.

[0095] In this embodiment of the invention, for example, the interface display effect of displaying the data details information corresponding to the target text template on the target details page can be referred to. Figure 5 , Figure 5 This is a schematic diagram of an interface for another AI-based intelligent text processing method disclosed in an embodiment of the present invention; wherein, as shown... Figure 5 As shown, the data details displayed on the target details page may include core data (i.e., the aforementioned core related data), material placement trends (i.e., the aforementioned creative placement trends), and category distribution. The creative placement trends can be displayed in the form of a line chart, with each type of creative promotion corresponding to one line. The category distribution can be displayed in the form of a pie chart, with each type of creative promotion corresponding to one pie chart. This embodiment of the invention does not impose any limitations on this.

[0096] Optionally, the second associated creative information for each associated text includes the number of second associated creatives, more specifically, the number of second associated creatives being the number of second associated creatives associated with the associated text; the second associated product information for each associated text includes the number of second associated products, more specifically, the number of second associated products being the number of second associated products associated with the associated text; the second popularity information for each associated text includes one or more combinations of the estimated exposure value of the advertising creative associated with the associated text, the growth rate of the advertising creative exposure, and the number of days of placement, which are not limited in this embodiment of the invention.

[0097] In this embodiment of the invention, for example, the text details information that displays the data details information corresponding to the target text template on the target details page can be found here. Figure 6 , Figure 6 This is a schematic diagram of an interface for another AI-based intelligent text processing method disclosed in an embodiment of the present invention; wherein, as shown... Figure 6 As shown, the text details displayed on the target details page may include text content (such as...). Figure 6 The displayed dialogue content), number of associated ad creatives, estimated ad creative exposure, ad creative exposure growth rate, number of days of campaign execution, and number of associated products (i.e., the number of the second associated products mentioned above); furthermore, the data details information corresponding to the target text template displayed on this target details page can also be divided into columns according to the creative promotion type (e.g., Figure 6The examples shown are "Product Advertising Materials," "Live Stream Traffic Generation Advertising Materials," and "CID Advertising Materials." Furthermore, on this target details page, you can also set filtering conditions for data details (such as...). Figure 6 The "category selection", "traffic platform", etc. shown in this embodiment of the invention are not limited.

[0098] As can be seen, this optional embodiment can respond to the details viewing operation of the preview page, display the target details page corresponding to the target text template, and display the data details information and / or the text details information corresponding to the target text template on the target details page based on preset details display parameters. It can efficiently and comprehensively display the associated text, creativity, products, popularity and other data of the text template, thereby helping users to further understand more details about the text template, and further improve the accuracy of users' decision-making in selecting creative text content during the creative creation process.

[0099] Example 3 Please see Figure 7 , Figure 7 This is a schematic diagram of the structure of an artificial intelligence-based intelligent creative text processing device disclosed in an embodiment of the present invention. Wherein, Figure 7 The described AI-based intelligent creative text processing device may include one of a smart terminal, a smart device, a smart system, and a server, wherein the server may be a local server or a cloud server, and this embodiment of the invention is not limited thereto. Figure 7 As shown, the AI-based intelligent creative text processing device may include: Display module 301 is used to respond to the template summary viewing operation and display the summary page corresponding to the template summary viewing operation. The display module 301 is also used to display template information corresponding to at least one summary text template on the summary page; Each summary text template satisfies the sorting conditions corresponding to the template summary viewing operation, and each summary text template also satisfies the pre-set text extraction conditions; the template information includes the template content, as well as the theme type and / or template association information; wherein, the template association information includes at least one of the following: associated text information, first associated creative information, first associated product information, and first popularity information.

[0100] As can be seen, the apparatus described in the embodiments of the present invention can respond to the template summary viewing operation, display the summary page corresponding to the template summary viewing operation, and display the template information corresponding to at least one summary text template on the summary page. It can efficiently filter and aggregate various text templates used in a large number of creative works based on the arrangement conditions, thereby improving the processing efficiency of creative text content and efficiently displaying the creative text analysis results. This helps users to efficiently browse the template content of various creative text templates as well as related popularity information and associated data, which in turn helps improve the efficiency and accuracy of users' decision-making in selecting creative text content during the creative creation process.

[0101] In an optional embodiment, such as Figure 8 As shown, the device may further include: The determination module 302 is used to determine the target text template corresponding to the text template viewing operation from all summary text templates in response to the text template viewing operation on the summary page. The display module 301 is also used to display the preview page corresponding to the target text template based on preset preview display parameters; The display module 301 is also used to display the associated preview information corresponding to the target text template on the preview page; The preset preview display parameters include at least the display method of the preview page; the associated preview information includes at least one of the following: core associated data, text preview information of at least one associated text, and creative preview information of at least one associated creative; wherein each associated text and each associated creative is associated with the target text template.

[0102] As can be seen, the apparatus described in this optional embodiment can respond to a text template viewing operation on a summary page, determine the target text template corresponding to the text template viewing operation from all summary text templates, and display the preview page corresponding to the target text template based on preset preview display parameters. Then, it displays the associated preview information corresponding to the target text template on the preview page. This can efficiently display the key information of the target text template for users to browse, and further improve the user's browsing efficiency of the general situation of the associated data of the text template. This is conducive to further improving the user's decision-making efficiency and accuracy in selecting creative text content during the creative creation process.

[0103] In this optional embodiment, optionally, the display module 301 is further configured to display the target details page corresponding to the target text template in response to a details viewing operation on the preview page; The display module 301 is also used to display data detail information and / or text detail information corresponding to the target text template on the target detail page based on preset detail display parameters; The data details include core related data, creative placement trends, and category distribution. The creative placement trend indicates the trend of changes in the number of related creatives placed for the target text template. The category distribution indicates the product category distribution of the related products for the target text template. The text details information includes at least one related text detail information; the text details information includes at least one of the following: text content, second related creative information, second related product information, and second popularity information.

[0104] As can be seen, the apparatus described in this optional embodiment is also capable of displaying the target details page corresponding to the target text template in response to a details viewing operation on the preview page, and displaying data details information and / or text details information corresponding to the target text template on the target details page based on preset details display parameters. This enables efficient and comprehensive display of data such as associated text, creativity, products, and popularity of the text template, thereby helping users to further understand more details about the text template and further improving the accuracy of users' decision-making in selecting creative text content during the creative creation process.

[0105] In an optional embodiment, each summary text template is a text template, and the text template can be generated in the following way: Perform text recognition operation on each newly added platform creative to obtain the text recognition result corresponding to each platform creative; the text recognition result includes the recognized text and the timestamp information corresponding to the recognized text. For each text recognition result, based on that text recognition result, extract the target text segment that meets the text extraction conditions from the recognized text of that text recognition result; wherein, the text extraction conditions include start and end time constraints and text integrity conditions; For all target text segments, perform a text merging operation to obtain at least one text set; each text set includes at least two similar texts, and all similar texts included in each text set satisfy a preset similarity condition; For each text set, determine the corresponding text template based on preset template conditions.

[0106] As can be seen, the apparatus described in this optional embodiment can first perform text recognition on the newly added platform creative to obtain the text recognition result, and then extract the target text fragments that meet the text extraction conditions from the text recognition result. A text merging operation is then performed on all target text fragments to obtain at least one text set. Then, the text template corresponding to each text set is determined. This can achieve efficient analysis and extraction of key text content in the platform creative, thereby improving the completeness of the restoration of the platform creative content while improving the accuracy of the extraction, aggregation, and filtering of creative text content, which in turn helps to improve the accuracy of text template generation.

[0107] In this optional embodiment, the text template generation method may also include: Before performing text merging operations on all target text fragments to obtain at least one text set, the topic type corresponding to each target text fragment is identified based on a pre-trained topic recognition model. Obtain the original additional related data corresponding to each platform's creative content; For each platform creative, the theme type corresponding to that platform creative and the original additional related data corresponding to that platform creative are synchronized to the determined search analysis engine.

[0108] As can be seen, the apparatus described in this optional embodiment can also identify the topic type corresponding to each target text segment based on the topic recognition model before merging the text set, and obtain the original additional related data corresponding to each platform creative. Then, for each platform creative, the topic type corresponding to the platform creative and the original additional related data are synchronized to the determined search analysis engine. This can classify the text according to topic type, making it easier for users to browse different creative text template styles. At the same time, it can improve the synchronization efficiency and comprehensiveness of the original additional related data of the platform creative, which is beneficial to the convenience and reliability of data acquisition in subsequent data analysis steps such as analyzing the similarity between texts.

[0109] In this optional embodiment, the specific method for performing a text merging operation on all target text fragments to obtain at least one text set may include: From all target text segments, determine the current identified text to be analyzed; The similarity query algorithm of the search analysis engine is invoked to filter out at least one candidate similar text that matches the currently identified text from all historical synchronized texts corresponding to the search analysis engine; wherein, the similarity between each candidate similar text and the currently identified text is greater than or equal to a preset similarity threshold. Calculate the average relative edit distance between each candidate similar text and the currently identified text; Target similar texts with an average relative edit distance less than or equal to a preset distance threshold are selected from all candidate similar texts, and the currently identified text and all target similar texts corresponding to the currently identified text are determined as a text set; wherein each similar text included in the text set is either the currently identified text or a target similar text; The operation of determining the currently identified text to be analyzed from all target fragment texts is repeated until every target fragment text has a corresponding text set.

[0110] As can be seen, the apparatus described in this optional embodiment can also, during the text merging operation, first determine the currently identified text to be analyzed, then call the similarity query algorithm of the search analysis engine to filter out at least one candidate similar text that matches the currently identified text from all historical synchronized texts corresponding to the search analysis engine, then calculate the average relative edit distance between each candidate similar text and the currently identified text, and filter out target similar texts whose average relative edit distance is less than or equal to a preset distance threshold, then determine the currently identified text and all target similar texts corresponding to the currently identified text as a text set, and repeat the above process until all newly added target fragment texts have corresponding text sets. This can improve the accuracy of similarity calculation and average relative edit distance calculation for the currently identified text, and improve the accuracy of target similar text filtering, thereby improving the efficiency and accuracy of text set generation, and further improving the accuracy of subsequent text template determination.

[0111] In this optional embodiment, optionally, the specific method for determining the text template corresponding to each text set based on preset template conditions may include: Determine the recognition time for each similar text; The text with the earliest identification time among all similar texts in the text set is determined as the text template corresponding to the text set. In this text set, all similar texts are related texts of the text template.

[0112] As can be seen, the apparatus described in this optional embodiment can also determine the earliest identified similar text in each text set as a text template, and merge other texts in the set under the text template to complete the aggregation of similar texts, thereby improving the efficiency of determining the text template and thus improving the aggregation efficiency of similar texts.

[0113] Example 4 Please see Figure 9 , Figure 9This is a schematic diagram of the structure of another intelligent creative text processing device based on artificial intelligence disclosed in an embodiment of the present invention. For example... Figure 9 As shown, the AI-based intelligent creative text processing device may include: Memory 401 storing executable program code; Processor 402 coupled to memory 401; The processor 402 calls the executable program code stored in the memory 401 to execute some or all of the steps in the AI-based creative text intelligent processing method described in Embodiment 1 or Embodiment 2 of the present invention.

[0114] Example 5 This invention discloses a computer storage medium storing computer instructions. When these computer instructions are invoked, they are used to execute some or all of the steps in the AI-based intelligent creative text processing method described in Embodiment 1 or Embodiment 2 of this invention.

[0115] Example 6 This invention discloses a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to perform some or all of the steps in the AI-based creative text intelligent processing method described in Embodiment 1 or Embodiment 2.

[0116] The device embodiments described above are merely illustrative. The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0117] Through the detailed description of the above embodiments, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, including read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-Erasable Programmable Read-Only Memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0118] Finally, it should be noted that the creative text intelligent processing method and apparatus based on artificial intelligence disclosed in the embodiments of the present invention are merely preferred embodiments of the present invention, and are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features therein; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An artificial intelligence-based creative text intelligent processing method, characterized in that, The method comprises: in response to a template summary viewing operation, displaying a summary page corresponding to the template summary viewing operation; displaying template information corresponding to at least one summary text template on the summary page; each of the summary text templates satisfies the arrangement condition corresponding to the template summary viewing operation, and each of the summary text templates also satisfies a pre-set text interception condition; the template information includes template content, and also includes theme type and / or template association information; the template association information includes at least one of associated text information, first associated creative information, first associated product information, and first heat information. 2.The AI-based creative text intelligent processing method according to claim 1, characterized in that, The method further comprises: in response to a text template viewing operation for the summary page, determining a target text template corresponding to the text template viewing operation from all the summary text templates, and displaying a preview page corresponding to the target text template based on a pre-set preview display parameter; displaying associated preview information corresponding to the target text template on the preview page; the pre-set preview display parameter at least includes a display mode of the preview page; the associated preview information includes at least one of core associated data, text preview information of at least one associated text, and creative preview information of at least one associated creative; each of the associated text and each of the associated creative is associated with the target text template. 3.The AI-based creative text intelligent processing method of claim 2, wherein, The method further comprises: in response to a detail viewing operation for the preview page, displaying a target detail page corresponding to the target text template; based on a pre-set detail display parameter, displaying data detail information corresponding to the target text template and / or text detail information corresponding to the target text template on the target detail page; the data detail information includes the core associated data, creative delivery trend, and category distribution; the creative delivery trend is used to represent the delivery quantity trend of the associated creative corresponding to the target text template; the category distribution is used to represent the product category distribution of the associated product corresponding to the target text template; the text detail information includes text detail information of at least one of the associated texts; the text detail information includes at least one of text content, second associated creative information, second associated product information, and second heat information. 4.The AI-based creative text intelligent processing method according to any one of claims 1-3, characterized in that, Each of the summary text templates is a text template, and the text template is generated by the following way: performing a text recognition operation on each newly added platform creative to obtain a text recognition result corresponding to each platform creative; the text recognition result includes recognized text and timestamp information corresponding to the recognized text; for each text recognition result, according to the text recognition result, a target fragment text satisfying the text interception condition is intercepted from the recognized text of the text recognition result; the text interception condition includes start and end time limit conditions and text integrity conditions; For all the target text segments, a text merging operation is performed to obtain at least one text set; each text set includes at least two similar texts, and all the similar texts included in each text set satisfy a preset similarity condition; For each text set, a text template corresponding to that text set is determined based on preset template conditions. 5.The AI-based creative text intelligent processing method according to claim 4, characterized in that, The text template generation method also includes: Before performing a text merging operation on all the target text fragments to obtain at least one text set, the topic type corresponding to each target text fragment is identified based on a pre-trained topic recognition model. Obtain the original additional associated data corresponding to each of the platform's creative ideas; For each platform creative, the topic type corresponding to the platform creative and the original additional related data corresponding to the platform creative are synchronized to the determined search analysis engine. 6.The AI-based creative text intelligent processing method according to claim 5, characterized in that, The step involves performing a text merging operation on all the target text fragments to obtain at least one text set, including: From all the target fragment texts, determine the current identified text to be analyzed; The similarity query algorithm of the search analysis engine is invoked to filter out at least one candidate similar text that matches the currently identified text from all historical synchronized texts corresponding to the search analysis engine; wherein the similarity between each candidate similar text and the currently identified text is greater than or equal to a preset similarity threshold. Calculate the average relative edit distance between each candidate similar text and the currently identified text; From all the candidate similar texts, target similar texts with an average relative edit distance less than or equal to a preset distance threshold are selected, and the currently identified text and all the target similar texts corresponding to the currently identified text are determined as a text set; wherein, each of the similar texts included in the text set is either the currently identified text or the target similar text; The operation of determining the current identified text to be analyzed from all the target fragment texts is repeated until each target fragment text has a corresponding text set. 7.The AI-based creative text intelligent processing method of claim 4, wherein, For each of the text sets, determining the text template corresponding to that text set based on preset template conditions includes: Determine the recognition time corresponding to each of the similar texts; The text with the earliest recognition time among all the similar texts in the text set is determined as the text template corresponding to the text set. In this text set, all similar texts are associated texts of the text template.

8. An artificial intelligence-based creative text intelligent processing apparatus, characterized by, The device includes: The display module is used to respond to the template summary viewing operation and display the summary page corresponding to the template summary viewing operation. The display module is also used to display template information corresponding to at least one summary text template on the summary page; Each of the summarized text templates satisfies the sorting conditions corresponding to the template summary viewing operation, and each of the summarized text templates also satisfies the pre-set text extraction conditions; the template information includes template content, and also includes theme type and / or template association information; wherein, the template association information includes at least one of associated text information, first associated creative information, first associated product information, and first popularity information. 9.A device for intelligent processing of creative text based on artificial intelligence, characterized in that, The device includes: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the AI-based creative text intelligent processing method as described in any one of claims 1-7.

10. A computer storage medium, characterized in that, The computer storage medium stores computer instructions, which, when invoked, are used to execute the AI-based intelligent creative text processing method as described in any one of claims 1-7.

Images