CN114021541A - Presentation generation method, device, equipment and storage medium - Google Patents
Presentation generation method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN114021541A CN114021541A CN202111366306.5A CN202111366306A CN114021541A CN 114021541 A CN114021541 A CN 114021541A CN 202111366306 A CN202111366306 A CN 202111366306A CN 114021541 A CN114021541 A CN 114021541A
- Authority
- CN
- China
- Prior art keywords
- presentation
- information
- template
- manuscript
- requirement
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 35
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 38
- 239000013598 vector Substances 0.000 claims description 61
- 238000004458 analytical method Methods 0.000 claims description 27
- 238000012545 processing Methods 0.000 claims description 13
- 238000013145 classification model Methods 0.000 claims description 10
- 238000011156 evaluation Methods 0.000 claims description 9
- 238000005065 mining Methods 0.000 claims description 8
- 238000010276 construction Methods 0.000 claims description 2
- 238000010606 normalization Methods 0.000 abstract description 5
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 6
- 230000005540 biological transmission Effects 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000005192 partition Methods 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000004806 packaging method and process Methods 0.000 description 1
- 230000002085 persistent effect Effects 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 238000012552 review Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/10—Text processing
- G06F40/166—Editing, e.g. inserting or deleting
- G06F40/186—Templates
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
- G06F16/353—Clustering; Classification into predefined classes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/40—Information retrieval; Database structures therefor; File system structures therefor of multimedia data, e.g. slideshows comprising image and additional audio data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/088—Non-supervised learning, e.g. competitive learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Databases & Information Systems (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Multimedia (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention relates to the field of artificial intelligence and discloses a method, a device, equipment and a storage medium for generating a presentation. The method comprises the following steps: acquiring demand information of a presentation to be generated, which is sent by a terminal; selecting a plurality of alternative presentation templates corresponding to the presentation to be generated by adopting a preset multilayer convolutional neural network model based on template information in the demand information; adopting a preset text recognition model to carry out semantic recognition on the manuscript information in the demand information to obtain the manuscript semantic information corresponding to the manuscript information; based on the semantic information of the manuscript, selecting an optimal presentation template with the highest matching degree with the style type of the presentation to be generated from all the alternative presentation templates; and constructing a corresponding demonstration manuscript by adopting the manuscript information and the optimal demonstration manuscript template, and returning the demonstration manuscript to the terminal. The invention realizes the automatic generation of the presentation, improves the normalization and the quality of the presentation and reduces the generation time of the presentation.
Description
Technical Field
The invention relates to the field of artificial intelligence, in particular to a method, a device, equipment and a storage medium for generating a presentation.
Background
With the development of science and technology, various office software layers are not abundant, but the change of the existing PPT (presentation) making form is not obvious, and the operation in front of a computer is still mainly performed artificially. Each time PPT is compiled, a proper template needs to be searched first, and a large amount of time is wasted; after the templates are selected, a large amount of time is spent for modifying the templates and the content, and finally the PPT generation shows uneven PPT quality according to the writing capability of different groups, namely the quality specification of PPT generation is difficult to guarantee.
Disclosure of Invention
The invention mainly aims to solve the technical problem that the quality specification of PPT generation is difficult to guarantee.
The invention provides a presentation generation method in a first aspect, which comprises the following steps: acquiring demand information of a presentation to be generated, which is sent by a terminal, wherein the demand information of the presentation comprises template information and corresponding document information; selecting a plurality of alternative presentation templates corresponding to the presentation to be generated by adopting a preset multilayer convolutional neural network model based on the template information; adopting a preset text recognition model to carry out semantic recognition on the manuscript information to obtain the manuscript semantic information corresponding to the manuscript information; based on the semantic information of the manuscript, selecting an optimal presentation template with the highest matching degree with the style type of the presentation to be generated from all the alternative presentation templates; and constructing a corresponding demonstration manuscript by adopting the manuscript information and the optimal demonstration manuscript template, and returning the demonstration manuscript to the terminal.
Optionally, in a first implementation manner of the first aspect of the present invention, selecting, based on the template information, a plurality of candidate presentation templates corresponding to the presentation to be generated by using a preset multilayer convolutional neural network model includes: analyzing a first requirement in the template information by using a preset multilayer convolutional neural network model to obtain corresponding first requirement information; analyzing a second requirement in the template information by using a multilayer convolutional neural network model according to the first requirement information and preset iteration parameters to obtain corresponding second requirement information; and carrying out iterative analysis processing on the first requirement information and the second requirement information to obtain a plurality of alternative presentation templates corresponding to the presentation to be generated.
Optionally, in a second implementation manner of the first aspect of the present invention, performing iterative analysis processing on the first requirement information and the second requirement information to obtain a plurality of alternative presentation templates corresponding to the presentation to be generated includes: judging whether the first requirement information meets a preset deep analysis evaluation condition; if the first requirement information does not meet the requirement, analyzing the second requirement information according to the multilayer convolutional neural network model to obtain information to be filled corresponding to the second requirement information; and selecting a plurality of alternative presentation templates corresponding to the presentation to be generated from the preset template library based on the information to be filled.
Optionally, in a third implementation manner of the first aspect of the present invention, performing semantic recognition on the document information by using a preset text recognition model, and obtaining document semantic information corresponding to the document information includes: converting the manuscript information into corresponding word vectors and position vectors by using a preset text recognition model; analyzing cross-sentence semantics corresponding to the manuscript information based on an information retrieval mining algorithm in a preset text recognition model, and determining an adjacent semantic vector corresponding to the manuscript information based on the cross-sentence semantics; and taking the word vector, the position vector and the adjacent semantic vector as the manuscript semantic information corresponding to the manuscript information.
Optionally, in a fourth implementation manner of the first aspect of the present invention, based on the document semantic information, selecting, from the candidate presentation templates, an optimal presentation template with a highest matching degree with a style type of a presentation to be generated includes: based on the word vector, the position vector and the adjacent semantic vector, adopting a preset classification model to correspondingly extract the part-of-speech characteristics, the entity characteristics and the context characteristics of the manuscript information; based on the part of speech characteristics, the entity characteristics and the context characteristics, identifying the style type of the presentation to be generated by using a classification model; and determining a content tag corresponding to the presentation to be generated according to the recognized style type, and selecting the optimal presentation template with the highest matching degree with the content tag from all the alternative presentation templates.
Optionally, in a fifth implementation manner of the first aspect of the present invention, constructing a corresponding presentation by using the document information and the optimal presentation template includes: extracting information fields and file resource fields in the manuscript information, and determining corresponding areas of the information fields and the file resource fields in the optimal presentation manuscript template; writing the information fields into corresponding areas in the optimal presentation template, searching corresponding file resources in a preset database according to the file resource fields, and inserting the file resources into the corresponding areas in the optimal presentation template to obtain corresponding presentations; and adjusting the styles of the information fields and the file resources in the presentation according to a preset style rule to obtain the corresponding presentation.
A second aspect of the present invention provides a presentation generating apparatus, including: the terminal comprises an acquisition module, a display module and a display module, wherein the acquisition module is used for acquiring the requirement information of the presentation to be generated, which is sent by the terminal, and the requirement information of the presentation comprises template information and corresponding document information; the selection module is used for selecting a plurality of alternative presentation templates corresponding to the presentation to be generated by adopting a preset multilayer convolutional neural network model based on the template information; the recognition module is used for performing semantic recognition on the manuscript information by adopting a preset text recognition model to obtain the manuscript semantic information corresponding to the manuscript information; the matching module is used for selecting an optimal presentation template with the highest matching degree with the style type of the presentation to be generated from all the alternative presentation templates based on the semantic information of the presentation; and the construction module is used for constructing the corresponding presentation by adopting the manuscript information and the optimal presentation template and returning the presentation to the terminal.
Optionally, in a first implementation manner of the second aspect of the present invention, the selecting module includes: the first analysis unit is used for analyzing a first requirement in the template information by using a preset multilayer convolutional neural network model to obtain corresponding first requirement information; the second analysis unit is used for analyzing a second requirement in the template information by using a multilayer convolutional neural network model according to the first requirement information and preset iteration parameters to obtain corresponding second requirement information; and the iteration unit is used for performing iteration analysis processing on the first requirement information and the second requirement information to obtain a plurality of alternative presentation templates corresponding to the presentation to be generated.
Optionally, in a second implementation manner of the second aspect of the present invention, the iteration unit is further configured to: judging whether the first requirement information meets a preset deep analysis evaluation condition; if the first requirement information does not meet the requirement, analyzing the second requirement information according to the multilayer convolutional neural network model to obtain information to be filled corresponding to the second requirement information; and selecting a plurality of alternative presentation templates corresponding to the presentation to be generated from the preset template library based on the information to be filled.
Optionally, in a third implementation manner of the second aspect of the present invention, the identification module includes: the conversion unit is used for converting the manuscript information into corresponding word vectors and position vectors by using a preset text recognition model; the retrieval and mining unit is used for analyzing cross-sentence semantics corresponding to the manuscript information based on an information retrieval and mining algorithm in the preset text recognition model and determining an adjacent semantic vector corresponding to the manuscript information based on the cross-sentence semantics; and the generating unit is used for taking the word vector, the position vector and the adjacent semantic vector as the manuscript semantic information corresponding to the manuscript information.
Optionally, in a fourth implementation manner of the second aspect of the present invention, the matching module includes: the classification unit is used for correspondingly extracting the part-of-speech characteristics, the entity characteristics and the context characteristics of the manuscript information by adopting a preset classification model based on the word vector, the position vector and the adjacent semantic vector; based on the part of speech characteristics, the entity characteristics and the context characteristics, identifying the style type of the presentation to be generated by using a classification model; and the matching unit is used for determining the content tags corresponding to the presentation to be generated based on the recognized style types and selecting the optimal presentation template with the highest matching degree with the content tags from the alternative presentation templates.
Optionally, in a fifth implementation manner of the second aspect of the present invention, the building module includes: the extraction unit is used for extracting the information fields and the file resource fields in the manuscript information and determining the corresponding areas of the information fields and the file resource fields in the optimal presentation template; the inserting unit is used for writing the information fields into corresponding areas in the optimal presentation template, searching corresponding file resources in a preset database according to the file resource fields, and inserting the file resources into the corresponding areas in the optimal presentation template to obtain corresponding presentations; and the adjusting unit is used for adjusting the styles of the information fields and the file resources in the presentation according to the preset style rules to obtain the corresponding presentation.
A third aspect of the present invention provides a presentation generating apparatus, comprising: a memory and at least one processor, the memory having instructions stored therein; the at least one processor invokes instructions in the memory to cause the presentation generation apparatus to perform the presentation generation method described above.
A fourth aspect of the present invention provides a computer-readable storage medium having stored therein instructions, which, when run on a computer, cause the computer to execute the presentation generation method described above.
According to the technical scheme, the required information of the presentation to be generated is acquired, wherein a plurality of required alternative presentation templates can be extracted through template information in the required information, the document semantic information in the required information can be identified through the document information in the required information, an optimal presentation template with the most matched style and type is further screened by combining the semantics of the alternative presentation templates and the presentation, and the final delay document is constructed to automatically generate the presentation. Meanwhile, the template style, the semantics of the content of the manuscript and the required template are considered, so that the generation efficiency of the presentation is improved, and the quality normalization of the generation of the presentation is also ensured.
Drawings
Fig. 1 is a schematic diagram of an embodiment of a presentation generation method according to an embodiment of the present invention;
fig. 2 is a schematic diagram of another embodiment of a presentation generation method according to an embodiment of the present invention;
fig. 3 is a schematic diagram of an embodiment of a presentation generation apparatus according to an embodiment of the present invention;
fig. 4 is a schematic diagram of another embodiment of a presentation generation apparatus according to an embodiment of the present invention;
fig. 5 is a schematic diagram of an embodiment of a presentation generating apparatus according to an embodiment of the present invention.
Detailed Description
The embodiment of the invention provides a method, a device, equipment and a storage medium for generating a presentation, which are used for acquiring demand information of the presentation to be generated, which is sent by a terminal; selecting a plurality of alternative presentation templates corresponding to the presentation to be generated by adopting a preset multilayer convolutional neural network model based on template information in the demand information; adopting a preset text recognition model to carry out semantic recognition on the manuscript information in the demand information to obtain the manuscript semantic information corresponding to the manuscript information; based on the semantic information of the manuscript, selecting an optimal presentation template with the highest matching degree with the style type of the presentation to be generated from all the alternative presentation templates; and constructing a corresponding demonstration manuscript by adopting the manuscript information and the optimal demonstration manuscript template, and returning the demonstration manuscript to the terminal. The invention realizes the automatic generation of the presentation, improves the normalization and the quality of the presentation and reduces the generation time of the presentation.
The terms "first," "second," "third," "fourth," and the like in the description and in the claims, as well as in the drawings, if any, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It will be appreciated that the data so used may be interchanged under appropriate circumstances such that the embodiments described herein may be practiced otherwise than as specifically illustrated or described herein. Furthermore, the terms "comprises," "comprising," or "having," and any variations thereof, are intended to cover non-exclusive inclusions, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
For convenience of understanding, a specific flow of the embodiment of the present invention is described below, and referring to fig. 1, a first embodiment of a presentation generating method according to the embodiment of the present invention includes:
101. acquiring demand information of a presentation to be generated, which is sent by a terminal, wherein the demand information of the presentation comprises template information and corresponding document information;
it is to be understood that the execution subject of the present invention may be a presentation generation apparatus, and may also be a terminal or a server, which is not limited herein. The embodiment of the present invention is described by taking a server as an execution subject.
In this embodiment, the corresponding template information and the corresponding document information are determined by obtaining a template style of a presentation application scene selected by a user at a terminal, and presentation requirement information obtained by packaging a presentation theme, a summary description, and the like, which need to be input in a template. The template information refers to which style of presentation template the user desires to obtain, and the document information includes document contents that the user desires to input in the presentation template. When the user has the requirement of making the demonstration manuscript, the user can send the expected demonstration manuscript template and the input manuscript content terminal.
102. Selecting a plurality of alternative presentation templates corresponding to the presentation to be generated by adopting a preset multilayer convolutional neural network model based on the template information;
in this embodiment, the selecting process of the multilayer convolutional neural network model may specifically include:
2.1) analyzing a first requirement in the template information by using a preset multilayer convolutional neural network model to obtain corresponding first requirement information;
2.2) analyzing a second requirement in the template information by utilizing a multilayer convolutional neural network model according to the first requirement information and preset iteration parameters to obtain corresponding second requirement information;
2.3) carrying out iterative analysis processing on the first requirement information and the second requirement information to obtain a plurality of alternative presentation templates corresponding to the presentation to be generated;
in this embodiment, the preset multilayer convolutional neural network model may be a combination of methods such as a multi-convolutional neural network model, depth limit unsupervised learning, a ridge function, and sparse classification, and is used to analyze the first requirement to obtain the first requirement information. Subsequently, when the second requirement is analyzed by combining the first requirement information and the preset iteration parameter, the iteration parameter comprises: the maximum number of iterations, the number of iterations increases. For example, if the initial iteration number is 0, the first requirement is analyzed, then 1 is added to the iteration number, the first iteration is obtained, and so on, until the maximum iteration number is reached, and the second requirement information is obtained.
103. Adopting a preset text recognition model to carry out semantic recognition on the manuscript information to obtain the manuscript semantic information corresponding to the manuscript information;
in this embodiment, the semantic information of the document includes words, sentences, and semantic information between different sentences in the document information. Specifically, the text recognition model for semantic recognition of the document information may adopt a natural language processing model to deeply mine the document semantics from a word level, a sentence level, and a cross-sentence level for the document contents.
3.1) converting the manuscript information into corresponding word vectors and position vectors by using a preset text recognition model;
3.2) analyzing cross-sentence semantics corresponding to the target document information based on an information retrieval mining algorithm in the preset text recognition model, and determining an adjacent semantic vector corresponding to the document information based on the cross-sentence semantics;
and 3.3) taking the word vector, the position vector and the adjacent semantic vector as the manuscript semantic information corresponding to the manuscript information.
In this embodiment, the word vector may be obtained by using mainstream word2vec, glove and other models, the position vector may be obtained by using a sine wave in a transform, and a word Frequency-inverse document Frequency algorithm, that is, a TF-idf (term Frequency inverse transform) algorithm, may be applied to perform cross-sentence semantic analysis of the document information, so as to determine a corresponding street proximity semantic vector.
Using a TF-IDF algorithm to generate a TF-IDF attribute queue corresponding to the manuscript information so as to convert the manuscript information into a word vector matrix, and performing part-of-speech analysis on each word sequence to obtain a part-of-speech frequency attribute queue corresponding to the target manuscript information; identifying an entity of each sentence sequence in the manuscript information by using a named entity identification technology, and obtaining an entity co-occurrence frequency attribute queue corresponding to the manuscript information by combining with a preset knowledge map triple; combining the TF-IDF attribute queue, the part-of-speech frequency attribute queue and the entity co-occurrence frequency attribute queue to obtain a target candidate word set; finally, the corresponding output of each word is formed by adding a word vector, a position vector and an adjacent semantic vector, and the result is the manuscript semantic information.
104. Based on the semantic information of the manuscript, selecting an optimal presentation template with the highest matching degree with the style type of the presentation to be generated from all the alternative presentation templates;
in this embodiment, each presentation template (including the alternative presentation template) indicates the applicable scene type. The optimal presentation template refers to the candidate presentation template that is selected finally and is most matched with the style type of the current presentation to be generated.
In addition, after the document semantic information of the alternative presentation is obtained, the corresponding content tags can be further determined, and screening is performed in the alternative presentation by taking the content tags as a standard, so that an optimal presentation template which is most matched with the style type of the presentation to be generated is screened out. After the optimal presentation template is determined, the document information can be integrated into the optimal presentation template, and the required presentation is finally obtained.
It should be noted that, since there may be pictures, audio, or video in addition to the text in one presentation, when the text portion is processed, the user may further process the text portion on the presentation, and add the contents of the pictures, audio, and video, etc. to be added to the presentation, so as to complete the presentation.
105. And constructing a corresponding demonstration manuscript by adopting the manuscript information and the optimal demonstration manuscript template, and returning the demonstration manuscript to the terminal.
In this embodiment, the document information is divided into a plurality of areas, for example, by using the content tag obtained before, the text part of the document voice mapped by the content tag in the document information is regarded as belonging to the text area corresponding to the content tag. For example, the current document information is divided into 4 paragraphs, and the 3 content tags obtained by speech recognition include: product name, promotional planning and dissemination topics. After the content tag is obtained, marking the content participating in the content tag extraction in the document information, for example, if the content tag of "product name" is obtained based on the first segment, marking the first segment as a first text area; and marking the second paragraph and the third paragraph as a second character area if the content label of the 'propaganda plan' is obtained based on the second paragraph and the third paragraph, marking the fourth paragraph as a third character area if the content label of the 'propagation subject' is obtained based on the fourth paragraph, and directly obtaining the character area corresponding to the manuscript information corresponding to each content label in the optimal presentation manuscript template, and filling to construct and obtain the corresponding delay manuscript.
In the embodiment of the invention, the required information of the presentation to be generated is acquired, wherein a plurality of required alternative presentation templates can be extracted through the template information in the required information, the document semantic information in the required information can be identified through the document information in the required information, an optimal presentation template with the most matched style type is further screened by combining the semantics of the alternative presentation templates and the presentation, and the final delay document is constructed to automatically generate the presentation. Meanwhile, the template style, the semantics of the content of the manuscript and the required template are considered, so that the generation efficiency of the presentation is improved, and the quality normalization of the generation of the presentation is also ensured.
Referring to fig. 2, a second embodiment of the method for generating a presentation according to the embodiment of the present invention includes:
201. acquiring demand information of a presentation to be generated, which is sent by a terminal, wherein the demand information of the presentation comprises template information and corresponding document information;
202. analyzing a first requirement in the template information by using a preset multilayer convolutional neural network model to obtain corresponding first requirement information;
203. analyzing a second requirement in the template information by using a multilayer convolutional neural network model according to the first requirement information and preset iteration parameters to obtain corresponding second requirement information;
204. judging whether the first requirement information meets a preset deep analysis evaluation condition;
205. if the first requirement information does not meet the requirement, analyzing the second requirement information according to the multilayer convolutional neural network model to obtain information to be filled corresponding to the second requirement information;
206. selecting a plurality of alternative presentation templates corresponding to the presentation to be generated from a preset template library based on the information to be filled;
in this embodiment, whether the first requirement information meets the preset deep analysis evaluation condition may be determined according to the following formula:
wherein, i is 1,2, … …, m, j is 1,2, … …, n. T is 1,2, … …, q, Lk ijtThe time delay of the automatic generation of the presentation template corresponding to the first requirement information is Ckit, the accuracy of the automatic generation of the presentation template corresponding to the first requirement information is Wk ijtAnd automatically generating a calculation power and transmission broadband cost ratio for the presentation template corresponding to the first requirement information, wherein m, n and q are respectively preset parameters. And m, n and q are values of vectors in three directions of the multilayer convolutional neural network model respectively.
The multilayer convolutional neural network model is in a 1,2, … h multidimensional space, and the plurality of depth analysis schemes migrate to the direction determined by the more optimized presentation template matching recommendation scheme according to the multilayer convolutional neural network model. Wherein, multilayer convolution neuron network contains: a presentation automatic generation delay L, a presentation template automatic selection accuracy C, and a presentation template power-to-transmission bandwidth cost ratio W (W is a presentation template generation edge server power/transmission bandwidth cost).
And if the first requirement information does not meet the preset deep analysis evaluation condition, continuously analyzing the second requirement information according to the preset iteration parameter and the first requirement information to determine the information to be filled for generating the alternative demonstration manuscript template. Specifically, the information to be filled in may be determined by the following formula:
wherein M isk ijtInformation to be filled corresponding to the first required information, Lk ijtA time delay automatically generated for the presentation template corresponding to the second requirement information, Ck ijtAccuracy, W, of automatic generation of a presentation template corresponding to the second requirement informationk ijtRatio of computational effort to cost of transmission bandwidth automatically generated for the presentation template corresponding to the second requirement information, Bk+1 ijtSparse unsupervised learning factor, L, corresponding to the second demand informationminGMinimum time delay, C, automatically generated for historical presentation templatesmaxGMaximum accuracy, W, for automatic generation of historical presentation templatesminGThe minimum computational power to transport bandwidth cost ratio automatically generated for the historical presentation template.
207. Converting the manuscript information into corresponding word vectors and position vectors by using a preset text recognition model;
208. analyzing cross-sentence semantics corresponding to the manuscript information based on an information retrieval mining algorithm in a preset text recognition model, and determining an adjacent semantic vector corresponding to the manuscript information based on the cross-sentence semantics;
209. taking the word vector, the position vector and the adjacent semantic vector as the corresponding document semantic information of the document information;
210. based on the word vector, the position vector and the adjacent semantic vector, adopting a preset classification model to correspondingly extract the part-of-speech characteristics, the entity characteristics and the context characteristics of the manuscript information;
211. based on the part of speech characteristics, the entity characteristics and the context characteristics, identifying the style type of the presentation to be generated by using a classification model;
212. determining a content tag corresponding to the presentation to be generated according to the recognized style type, and selecting an optimal presentation template with the highest matching degree with the content tag from all the alternative presentation templates;
in this embodiment, the part-of-speech frequency information may be determined by the word vector to determine the part-of-speech features expressed in the document information; the single words, the words and the sentences in the manuscript information can be determined to be divided through the position vector so as to determine the semantic entity characteristics of each single word, word and sentence in the whole manuscript information; the context relationship among each single word, word and sentence obtained in the previous step can be determined through the context characteristics, the analysis of the actual expression significance of the whole manuscript information is facilitated, and the style type to be expressed by the manuscript information is further identified from the parts of speech, the entity and the context relationship, namely the style type of the generated demonstration manuscript to be generated.
In this embodiment, the content tags are keywords or keyword groups capable of indicating semantic information of the document, and may include product names, work plans, history reviews, thank you and words, and the like. After determining the document semantic information of the document information, the terminal also needs to condense the analyzed document semantic information into one or more content tags, that is, convert the mined document semantic information into one or more keywords as the content tags capable of indicating the document semantic information.
Because a copy of the document information usually expresses multiple meanings, that is, a plurality of content tags need to be marked, the text in the same copy of the document information can be divided into areas according to the content tags, and each area text is associated with one content tag.
213. Extracting information fields and file resource fields in the manuscript information, and determining corresponding areas of the information fields and the file resource fields in the optimal presentation manuscript template;
214. writing the information fields into corresponding areas in the optimal presentation template, searching corresponding file resources in a preset database according to the file resource fields, and inserting the file resources into the corresponding areas in the optimal presentation template to obtain corresponding presentations;
215. and adjusting the styles of the information fields and the file resources in the presentation according to preset style rules to obtain a corresponding presentation, and returning the presentation to the terminal.
In this embodiment, the file resource may include a picture, a text report, a report, and/or a voice. For the file resource of the picture, the size, position, arrangement and other patterns of the picture can be adjusted according to the number of the pictures added in the area corresponding to the file resource field. For example, when there are 1 picture in the area corresponding to the file resource field, the picture occupies the height of the entire area, when there are 2 or 3 pictures in the area, the pictures may be placed side by side, when there are 4 pictures in the area, the pictures may be arranged side by side in two rows, and so on.
In this embodiment, the information field includes a text, and may also include a processing status field of a flow task of the presentation, and may be inserted into the corresponding status prompt area, and the filling color of the corresponding area is adjusted according to the content of the processing status field. That is, the fill color of the area is adjusted according to the preset processing state, and the processing state of the flow task of the current presentation is represented by the fill color.
In the embodiment of the invention, when the alternative presentation template is generated based on the template information, the alternative presentation template can be screened through a multilayer convolutional neural network, and the presentation of the presentation between edge computers is ensured by considering the time delay, the accuracy, the calculation and the transmission broadband cost ratio generated by the template, so that each alternative presentation template can meet the deep analysis evaluation condition, the template can meet the condition of edge calculation, and the content of the presentation information is partitioned through the content tag, so that each partition can be filled with the content of the corresponding semantic manuscript, and the quality of the presentation is improved.
In the above description of the method for generating a presentation in the embodiment of the present invention, referring to fig. 3, a presentation generating apparatus in the embodiment of the present invention is described below, where an embodiment of the presentation generating apparatus in the embodiment of the present invention includes:
the acquiring module 301 is configured to acquire requirement information of a presentation to be generated, where the requirement information of the presentation includes template information and corresponding document information;
a selecting module 302, configured to select, based on the template information, a plurality of candidate presentation templates corresponding to the presentation to be generated by using a preset multilayer convolutional neural network model;
the identification module 303 is configured to perform semantic identification on the document information by using a preset text identification model to obtain document semantic information corresponding to the document information;
the matching module 304 is used for selecting an optimal presentation template with the highest matching degree with the style type of the presentation to be generated from all the alternative presentation templates based on the semantic information of the presentation;
and the constructing module 305 is configured to construct a corresponding presentation by using the document information and the optimal presentation template, and return the presentation to the terminal.
In the embodiment of the invention, the required information of the presentation to be generated is acquired, wherein a plurality of required alternative presentation templates can be extracted through the template information in the required information, the document semantic information in the required information can be identified through the document information in the required information, an optimal presentation template with the most matched style type is further screened by combining the semantics of the alternative presentation templates and the presentation, and the final delay document is constructed to automatically generate the presentation. Meanwhile, the template style, the semantics of the content of the manuscript and the required template are considered, so that the generation efficiency of the presentation is improved, and the quality normalization of the generation of the presentation is also ensured.
Referring to fig. 4, another embodiment of the presentation generation apparatus according to the embodiment of the present invention includes:
the acquiring module 301 is configured to acquire requirement information of a presentation to be generated, where the requirement information of the presentation includes template information and corresponding document information;
a selecting module 302, configured to select, based on the template information, a plurality of candidate presentation templates corresponding to the presentation to be generated by using a preset multilayer convolutional neural network model;
the identification module 303 is configured to perform semantic identification on the document information by using a preset text identification model to obtain document semantic information corresponding to the document information;
the matching module 304 is used for selecting an optimal presentation template with the highest matching degree with the style type of the presentation to be generated from all the alternative presentation templates based on the semantic information of the presentation;
and the constructing module 305 is configured to construct a corresponding presentation by using the document information and the optimal presentation template, and return the presentation to the terminal.
Specifically, the selecting module 302 includes:
the first analysis unit 3021 is configured to analyze a first requirement in the template information by using a preset multilayer convolutional neural network model to obtain corresponding first requirement information;
the second analysis unit 3022 is configured to analyze a second requirement in the template information by using the multilayer convolutional neural network model according to the first requirement information and a preset iteration parameter, so as to obtain corresponding second requirement information;
the iteration unit 3023 is configured to perform iterative analysis processing on the first requirement information and the second requirement information to obtain a plurality of alternative presentation templates corresponding to the presentation to be generated.
Specifically, the iteration unit 3023 is further configured to:
judging whether the first requirement information meets a preset deep analysis evaluation condition;
if the first requirement information does not meet the requirement, analyzing the second requirement information according to the multilayer convolutional neural network model to obtain information to be filled corresponding to the second requirement information;
and selecting a plurality of alternative presentation templates corresponding to the presentation to be generated from the preset template library based on the information to be filled.
Specifically, the identification module 303 includes:
a converting unit 3031, configured to convert the document information into corresponding word vectors and position vectors by using a preset text recognition model;
a retrieval mining unit 3032, configured to analyze cross-sentence semantics corresponding to the document information based on an information retrieval mining algorithm in the preset text recognition model, and determine an adjacent semantic vector corresponding to the document information based on the cross-sentence semantics;
a generating unit 3033 is configured to use the word vector, the position vector, and the adjacent semantic vector as document semantic information corresponding to the document information.
Specifically, the matching module 304 includes:
a classification unit 3041, based on the word vector, the position vector and the adjacent semantic vector, extracting part-of-speech characteristics, entity characteristics and context characteristics of the manuscript information by using a preset classification model; based on the part of speech characteristics, the entity characteristics and the context characteristics, identifying the style type of the presentation to be generated by using a classification model;
the matching unit 3042 determines a content tag corresponding to the presentation to be generated based on the identified style type, and selects an optimal presentation template with the highest matching degree with the content tag from the candidate presentation templates.
Specifically, the building block 305 includes:
the extracting unit 3051, configured to extract an information field and a file resource field in the document information, and determine a corresponding area of the information field and the file resource field in the optimal presentation template;
the inserting unit 3052, configured to write the information field into the corresponding region in the optimal presentation template, search for a corresponding file resource in the preset database according to the file resource field, and insert the file resource into the corresponding region in the optimal presentation template to obtain a corresponding presentation;
and the adjusting unit 3053, configured to adjust the style of the information field and the file resource in the presentation according to a preset style rule, to obtain a corresponding presentation.
In the embodiment of the invention, when the alternative presentation template is generated based on the template information, the alternative presentation template can be screened through a multilayer convolutional neural network, and the presentation of the presentation between edge computers is ensured by considering the time delay, the accuracy, the calculation and the transmission broadband cost ratio generated by the template, so that each alternative presentation template can meet the deep analysis evaluation condition, the template can meet the condition of edge calculation, and the content of the presentation information is partitioned through the content tag, so that each partition can be filled with the content of the corresponding semantic manuscript, and the quality of the presentation is improved.
Fig. 3 and fig. 4 describe the presentation generating apparatus in the embodiment of the present invention in detail from the perspective of the modular functional entity, and the presentation generating apparatus in the embodiment of the present invention is described in detail from the perspective of hardware processing.
Fig. 5 is a schematic structural diagram of a presentation generating apparatus according to an embodiment of the present invention, where the presentation generating apparatus 500 may have a relatively large difference due to different configurations or performances, and may include one or more processors (CPUs) 510 (e.g., one or more processors) and a memory 520, and one or more storage media 530 (e.g., one or more mass storage devices) for storing applications 533 or data 532. Memory 520 and storage media 530 may be, among other things, transient or persistent storage. The program stored in the storage medium 530 may include one or more modules (not shown), each of which may include a series of instruction operations for the presentation generation apparatus 500. Further, the processor 510 may be configured to communicate with the storage medium 530, and execute a series of instruction operations in the storage medium 530 on the presentation generating apparatus 500.
The presentation generating device 500 may also include one or more power supplies 540, one or more wired or wireless network interfaces 550, one or more input-output interfaces 560, and/or one or more operating systems 531, such as Windows server, Mac OS X, Unix, Linux, FreeBSD, and the like. Those skilled in the art will appreciate that the presentation generating device configuration shown in fig. 5 does not constitute a limitation of the presentation generating device and may include more or fewer components than those shown, or some components may be combined, or a different arrangement of components.
The present invention further provides a presentation generating device, where the computer device includes a memory and a processor, where the memory stores computer readable instructions, and the computer readable instructions, when executed by the processor, cause the processor to execute the steps of the presentation generating method in the foregoing embodiments.
The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium, and which may also be a volatile computer-readable storage medium, having stored therein instructions, which, when executed on a computer, cause the computer to perform the steps of the presentation generation method.
It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: various media capable of storing program codes, such as a usb disk, a removable hard disk, a read-only memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.
The above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A presentation generation method, comprising:
acquiring demand information of a presentation to be generated, which is sent by a terminal, wherein the demand information of the presentation comprises template information and corresponding document information;
selecting a plurality of alternative presentation templates corresponding to the presentation to be generated by adopting a preset multilayer convolutional neural network model based on the template information;
performing semantic recognition on the manuscript information by adopting a preset text recognition model to obtain manuscript semantic information corresponding to the manuscript information;
based on the semantic information of the manuscript, selecting an optimal presentation template with the highest matching degree with the style type of the presentation to be generated from each alternative presentation template;
and constructing a corresponding presentation by adopting the document information and the optimal presentation template, and returning the presentation to the terminal.
2. The method for generating a presentation according to claim 1, wherein the selecting, based on the template information, a plurality of candidate presentation templates corresponding to the presentation to be generated using a preset multilayer convolutional neural network model comprises:
analyzing a first requirement in the template information by using a preset multilayer convolutional neural network model to obtain corresponding first requirement information;
analyzing a second requirement in the template information by using the multilayer convolutional neural network model according to the first requirement information and preset iteration parameters to obtain corresponding second requirement information;
and performing iterative analysis processing on the first requirement information and the second requirement information to obtain a plurality of alternative presentation templates corresponding to the presentation to be generated.
3. The method for generating a presentation according to claim 2, wherein the performing iterative analysis processing on the first requirement information and the second requirement information to obtain a plurality of candidate presentation templates corresponding to the presentation to be generated comprises:
judging whether the first requirement information meets a preset depth analysis evaluation condition or not;
if the first requirement information does not meet the requirement, analyzing the second requirement information according to the multilayer convolutional neural network model to obtain information to be filled corresponding to the second requirement information;
and selecting a plurality of alternative presentation templates corresponding to the presentation to be generated from a preset template library based on the information to be filled.
4. The method for generating a presentation according to claim 1, wherein the performing semantic recognition on the document information by using a preset text recognition model to obtain document semantic information corresponding to the document information comprises:
converting the manuscript information into corresponding word vectors and position vectors by using a preset text recognition model;
analyzing cross-sentence semantics corresponding to the manuscript information based on an information retrieval mining algorithm in a preset text recognition model, and determining an adjacent semantic vector corresponding to the manuscript information based on the cross-sentence semantics;
and taking the word vector, the position vector and the adjacent semantic vector as the manuscript semantic information corresponding to the manuscript information.
5. The presentation generation method of claim 4, wherein selecting, based on the document semantic information, an optimal presentation template from the alternative presentation templates that matches the genre type of the presentation to be generated to a highest degree comprises:
based on the word vector, the position vector and the adjacent semantic vector, adopting a preset classification model to correspondingly extract part-of-speech characteristics, entity characteristics and context characteristics corresponding to the manuscript information;
based on the part of speech characteristics, the entity characteristics and the context characteristics, identifying the style type of the presentation to be generated by using the classification model;
and determining a content tag corresponding to the presentation to be generated according to the identified style type, and selecting an optimal presentation template with the highest matching degree with the content tag from the alternative presentation templates.
6. The method for generating a presentation according to any one of claims 1 to 5, wherein the constructing a corresponding presentation using the document information and the optimal presentation template comprises:
extracting an information field and a file resource field in the manuscript information, and determining a corresponding area of the information field and the file resource field in an optimal presentation template;
writing the information fields into corresponding areas in the optimal presentation template, searching corresponding file resources in a preset database according to the file resource fields, and inserting the file resources into the corresponding areas in the optimal presentation template to obtain corresponding presentations;
and adjusting the styles of the information fields and the file resources in the presentation according to a preset style rule to obtain a corresponding presentation.
7. A presentation generation apparatus, comprising:
the terminal comprises an acquisition module, a display module and a display module, wherein the acquisition module is used for acquiring the requirement information of the presentation to be generated, which is sent by the terminal, and the requirement information of the presentation comprises template information and corresponding document information;
the selection module is used for selecting a plurality of alternative presentation templates corresponding to the presentation to be generated by adopting a preset multilayer convolutional neural network model based on the template information;
the recognition module is used for performing semantic recognition on the manuscript information by adopting a preset text recognition model to obtain manuscript semantic information corresponding to the manuscript information;
the matching module is used for selecting the optimal presentation template with the highest matching degree with the style type of the presentation to be generated from each alternative presentation template based on the document semantic information;
and the construction module is used for constructing a corresponding presentation by adopting the manuscript information and the optimal presentation template and returning the presentation to the terminal.
8. The presentation generation apparatus of claim 7, wherein the selection module comprises:
the first analysis unit is used for analyzing a first requirement in the template information by using a preset multilayer convolutional neural network model to obtain corresponding first requirement information;
the second analysis unit is used for analyzing a second requirement in the template information by using the multilayer convolutional neural network model according to the first requirement information and preset iteration parameters to obtain corresponding second requirement information;
and the iteration unit is used for performing iteration analysis processing on the first requirement information and the second requirement information to obtain a plurality of alternative presentation templates corresponding to the presentation to be generated.
9. A presentation generating apparatus, characterized by comprising: a memory and at least one processor, the memory having instructions stored therein;
the at least one processor invokes the instructions in the memory to cause the presentation generation apparatus to perform the steps of the presentation generation method of any of claims 1-6.
10. A computer-readable storage medium having instructions stored thereon, wherein the instructions, when executed by a processor, implement the steps of the presentation generation method according to any one of claims 1-6.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111366306.5A CN114021541A (en) | 2021-11-18 | 2021-11-18 | Presentation generation method, device, equipment and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111366306.5A CN114021541A (en) | 2021-11-18 | 2021-11-18 | Presentation generation method, device, equipment and storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN114021541A true CN114021541A (en) | 2022-02-08 |
Family
ID=80065175
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111366306.5A Pending CN114021541A (en) | 2021-11-18 | 2021-11-18 | Presentation generation method, device, equipment and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114021541A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115396726A (en) * | 2022-08-01 | 2022-11-25 | 陈兵 | Demonstration manuscript generation system and method for commercial live broadcast |
-
2021
- 2021-11-18 CN CN202111366306.5A patent/CN114021541A/en active Pending
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115396726A (en) * | 2022-08-01 | 2022-11-25 | 陈兵 | Demonstration manuscript generation system and method for commercial live broadcast |
CN115396726B (en) * | 2022-08-01 | 2024-05-07 | 陈兵 | Demonstration manuscript generation system and method for commercial live broadcast |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111753060B (en) | Information retrieval method, apparatus, device and computer readable storage medium | |
CN108804641B (en) | Text similarity calculation method, device, equipment and storage medium | |
CN111444320B (en) | Text retrieval method and device, computer equipment and storage medium | |
CN111666401B (en) | Document recommendation method, device, computer equipment and medium based on graph structure | |
CN110929038B (en) | Knowledge graph-based entity linking method, device, equipment and storage medium | |
CN111539197B (en) | Text matching method and device, computer system and readable storage medium | |
CN109376222B (en) | Question-answer matching degree calculation method, question-answer automatic matching method and device | |
KR20200094627A (en) | Method, apparatus, device and medium for determining text relevance | |
CN112819023B (en) | Sample set acquisition method, device, computer equipment and storage medium | |
US12038970B2 (en) | Training image and text embedding models | |
US20170344822A1 (en) | Semantic representation of the content of an image | |
CN113298197B (en) | Data clustering method, device, equipment and readable storage medium | |
JPWO2019106965A1 (en) | Information processing equipment, information processing methods, and programs | |
KR20210129465A (en) | Apparatus for managing laboratory note and method for searching laboratory note using thereof | |
CN113051914A (en) | Enterprise hidden label extraction method and device based on multi-feature dynamic portrait | |
CN111078835A (en) | Resume evaluation method and device, computer equipment and storage medium | |
CN111274822A (en) | Semantic matching method, device, equipment and storage medium | |
CN112070550A (en) | Keyword determination method, device and equipment based on search platform and storage medium | |
CN112632258A (en) | Text data processing method and device, computer equipment and storage medium | |
CN114547257B (en) | Class matching method and device, computer equipment and storage medium | |
US20090327877A1 (en) | System and method for disambiguating text labeling content objects | |
CN110909768B (en) | Method and device for acquiring marked data | |
CN113821657A (en) | Artificial intelligence-based image processing model training method and image processing method | |
CN115640394A (en) | Text classification method, text classification device, computer equipment and storage medium | |
US20230282018A1 (en) | Generating weighted contextual themes to guide unsupervised keyphrase relevance models |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |