CN107480118B - Text editing method and device - Google Patents
Text editing method and device Download PDFInfo
- Publication number
- CN107480118B CN107480118B CN201710702071.XA CN201710702071A CN107480118B CN 107480118 B CN107480118 B CN 107480118B CN 201710702071 A CN201710702071 A CN 201710702071A CN 107480118 B CN107480118 B CN 107480118B
- Authority
- CN
- China
- Prior art keywords
- text
- editing
- semantic
- edited
- slots
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 52
- 238000000605 extraction Methods 0.000 claims description 55
- 239000013598 vector Substances 0.000 claims description 16
- 238000010276 construction Methods 0.000 claims description 12
- 230000003993 interaction Effects 0.000 claims description 10
- 238000002372 labelling Methods 0.000 claims description 8
- 230000011218 segmentation Effects 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 5
- 241000283084 Balaenoptera musculus Species 0.000 description 3
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000010420 art technique Methods 0.000 description 1
- 239000003337 fertilizer Substances 0.000 description 1
- 239000000945 filler Substances 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000002715 modification method Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/284—Lexical analysis, e.g. tokenisation or collocates
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Document Processing Apparatus (AREA)
- Machine Translation (AREA)
Abstract
The invention discloses a text editing method and a text editing device, wherein the method comprises the following steps: acquiring an original text to be edited; receiving user edited voice data; performing voice recognition on the voice data edited by the user to obtain an editing command; semantic understanding is carried out on the editing command so as to determine the value of the current text to be edited and the editing state; the editing state is used for describing whether the position of the current text to be edited in the original text can be determined or not; determining whether to execute the editing command according to the value of the current text to be edited and the editing state; if yes, editing the original text according to the editing command; otherwise, the user is interacted with to obtain more values corresponding to the editing state of the text to be edited. The invention can improve the text editing efficiency and ensure the accuracy of editing operation.
Description
Technical Field
The invention relates to the field of voice signal processing, in particular to a text editing method and device.
Background
The text editing mainly refers to carrying out font formats, typesetting processing, error modification and the like on the text, in the prior art, the text editing is generally completed by manually utilizing a keyboard and a mouse, firstly, the position of a cursor in an original text of the text to be edited is positioned, and then the corresponding text in the original text is edited. For example, for text data "weather is good for today's blue whale, i intend to go to that play for several days. "blue whale" is required to be changed into Nanjing, nanjing is also required to be thickened, weather is underlined, and the like.
The editing method needs manual operation, when more original texts need to be edited, the editing efficiency is low, errors are easy to occur, and the labor cost is high.
Disclosure of Invention
The invention provides a text editing method and device, which are used for solving the problems of low efficiency and high cost of the existing text editing.
Therefore, the invention provides the following technical scheme:
a text editing method, comprising:
Acquiring an original text to be edited;
Receiving user edited voice data;
performing voice recognition on the voice data edited by the user to obtain an editing command;
Semantic understanding is carried out on the editing command so as to determine the value of the current text to be edited and the editing state; the editing state is used for describing whether the position of the current text to be edited in the original text can be determined or not;
Determining whether the editing command can be executed or not according to the value of the current text to be edited and the editing state;
if yes, editing the original text according to the editing command;
Otherwise, the user is interacted with to obtain more values corresponding to the editing state of the text to be edited.
Preferably, the method further comprises: presetting the type of a semantic slot;
The semantic understanding of the editing command to determine the current text to be edited and the editing state comprises the following steps:
Carrying out semantic understanding on the editing command, and determining a semantic slot and a type contained in the editing command;
and determining the current text to be edited and the editing state according to the semantic slot and the type thereof contained in the editing command.
Preferably, the types of the semantic slots include a semantic slot for describing a user editing state, a semantic slot for describing a user editing operation, and a semantic slot for describing an editing text;
The semantic slots for describing user editing status include any one or more of:
the sequence semantic slot is used for describing the sequence of the text to be edited in the original text;
the type semantic slot is used for describing the type of the text to be edited;
the Chinese character structure semantic slot is used for describing a Chinese character structure of a text to be edited;
A radical semantic slot for describing a radical of a text to be edited;
The Chinese character description semantic slot is used for limiting a text to be edited;
The reference semantic slot is used for describing a reference text of the text to be edited in the original text;
a global position semantic slot, wherein the global position semantic slot refers to a global position of a text to be edited in an original text;
a relative position semantic slot, which refers to the relative position of the text to be edited in the original text;
The semantic slots for describing user editing operations include any one or more of: adding a semantic slot, deleting the semantic slot, modifying the font size semantic slot, modifying the font color semantic slot, modifying the font thickness semantic slot, and line feed semantic slot;
the semantic slots for describing the edited text include: and a semantic slot to be edited and a semantic slot after editing.
Preferably, the method further comprises:
Pre-constructing a semantic slot extraction model;
the semantic understanding of the editing command is carried out, and determining the semantic slots and the types thereof contained in the editing command comprises the following steps:
vectorizing each word in the editing command and then extracting features as semantic slots;
inputting the semantic slot extraction features into the semantic slot extraction model, and obtaining the semantic slots and the types thereof contained in the editing command according to the output of the semantic slot extraction model.
Preferably, the semantic slot extraction model is constructed as follows:
Setting an editing command template, wherein the editing command template comprises one or more semantic slots, and each semantic slot has a determined type;
Collecting a plurality of fill text for filling the semantic slots;
filling the filling text into the semantic slots to obtain a large amount of editing command data and using the editing command data as training data, and marking whether each word in the training data is a semantic slot or not and the type of the semantic slot;
Word segmentation is carried out on the training data, and word vectors of each word are obtained;
and training according to the word vector and the labeling information to obtain a semantic slot extraction model.
Preferably, the method further comprises:
pre-constructing a text editing grammar library, wherein the text editing grammar library comprises a plurality of grammar templates;
the semantic understanding of the editing command is carried out, and determining the semantic slots and the types thereof contained in the editing command comprises the following steps:
and matching the editing command with each grammar template in the text editing grammar library, and determining a semantic slot and the type thereof contained in the editing command according to a matching result.
Preferably, the text editing grammar library is constructed as follows:
Constructing a basic grammar template, wherein the basic grammar template comprises one or more semantic slots, and each semantic slot is of a determined type;
Expanding the basic grammar template to obtain an expanded grammar template;
and generating a text editing grammar library from the basic grammar template and the extended grammar template.
Preferably, the determining the value of the current text to be edited and the editing state according to the semantic slot and the type thereof included in the editing command includes:
And taking the text in the semantic slot to be edited contained in the editing command as the current text to be edited, and obtaining the value of the current editing state according to the text in the semantic slot for describing the editing state of the user contained in the editing command.
Preferably, the determining whether to execute the editing command according to the current text to be edited and the value of the editing state includes:
Performing fuzzy matching on the current text to be edited and the original text to obtain a primary matching position of the text to be edited in the original text;
If one of the primary matching positions is available, determining that the editing command can be executed;
Otherwise, adding the value of the current editing state into fuzzy matching of the current text to be edited and the original text to obtain a secondary matching position of the text to be edited in the original text;
If one of the secondary matching positions exists, determining that the editing command can be executed;
Otherwise, it is determined that the edit command cannot be executed.
A text editing apparatus comprising:
The text acquisition module is used for acquiring an original text to be edited;
the voice receiving module is used for receiving voice data edited by a user;
the voice recognition module is used for carrying out voice recognition on the voice data edited by the user to obtain an editing command;
the semantic understanding module is used for carrying out semantic understanding on the editing command so as to determine the value of the current text to be edited and the editing state; the editing state is used for describing whether the position of the current text to be edited in the original text can be determined or not;
the judging module is used for determining whether the editing command can be executed or not according to the values of the current text to be edited and the editing state;
the editing module is used for editing the original text according to the editing command after the judging module determines that the editing command can be executed;
And the interaction module is used for interacting with a user after the judgment module determines that the editing command can not be executed so as to acquire more values corresponding to the editing state of the text to be edited currently.
Preferably, the apparatus further comprises:
The semantic slot setting module is used for presetting the type of the semantic slot;
The semantic understanding module comprises:
The semantic slot information determining module is used for carrying out semantic understanding on the editing command and determining a semantic slot and a type thereof contained in the editing command;
and the editing information determining module is used for determining the current text to be edited and the editing state according to the semantic slots and the types thereof contained in the editing command.
Preferably, the types of the semantic slots include a semantic slot for describing a user editing state, a semantic slot for describing a user editing operation, and a semantic slot for describing an editing text;
The semantic slots for describing user editing status include any one or more of:
the sequence semantic slot is used for describing the sequence of the text to be edited in the original text;
the type semantic slot is used for describing the type of the text to be edited;
the Chinese character structure semantic slot is used for describing a Chinese character structure of a text to be edited;
A radical semantic slot for describing a radical of a text to be edited;
The Chinese character description semantic slot is used for limiting a text to be edited;
The reference semantic slot is used for describing a reference text of the text to be edited in the original text;
a global position semantic slot, wherein the global position semantic slot refers to a global position of a text to be edited in an original text;
a relative position semantic slot, which refers to the relative position of the text to be edited in the original text;
The semantic slots for describing user editing operations include any one or more of: adding a semantic slot, deleting the semantic slot, modifying the font size semantic slot, modifying the font color semantic slot, modifying the font thickness semantic slot, and line feed semantic slot;
the semantic slots for describing the edited text include: and a semantic slot to be edited and a semantic slot after editing.
Preferably, the apparatus further comprises:
The semantic slot extraction model construction module is used for constructing a semantic slot extraction model in advance;
the semantic slot information determining module includes:
the semantic slot extraction feature generation unit is used for vectorizing each word in the editing command and then taking the vectorized word as a semantic slot extraction feature;
the semantic slot information extraction unit is used for inputting the semantic slot extraction features into the semantic slot extraction model, and obtaining the semantic slots and the types thereof contained in the editing command according to the output of the semantic slot extraction model.
Preferably, the semantic slot extraction model building module includes:
The template setting unit is used for setting an editing command template, wherein the editing command template comprises one or more semantic slots, and each semantic slot is provided with a determined type;
a text collection unit for collecting a plurality of filled texts for filling the semantic slots;
The semantic slot filling unit is used for filling the filling text into the semantic slots to obtain a large amount of editing command data and taking the editing command data as training data, and marking whether each word in the training data is a semantic slot or not and the type of the semantic slot;
The word vector obtaining unit is used for segmenting the training data and obtaining word vectors of each word;
and the training unit is used for training to obtain a semantic slot extraction model according to the word vector and the labeling information.
Preferably, the apparatus further comprises:
the text editing grammar library construction module is used for pre-constructing a text editing grammar library, and the text editing grammar library comprises a plurality of grammar templates;
The semantic slot information determining module is specifically configured to match the editing command with each grammar template in the text editing grammar library, and determine a semantic slot and a type thereof included in the editing command according to a matching result.
Preferably, the text editing grammar library construction module includes:
The basic grammar template construction unit is used for constructing a basic grammar template, and the basic grammar template comprises one or more semantic slots, and each semantic slot is provided with a determined type;
the expansion unit is used for expanding the basic grammar template to obtain an expanded grammar template;
and the grammar library generating unit is used for generating a text editing grammar library from the basic grammar template and the extended grammar template.
Preferably, the editing information determining module is specifically configured to take a text in a semantic slot to be edited, which is included in the editing command, as a current text to be edited, and obtain a value of the current editing state according to the text in the semantic slot, which is included in the editing command and is used for describing the editing state of the user.
Preferably, the judging module includes:
The first matching unit is used for carrying out fuzzy matching on the current text to be edited and the original text to obtain a primary matching position of the text to be edited in the original text;
a judging unit configured to determine that the editing command can be executed when one of the primary matching positions is present; otherwise, triggering the second matching unit to perform fuzzy matching operation;
the second matching unit is used for adding the value of the current editing state into fuzzy matching of the current text to be edited and the original text to obtain a secondary matching position of the text to be edited in the original text;
The judging unit is further used for determining that the editing command can be executed when one secondary matching position exists; otherwise, it is determined that the edit command cannot be executed.
According to the text editing method and device provided by the embodiment of the invention, voice recognition is carried out on voice data edited by a user to obtain an editing command; then carrying out semantic understanding on the editing command to determine the current text to be edited and the editing state, determining whether to execute the editing command according to the editing state, and if so, directly editing the original text; otherwise, asking a question to the user, further determining the value of the editing state of the user, and further acquiring more accurate position information of the text to be edited in the original text. Because the corresponding user editing command is executed, the whole process is automatically executed, the text editing efficiency is greatly improved, the accuracy of editing operation is ensured, and the user experience is improved.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments described in the present application, and other drawings may be obtained according to these drawings for a person having ordinary skill in the art.
FIG. 1 is a flow chart of a text editing method according to an embodiment of the present invention;
FIG. 2 is a flow chart of constructing a semantic slot extraction model in an embodiment of the present invention;
FIG. 3 is a flow chart of constructing a text editing grammar library in an embodiment of the invention;
FIG. 4 is a schematic diagram of a text editing apparatus according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of a structure of a model-based semantic slot information determination module according to an embodiment of the present invention;
Fig. 6 is a schematic structural diagram of a grammar-based semantic slot information determining module according to an embodiment of the present invention.
Detailed Description
In order to make the solution of the embodiment of the present invention better understood by those skilled in the art, the embodiment of the present invention is further described in detail below with reference to the accompanying drawings and embodiments.
In consideration of convenience and high efficiency of voice interaction, the text data is edited by voice in a fast and efficient mode, so that the text editing method and device provided by the embodiment of the invention aim at the problem of low efficiency in the prior art that text editing is performed manually by using a keyboard and a mouse, realize automatic text editing by voice, and improve text editing efficiency; and under the condition that the position of the text to be edited in the original text can be accurately determined, the original text is edited, otherwise, more accurate position information of the text to be edited in the original text is acquired through further interaction, and the accuracy of editing operation is ensured.
As shown in fig. 1, a flowchart of a text editing method according to an embodiment of the present invention includes the following steps:
step 101, obtaining an original text to be edited.
The original text to be edited can be obtained according to voice input information or photographing input information of a user; for the user voice input information, the voice input information can be identified as a text by utilizing a voice identification technology, and the text is used as an original text to be edited; for the user to take a photograph to input information, a corresponding text can be obtained by utilizing an image recognition technology, and the text is used as an original text to be edited. Of course, the original text to be edited may also be text data directly provided by the user. The voice recognition and the image recognition may use some existing related technologies, and the embodiments of the present invention are not limited thereto.
Step 102, receiving user edited voice data.
The user editing voice data refers to data corresponding to voice spoken when the user modifies text data, if the original text to be edited is "help me to order a train ticket with a mixture of fertilizer to Nanjing", the user wants to modify Nanjing into Beijing, and the corresponding user editing voice data is "modify Nanjing into Beijing".
And step 103, performing voice recognition on the voice data edited by the user to obtain an editing command.
The editing command is a recognition text obtained by voice recognition. The speech recognition may employ related prior art techniques, which are not described in detail herein.
104, Carrying out semantic understanding on the editing command to determine the value of the current text to be edited and the editing state; the editing state is used for describing whether the position of the current text to be edited in the original text can be determined.
Firstly, carrying out semantic understanding on the editing command, and determining a semantic slot and a type thereof contained in the editing command; and then determining the current text to be edited and the editing state according to the semantic slots and the types thereof contained in the editing command.
The types of the semantic slots include a semantic slot for describing the user editing state, a semantic slot for describing the user editing operation, and a semantic slot for describing the editing text. These types of semantic slots are each described in detail below.
1) The semantic slots for describing user edit status include any one or more of the following:
A sequential semantic slot for describing the order of the text to be edited in the original text, such as first, second, last, penultimate, etc.;
the type semantic slot is used for describing the type of the text to be edited, such as Chinese characters, english, numbers, chinese capitals, arabic numbers, number strings and the like;
the Chinese character structure semantic slot is used for describing a Chinese character structure of a text to be edited, such as a bow, a Wen Daoliu, an ear east, a three-horizontal-vertical king and the like;
radical semantic slots for describing radicals of text to be edited, such as single person side, vertical knife side, wang Zibang, treasured header, word side, etc.;
a Chinese character description semantic slot for defining text to be edited, such as our me, science, china, etc.;
the reference semantic slot is used for describing a reference text of a text to be edited in an original text, such as a Beijing which I want to go to the back, wherein the "I want to go" is the reference text;
A global position semantic slot, wherein the global position semantic slot refers to a global position of a text to be edited in an original text, such as chapters, paragraphs, sentences and the like of the text to be edited in the original text; nanjing in the second sentence of the third paragraph;
The relative position semantic slot refers to the relative position of the text to be edited in the original text, and if the text to be edited is to go to the south Beijing at the back, the text to be edited is the relative position semantic slot at the back;
2) Semantic slots for describing user editing operations, such semantic slots for describing specific editing operations of a user, may include, for example, any one or more of the following: adding a semantic slot, deleting the semantic slot, modifying the font size semantic slot, modifying the font color semantic slot, modifying the font thickness semantic slot, and line feed semantic slot;
3) The semantic slots for describing the edit text include: and a semantic slot to be edited and a semantic slot after editing. The text in the semantic slot to be edited is the text to be edited, and the text in the semantic slot after editing is the text after editing.
In determining the semantic slots and their types contained in the edit command, a model-based method or a grammar-based method may be employed, and the specific process will be described in detail later.
After the semantic slots and the types thereof contained in the editing command are obtained, the current text to be edited and the editing state can be determined directly according to the semantic slots and the types thereof contained in the editing command. Specifically, the text in the semantic slot to be edited contained in the editing command is the current text to be edited; and obtaining the value of the current editing state according to the text in the semantic slot for describing the editing state of the user, which is contained in the editing command.
If the user edit command "modify me to be later Nanjing to Beijing", the extracted semantic slot type is "modify/O I/RefText to be later/RefText to be later/Nanjing/OriText modification/O to be/O Beijing/Text" of RelativePos, wherein O represents a non-semantic slot, oriText is a semantic slot to be edited, refText is a reference semantic slot, and Text is an edited semantic slot;
The current text to be edited is Nanjing, and the value of the editing state is:
Sequential semantic slots: null
Type semantic slots: null
Chinese character structure semantic slot: null
Radical semantic slots: null
Chinese character description semantic slot: null
Reference semantic slots: i want to go to
Global location semantic slots: null
Relative position semantic slots: rear face
Step 105, determining whether the editing command can be executed according to the value of the current text to be edited and the editing state; if yes, go to step 106; otherwise, step 107 is performed.
Specifically, fuzzy matching is carried out on the current edited text and the original text, so that primary matching positions of the text to be edited in the original text are obtained, and one or more primary matching positions are provided.
If only one primary matching position exists, determining the position of the current text to be edited in the original text as the primary matching position; since the position of the text currently to be edited can be accurately determined, in this case, a user editing command, i.e., editing the original text according to the editing command, can be performed.
If there are multiple matching positions at a time, the specific position of the text to be edited in the original text can be further determined according to the value of the current editing state. Specifically, the value of the current editing state can be directly added into fuzzy matching of the current text to be edited and the original text, and if only one secondary matching position is matched, the editing command can be determined to be executed; otherwise, the user needs to interact to obtain more values of the editing state until the accurate position of the current text to be edited in the original text is obtained.
And 106, editing the original text according to the editing command, and returning to the step 102.
If the user editing command is "modify me to be the subsequent Nanjing to Beijing", the text to be edited is determined to be the Nanjing, and the Nanjing at the position is directly modified to be the Beijing after the position of the Nanjing to be edited in the original text is determined; the specific modification method may employ existing related techniques, and will not be described in detail herein.
And step 107, interacting with a user to obtain more values corresponding to the editing state of the current text to be edited.
When the user interacts with the user, questions can be asked to the user according to the value condition of each semantic slot in the current user editing state and the fuzzy matching result after the semantic slot is added in the user editing state, for example, the 'Nanjing to be later modified to Beijing' is still a plurality of secondary matching results after the semantic slot is added in the current editing state, for example, the 'Nanjing' appears in the first section of the original text, the 'Nanjing', 'blue whale' appears in the third section of the original text; if the user can be asked of the problems of Nanjing of which section, jiangsu province can be done, and the like, the value of the semantic slot in more editing states can be continuously obtained; and adding the obtained semantic slot value of the user editing state after each questioning into the matching of the text to be edited and the original text until the accurate position of the text to be edited in the original text is obtained.
It should be noted that when the user edits the text to be edited, the number of interactions cannot be too large, otherwise, the user experience is poor, so that the number of interactions can be controlled within two times in order to improve the user experience; of course, multiple interactions may also be performed to obtain more accurate information, determining the user intent.
In the foregoing, in the embodiment of the present invention, when determining the semantic slots and the types thereof included in the editing command, a model-based method or a grammar-based method may be used, and the two methods are described in detail below.
(1) Model-based method for determining semantic slots and types of semantic slots contained in editing commands
In this method, a semantic slot extraction model needs to be built in advance, as shown in fig. 2, which is a flowchart for building the semantic slot extraction model in the embodiment of the present invention, and includes the following steps:
Step 21, setting an editing command template, wherein the editing command template comprises one or more semantic slots, and each semantic slot has a determined type.
If "change a to B" is an edit command template, where a and B are semantic slots, specifically, the type of semantic slot a is a semantic slot to be edited and the type of semantic slot B is a post-editing semantic slot.
A number of filler text is collected for filling the semantic slots, step 22.
And step 23, filling the filling text into the semantic slots to obtain a large amount of editing command data and using the editing command data as training data, and marking whether each word in the training data is a semantic slot and the type of the semantic slot.
In the information labeling, if the word in the edit command data is not a semantic slot, a uniform symbol representation, such as a character O, may be used to represent a non-semantic slot.
And step 24, word segmentation is carried out on the training data, and word vectors of each word are obtained.
And step 25, training according to the word vectors and the labeling information to obtain a semantic slot extraction model.
When the method is specifically used for training, word vectors of all words in the user editing command can be used as models to be input, whether the words in the editing command are semantic slots and the types of the semantic slots are predicted, prediction results and labeling information of the semantic slots of each editing command are utilized to update model parameters, and a semantic slot extraction model is obtained after training is finished.
Based on the semantic slot extraction model, when determining the semantic slots and the types thereof contained in the editing command, vectorizing each word in the editing command obtained in the step 103 to be used as a semantic slot extraction feature; and then inputting the semantic slot extraction features into the semantic slot extraction model, and obtaining the semantic slots and types thereof contained in the editing command according to the output of the semantic slot extraction model. If the current user editing command is "modifying Nanjing to Beijing", the semantic slot extraction model is utilized, and the obtained semantic slot extraction result is: "modifying/O Nanjing/OriText to/O Beijing/Text", wherein O represents a non-semantic slot, oriText is a semantic slot, the type is a semantic slot to be edited, and Text is a semantic slot, the type is an edited semantic slot.
(2) Grammar-based method for determining semantic slots and types contained in editing commands
In this method, a text editing grammar library needs to be built in advance, as shown in fig. 3, which is a flowchart for building the text editing grammar library in the embodiment of the invention, and includes the following steps:
Step 31, constructing a basic grammar template, wherein the basic grammar template comprises one or more semantic slots, and each semantic slot has a determined type.
And step 32, expanding the basic grammar template to obtain an expanded grammar template.
The basic grammar templates are as follows: modifying A into B, it can be expanded into 'modifying A into B, changing A into B, I want to change A into B, A into B', etc., thus obtaining expanded grammar template, the semantic slots in the expanded grammar template and the type thereof are the same as those in the corresponding basic grammar template.
And step 33, generating a text editing grammar library from the basic grammar template and the extended grammar template.
Based on the text editing grammar library, when determining the semantic slots and the types thereof contained in the editing command, the editing command obtained in the step 103 can be matched with each grammar template in the text editing grammar library, and the semantic slots and the types thereof contained in the editing command can be determined according to the matching result. The specific matching method may use existing related technologies, for example, an exact matching or fuzzy matching method may be used, which will not be described in detail herein.
According to the text editing method provided by the embodiment of the invention, voice recognition is carried out on voice data edited by a user to obtain an editing command; then carrying out semantic understanding on the editing command to determine the current text to be edited and the editing state, determining whether to execute the editing command according to the editing state, and if so, directly editing the original text; otherwise, asking a question to the user, further determining the value of the editing state of the user, and further acquiring more accurate position information of the text to be edited in the original text. Because the corresponding user editing command is executed, the whole process is automatically executed, the text editing efficiency is greatly improved, the accuracy of editing operation is ensured, and the user experience is improved.
Correspondingly, the embodiment of the invention also provides a text editing device, and as shown in fig. 4, the device is a schematic structure diagram.
In this embodiment, the apparatus comprises:
A text obtaining module 401, configured to obtain an original text to be edited;
A voice receiving module 402, configured to receive user edited voice data;
a voice recognition module 403, configured to perform voice recognition on the user edited voice data to obtain an editing command;
the semantic understanding module 404 is configured to perform semantic understanding on the editing command to determine a value of a current text to be edited and an editing state; the editing state is used for describing whether the position of the current text to be edited in the original text can be determined or not;
a judging module 405, configured to determine whether the editing command can be executed according to the value of the current text to be edited and the editing state;
the editing module 406 is configured to edit the original text according to the editing command after the judging module determines that the editing command can be executed;
And the interaction module 407 is configured to interact with a user to obtain more values corresponding to the editing state of the text to be edited currently after the judgment module determines that the editing command cannot be executed.
The above-mentioned voice receiving module 402 may specifically perform voice recognition on the user-edited voice data by using a conventional technology, to obtain a recognition text, where the recognition text is an editing command of the user.
The semantic understanding module 404 performs semantic understanding on the editing command, mainly based on preset various semantic slots and types thereof, to determine the current text to be edited and the editing state. The type of the semantic slot may be preset in the device, and of course, for convenience of operation, a corresponding semantic slot setting module (not shown) may also be set in the device to preset the type of the semantic slot.
In the embodiment of the present invention, the types of semantic slots include a semantic slot for describing a user editing state, a semantic slot for describing a user editing operation, and a semantic slot for describing an editing text, and specific meanings of the semantic slots of each type are described in detail above and are not described herein.
Accordingly, as shown in fig. 4, the semantic understanding module 404 may include a semantic slot information determination module 441 and an edit information determination module 442. The semantic slot information determining module 441 is configured to perform semantic understanding on the editing command, and determine a semantic slot and a type thereof included in the editing command; the editing information determining module 442 is configured to determine a current text to be edited and an editing state according to the semantic slots and the types thereof included in the editing command.
The semantic slot information determining module 441 may employ a model-based method or a grammar-based method in determining the semantic slots and types thereof included in the editing commands, which will be described in detail below.
As shown in fig. 5, a schematic diagram of a structure of the semantic slot information determining module when the model-based method determines the semantic slots and types thereof contained in the editing command is shown.
In this embodiment, the semantic slot extraction model 50 is pre-constructed by the semantic slot extraction model construction module 51, and the semantic slot extraction model construction module 51 may be used as a part of the apparatus of the embodiment of the present invention, or may be independent of the apparatus, which is not limited to the embodiment of the present invention.
Accordingly, in this embodiment, the semantic slot information determination module 441 includes:
A semantic slot extraction feature generation unit 52, configured to vectorize each word in the editing command and then use the vectorized word as a semantic slot extraction feature;
The semantic slot information extraction unit 53 is configured to input the semantic slot extraction feature into the semantic slot extraction model, and obtain a semantic slot and a type thereof included in the editing command according to an output of the semantic slot extraction model.
One specific structure of the semantic slot extraction model building module 51 includes the following units:
The template setting unit is used for setting an editing command template, wherein the editing command template comprises one or more semantic slots, and each semantic slot is provided with a determined type;
a text collection unit for collecting a plurality of filled texts for filling the semantic slots;
The semantic slot filling unit is used for filling the filling text into the semantic slots to obtain a large amount of editing command data and taking the editing command data as training data, and marking whether each word in the training data is a semantic slot or not and the type of the semantic slot;
The word vector obtaining unit is used for segmenting the training data and obtaining word vectors of each word;
and the training unit is used for training to obtain a semantic slot extraction model according to the word vector and the labeling information.
Referring to fig. 6, a schematic diagram of a structure of a semantic slot information determining module when determining a semantic slot and a type thereof included in an edit command according to a grammar-based method is shown.
In this embodiment, the text editing grammar library 60 is pre-built by the text editing grammar library building module 61, and the text editing grammar library building module 61 may be a part of the apparatus of the embodiment of the present invention, or may be independent of the apparatus, which is not limited to the embodiment of the present invention. The text editing grammar library 60 includes a plurality of grammar templates.
Accordingly, in this embodiment, the semantic slot information determining module 441 matches the editing command with each grammar template in the text editing grammar library, and determines the semantic slots and types thereof included in the editing command according to the matching result.
One specific structure of the text editing grammar library construction module 61 includes the following units:
The basic grammar template construction unit is used for constructing a basic grammar template, and the basic grammar template comprises one or more semantic slots, and each semantic slot is provided with a determined type;
the expansion unit is used for expanding the basic grammar template to obtain an expanded grammar template;
and the grammar library generating unit is used for generating a text editing grammar library from the basic grammar template and the extended grammar template.
The semantic slot information determination module 441 obtains the semantic slots and types thereof contained in the edit command based on whatever manner is described above,
The editing information determining module 442 may determine the value of the current text to be edited and the value of the editing state according to the semantic slot and the type thereof, specifically, take the text in the semantic slot to be edited included in the editing command as the current text to be edited, and obtain the value of the current editing state according to the text in the semantic slot for describing the editing state of the user included in the editing command.
The judging module in fig. 4 may specifically include the following units:
The first matching unit is used for carrying out fuzzy matching on the current text to be edited and the original text to obtain a primary matching position of the text to be edited in the original text;
A judging unit, configured to determine that the editing command can be executed when there is one of the primary matching positions, and at this time, output an instruction triggering the editing module 406 to start editing; otherwise, triggering the second matching unit to perform fuzzy matching operation;
the second matching unit is used for adding the value of the current editing state into fuzzy matching of the current text to be edited and the original text to obtain a secondary matching position of the text to be edited in the original text;
The judging unit is further configured to determine that the editing command can be executed when one of the secondary matching positions exists, and at this time, may output an instruction triggering the editing module 406 to start editing; otherwise, it is determined that the edit command cannot be executed, and at this time, an instruction triggering the interaction module 407 to interact may be output.
The specific operation of the editing module 406 for editing the original text according to the editing command may be implemented by using the prior art.
When the interaction module 407 interacts with the user, a question can be asked to the user according to the value condition of each semantic slot in the current user editing state and the fuzzy matching result after the semantic slot is added into the user editing state. And adding the obtained semantic slot value of the user editing state after each questioning into the matching of the text to be edited and the original text until the accurate position of the text to be edited in the original text is obtained.
The text editing device provided by the embodiment of the invention carries out voice recognition on voice data edited by a user to obtain an editing command; then carrying out semantic understanding on the editing command to determine the current text to be edited and the editing state, determining whether to execute the editing command according to the editing state, and if so, directly editing the original text; otherwise, asking a question to the user, further determining the value of the editing state of the user, and further acquiring more accurate position information of the text to be edited in the original text. Because the corresponding user editing command is executed, the whole process is automatically executed, the text editing efficiency is greatly improved, the accuracy of editing operation is ensured, and the user experience is improved.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments. In particular, for the device embodiments, since they are substantially similar to the method embodiments, the description is relatively simple, and reference is made to the description of the method embodiments for relevant points. The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.
The foregoing detailed description of embodiments of the invention, wherein specific embodiments are described, is provided to assist in understanding the methods and systems of the invention; meanwhile, as those skilled in the art will have variations in the specific embodiments and application scope in accordance with the ideas of the present invention, the present description should not be construed as limiting the present invention in view of the above.
Claims (18)
1. A text editing method, comprising:
Acquiring an original text to be edited;
Receiving user edited voice data;
performing voice recognition on the voice data edited by the user to obtain an editing command;
Semantic understanding is carried out on the editing command so as to determine the value of the current text to be edited and the editing state; the editing state is used for describing whether the position of the current text to be edited in the original text can be determined or not; the determining the value of the current text to be edited and the editing state comprises the following steps:
determining a semantic slot and a type thereof contained in the editing command; the semantic slots comprise semantic slots for describing the editing state of a user and semantic slots for describing editing texts;
Obtaining the current text to be edited based on the semantic slots used for describing the edited text in the semantic slot classification;
obtaining the value of the current editing state based on the semantic slots used for describing the editing state of the user in the semantic slot classification;
wherein the semantic slots for describing user editing status include any one or more of:
the sequence semantic slot is used for describing the sequence of the text to be edited in the original text;
the type semantic slot is used for describing the type of the text to be edited;
the Chinese character structure semantic slot is used for describing a Chinese character structure of a text to be edited;
A radical semantic slot for describing a radical of a text to be edited;
The Chinese character description semantic slot is used for limiting a text to be edited;
The reference semantic slot is used for describing a reference text of the text to be edited in the original text;
A global position semantic slot, wherein the global position semantic slot refers to a global position of a text to be edited in an original text;
a relative position semantic slot, which refers to the relative position of the text to be edited in the original text;
The semantic slots for describing the edited text include: a semantic slot to be edited and a semantic slot after editing;
Determining whether the editing command can be executed or not according to the value of the current text to be edited and the editing state;
if yes, editing the original text according to the editing command;
Otherwise, the user is interacted with to obtain more values corresponding to the editing state of the text to be edited.
2. The method according to claim 1, wherein the method further comprises: presetting the type of a semantic slot;
The semantic understanding of the editing command to determine the current text to be edited and the editing state comprises the following steps:
Carrying out semantic understanding on the editing command, and determining a semantic slot and a type contained in the editing command;
and determining the current text to be edited and the editing state according to the semantic slot and the type thereof contained in the editing command.
3. The method of claim 2, wherein the types of semantic slots further comprise semantic slots for describing user editing operations;
the semantic slots for describing user editing operations include any one or more of: adding semantic slots, deleting semantic slots, modifying font size semantic slots, modifying font color semantic slots, modifying font thickness semantic slots, line feed semantic slots, and line retract semantic slots.
4. The method according to claim 2, wherein the method further comprises:
Pre-constructing a semantic slot extraction model;
the semantic understanding of the editing command is carried out, and determining the semantic slots and the types thereof contained in the editing command comprises the following steps:
vectorizing each word in the editing command and then extracting features as semantic slots;
inputting the semantic slot extraction features into the semantic slot extraction model, and obtaining the semantic slots and the types thereof contained in the editing command according to the output of the semantic slot extraction model.
5. The method of claim 4, wherein the semantic slot extraction model is constructed as follows:
Setting an editing command template, wherein the editing command template comprises one or more semantic slots, and each semantic slot has a determined type;
Collecting a plurality of fill text for filling the semantic slots;
filling the filling text into the semantic slots to obtain a large amount of editing command data and using the editing command data as training data, and marking whether each word in the training data is a semantic slot or not and the type of the semantic slot;
Word segmentation is carried out on the training data, and word vectors of each word are obtained;
and training according to the word vector and the labeling information to obtain a semantic slot extraction model.
6. The method according to claim 2, wherein the method further comprises:
pre-constructing a text editing grammar library, wherein the text editing grammar library comprises a plurality of grammar templates;
the semantic understanding of the editing command is carried out, and determining the semantic slots and the types thereof contained in the editing command comprises the following steps:
and matching the editing command with each grammar template in the text editing grammar library, and determining a semantic slot and the type thereof contained in the editing command according to a matching result.
7. The method of claim 6, wherein the text editing grammar library is constructed in the following manner:
Constructing a basic grammar template, wherein the basic grammar template comprises one or more semantic slots, and each semantic slot is of a determined type;
Expanding the basic grammar template to obtain an expanded grammar template;
and generating a text editing grammar library from the basic grammar template and the extended grammar template.
8. A method according to claim 3, wherein determining the value of the current text to be edited and the editing status according to the semantic slots and the type thereof included in the editing command comprises:
And taking the text in the semantic slot to be edited contained in the editing command as the current text to be edited, and obtaining the value of the current editing state according to the text in the semantic slot for describing the editing state of the user contained in the editing command.
9. The method according to any one of claims 1 to 8, wherein determining whether to execute the editing command according to the current text to be edited and the value of the editing state includes:
Performing fuzzy matching on the current text to be edited and the original text to obtain a primary matching position of the text to be edited in the original text;
If one of the primary matching positions is available, determining that the editing command can be executed;
Otherwise, adding the value of the current editing state into fuzzy matching of the current text to be edited and the original text to obtain a secondary matching position of the text to be edited in the original text;
If one of the secondary matching positions exists, determining that the editing command can be executed;
Otherwise, it is determined that the edit command cannot be executed.
10. A text editing apparatus, comprising:
The text acquisition module is used for acquiring an original text to be edited;
the voice receiving module is used for receiving voice data edited by a user;
the voice recognition module is used for carrying out voice recognition on the voice data edited by the user to obtain an editing command;
The semantic understanding module is used for carrying out semantic understanding on the editing command so as to determine the value of the current text to be edited and the editing state; the editing state is used for describing whether the position of the current text to be edited in the original text can be determined or not; the determining the value of the current text to be edited and the editing state comprises the following steps:
determining a semantic slot and a type thereof contained in the editing command; the semantic slots comprise semantic slots for describing the editing state of a user and semantic slots for describing editing texts;
Obtaining the current text to be edited based on the semantic slots used for describing the edited text in the semantic slot classification;
obtaining the value of the current editing state based on the semantic slots used for describing the editing state of the user in the semantic slot classification;
wherein the semantic slots for describing user editing status include any one or more of:
the sequence semantic slot is used for describing the sequence of the text to be edited in the original text;
the type semantic slot is used for describing the type of the text to be edited;
the Chinese character structure semantic slot is used for describing a Chinese character structure of a text to be edited;
A radical semantic slot for describing a radical of a text to be edited;
The Chinese character description semantic slot is used for limiting a text to be edited;
The reference semantic slot is used for describing a reference text of the text to be edited in the original text;
A global position semantic slot, wherein the global position semantic slot refers to a global position of a text to be edited in an original text;
a relative position semantic slot, which refers to the relative position of the text to be edited in the original text;
The semantic slots for describing the edited text include: a semantic slot to be edited and a semantic slot after editing;
the judging module is used for determining whether the editing command can be executed or not according to the values of the current text to be edited and the editing state;
the editing module is used for editing the original text according to the editing command after the judging module determines that the editing command can be executed;
And the interaction module is used for interacting with a user after the judgment module determines that the editing command can not be executed so as to acquire more values corresponding to the editing state of the text to be edited currently.
11. The apparatus of claim 10, wherein the apparatus further comprises:
The semantic slot setting module is used for presetting the type of the semantic slot;
The semantic understanding module comprises:
The semantic slot information determining module is used for carrying out semantic understanding on the editing command and determining a semantic slot and a type thereof contained in the editing command;
and the editing information determining module is used for determining the current text to be edited and the editing state according to the semantic slots and the types thereof contained in the editing command.
12. The apparatus of claim 11, wherein the types of semantic slots further comprise semantic slots for describing user editing operations;
the semantic slots for describing user editing operations include any one or more of: adding semantic slots, deleting semantic slots, modifying font size semantic slots, modifying font color semantic slots, modifying font thickness semantic slots, line feed semantic slots, and line retract semantic slots.
13. The apparatus of claim 11, wherein the apparatus further comprises:
The semantic slot extraction model construction module is used for constructing a semantic slot extraction model in advance;
the semantic slot information determining module includes:
the semantic slot extraction feature generation unit is used for vectorizing each word in the editing command and then taking the vectorized word as a semantic slot extraction feature;
the semantic slot information extraction unit is used for inputting the semantic slot extraction features into the semantic slot extraction model, and obtaining the semantic slots and the types thereof contained in the editing command according to the output of the semantic slot extraction model.
14. The apparatus of claim 13, wherein the semantic slot extraction model building module comprises:
The template setting unit is used for setting an editing command template, wherein the editing command template comprises one or more semantic slots, and each semantic slot is provided with a determined type;
a text collection unit for collecting a plurality of filled texts for filling the semantic slots;
The semantic slot filling unit is used for filling the filling text into the semantic slots to obtain a large amount of editing command data and taking the editing command data as training data, and marking whether each word in the training data is a semantic slot or not and the type of the semantic slot;
The word vector obtaining unit is used for segmenting the training data and obtaining word vectors of each word;
and the training unit is used for training to obtain a semantic slot extraction model according to the word vector and the labeling information.
15. The apparatus of claim 11, wherein the apparatus further comprises:
the text editing grammar library construction module is used for pre-constructing a text editing grammar library, and the text editing grammar library comprises a plurality of grammar templates;
The semantic slot information determining module is specifically configured to match the editing command with each grammar template in the text editing grammar library, and determine a semantic slot and a type thereof included in the editing command according to a matching result.
16. The apparatus of claim 15, wherein the text editing grammar library construction module comprises:
The basic grammar template construction unit is used for constructing a basic grammar template, and the basic grammar template comprises one or more semantic slots, and each semantic slot is provided with a determined type;
the expansion unit is used for expanding the basic grammar template to obtain an expanded grammar template;
and the grammar library generating unit is used for generating a text editing grammar library from the basic grammar template and the extended grammar template.
17. The apparatus of claim 12, wherein the device comprises a plurality of sensors,
The editing information determining module is specifically configured to take a text in a semantic slot to be edited, which is included in the editing command, as a current text to be edited, and obtain a value of the current editing state according to the text in the semantic slot, which is included in the editing command and is used for describing the editing state of the user.
18. The apparatus according to any one of claims 10 to 17, wherein the determining module comprises:
The first matching unit is used for carrying out fuzzy matching on the current text to be edited and the original text to obtain a primary matching position of the text to be edited in the original text;
a judging unit configured to determine that the editing command can be executed when one of the primary matching positions is present; otherwise, triggering the second matching unit to perform fuzzy matching operation;
the second matching unit is used for adding the value of the current editing state into fuzzy matching of the current text to be edited and the original text to obtain a secondary matching position of the text to be edited in the original text;
The judging unit is further used for determining that the editing command can be executed when one secondary matching position exists; otherwise, it is determined that the edit command cannot be executed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710702071.XA CN107480118B (en) | 2017-08-16 | 2017-08-16 | Text editing method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710702071.XA CN107480118B (en) | 2017-08-16 | 2017-08-16 | Text editing method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107480118A CN107480118A (en) | 2017-12-15 |
| CN107480118B true CN107480118B (en) | 2024-05-31 |
Family
ID=60600480
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710702071.XA Active CN107480118B (en) | 2017-08-16 | 2017-08-16 | Text editing method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107480118B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110321534B (en) * | 2018-03-28 | 2023-11-24 | 科大讯飞股份有限公司 | Text editing method, device, equipment and readable storage medium |
| CN109388792B (en) * | 2018-09-30 | 2020-11-24 | 珠海格力电器股份有限公司 | Text processing method, device, equipment, computer equipment and storage medium |
| CN109087645B (en) * | 2018-10-24 | 2021-04-30 | 科大讯飞股份有限公司 | Decoding network generation method, device, equipment and readable storage medium |
| CN111046674B (en) * | 2019-12-20 | 2024-05-31 | 科大讯飞股份有限公司 | Semantic understanding method and device, electronic equipment and storage medium |
| CN113378520B (en) * | 2021-04-20 | 2024-04-09 | 北京灵伴即时智能科技有限公司 | Text editing method and system |
| CN113723050A (en) * | 2021-07-23 | 2021-11-30 | 内蒙古工业大学 | Word document processing system and method based on voice interaction |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103106061A (en) * | 2013-03-05 | 2013-05-15 | 北京车音网科技有限公司 | Voice input method and device |
| CN105869632A (en) * | 2015-01-22 | 2016-08-17 | 北京三星通信技术研究有限公司 | Speech recognition-based text revision method and device |
| CN106534548A (en) * | 2016-11-17 | 2017-03-22 | 科大讯飞股份有限公司 | Voice error correction method and device |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8719014B2 (en) * | 2010-09-27 | 2014-05-06 | Apple Inc. | Electronic device with text error correction based on voice recognition data |
-
2017
- 2017-08-16 CN CN201710702071.XA patent/CN107480118B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103106061A (en) * | 2013-03-05 | 2013-05-15 | 北京车音网科技有限公司 | Voice input method and device |
| CN105869632A (en) * | 2015-01-22 | 2016-08-17 | 北京三星通信技术研究有限公司 | Speech recognition-based text revision method and device |
| CN106534548A (en) * | 2016-11-17 | 2017-03-22 | 科大讯飞股份有限公司 | Voice error correction method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107480118A (en) | 2017-12-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107480118B (en) | Text editing method and device | |
| CN108090400B (en) | Image text recognition method and device | |
| CN111738251B (en) | Optical character recognition method and device fused with language model and electronic equipment | |
| CN108959242B (en) | Target entity identification method and device based on part-of-speech characteristics of Chinese characters | |
| CN107039034B (en) | Rhythm prediction method and system | |
| CN106503231B (en) | Search method and device based on artificial intelligence | |
| CN110765759B (en) | Intention recognition method and device | |
| CN106227721B (en) | Chinese Prosodic Hierarchy forecasting system | |
| WO2017075957A1 (en) | Recognition rate determining method and device | |
| CN105975625A (en) | Chinglish inquiring correcting method and system oriented to English search engine | |
| CN103680498A (en) | Speech recognition method and speech recognition equipment | |
| CN103646019A (en) | Method and device for fusing multiple machine translation systems | |
| CN108052499A (en) | Text error correction method, device and computer-readable medium based on artificial intelligence | |
| DE602004004310T2 (en) | System with combined statistical and rule-based grammar model for speech recognition and understanding | |
| CN106648819A (en) | Internationalized code conversion method based on editor | |
| CN106297764A (en) | A kind of multilingual mixed Chinese language treatment method and system | |
| CN109977398A (en) | A kind of speech recognition text error correction method of specific area | |
| CN114528394B (en) | Text triple extraction method and device based on mask language model | |
| CN112216267B (en) | Prosody prediction method, device, equipment and storage medium | |
| CN107748744A (en) | Method and device for establishing drawing box knowledge base | |
| CN109992770A (en) | A kind of Laotian name entity recognition method based on combination neural net | |
| CN106610937A (en) | Information theory-based Chinese automatic word segmentation method | |
| CN106156013B (en) | A kind of two-part machine translation method that regular collocation type phrase is preferential | |
| CN112560510A (en) | Translation model training method, device, equipment and storage medium | |
| Serrano et al. | Interactive handwriting recognition with limited user effort |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |