CN109783677A - Answering method, return mechanism, electronic equipment and computer readable storage medium - Google Patents
Answering method, return mechanism, electronic equipment and computer readable storage medium Download PDFInfo
- Publication number
- CN109783677A CN109783677A CN201910052612.8A CN201910052612A CN109783677A CN 109783677 A CN109783677 A CN 109783677A CN 201910052612 A CN201910052612 A CN 201910052612A CN 109783677 A CN109783677 A CN 109783677A
- Authority
- CN
- China
- Prior art keywords
- picture
- semantic
- input information
- user
- reply
- 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
Abstract
The present invention provides a kind of answering method, return mechanism, electronic equipment and computer readable storage medium, the answering method includes: semantic processes step, processing based on semantic understanding is carried out to the input information of user's input, to generate the semantic processes result of input information；And reply and obtain step, by carrying out the semantic processes result based on the retrieval process for replying picture database, to obtain the picture of matching input information, using as it is to be output to user for the reply of input information.The present invention is based on the semantic processes to input information as a result, picture concerned is retrieved from replying in picture database, so that solving the problems, such as can not intelligent replying picture in human-computer interaction.
Description
Technical field
The present embodiments relate to answering method, return mechanism, electronic equipments and computer-readable based on text interaction
Storage medium, and in particular, to one kind is matched based on input content intelligent replying and input content in human-computer interaction scene
The method of picture.
Background technique
Currently in the interaction of the person to person based on text (for example, Internet chat, operational line consulting), hair picture is (outstanding
It is cardon, expression figure, entertaining figure etc.) interactive and interest means are promoted as a kind of, it is widely used by user.
And man-machine text interactive system has started to be widely used in all trades and professions at present.Currently in the interaction of man-machine automation text,
The content that machine is replied is replied both for the text that context is made, or simple expression fixed on a small quantity is (such as
Emoji expression).
Lack a kind of method at present, can be realized under man-machine automation text interaction scenarios, can be inputted according to user
Content replys appropriate picture.
Summary of the invention
The present invention is had developed in view of the above problem in the prior art.The present invention is intended to provide one kind can be man-machine automatic
Change the method that intelligent replying is realized in interactive process.
The first aspect of the present invention provides the answering method in a kind of human-computer interaction, and the answering method includes: semantic place
Step is managed, the processing based on semantic understanding is carried out to the input information of user's input, to generate the semantic processes knot of input information
Fruit；And reply and obtain step, by come based on the retrieval process for replying picture database to the semantic processes result
Obtain the picture that can match input information, using as it is to be output to user for the reply of input information.
Preferably, it is described reply picture database by by picture and reflection image content picture describe it is associated in a manner of
Manage picture.
Preferably, picture description be by any one of artificial addition manner and machine learning mode mode come
It generates.
Preferably, it includes: similarity calculation step that the reply, which obtains step, calculates the semantic processes result and each figure
Semantic similarity between piece description；And result return step, it will be associated with the semantic similarity for meeting predetermined condition
Picture is as the picture for matching the input information.
Preferably, the reply obtains step further include: and sequence step is ranked up calculated semantic similarity,
Wherein, in the result return step, by with N before ranking and meet the associated picture of the semantic similarity of predetermined condition and make
For the picture for matching the input information, the N is positive integer.
Preferably, the answering method is before the semantic processes step further include: pre-treatment step inputs user
Input information pre-processed, wherein in the semantic processes step, pretreated information is carried out based on semantic reason
The processing of solution, to generate the semantic processes result of input information.
Preferably, the pretreatment includes the removal of punctuate, the removal of meaningless symbol, the removal of messy code, capital and small letter turn
Change and simplified traditional font conversion at least one of.
Preferably, the processing based on semantic understanding includes participle, synonym replacement, the filtering of invalid word, mood point
At least one of analysis and contextual relation.
Preferably, the input information is the text information of user's input, by the voice messaging of conversion user's input
The text information of generation and the text information of user's input are composed with the text information for being converted into user speech information
One of text information.
The second aspect of the present invention provides the return mechanism in a kind of human-computer interaction, and the return mechanism includes: semantic place
Unit is managed, the processing based on semantic understanding is carried out to the input information of user's input, to generate the semantic processes knot of input information
Fruit；And obtaining unit is replied, by come based on the retrieval process for replying picture database to the semantic processes result
Obtain the picture that can match input information, using as it is to be output to user for the reply of input information.
Preferably, it is described reply picture database by by picture and reflection image content picture describe it is associated in a manner of
Manage picture.
Preferably, picture description be by any one of artificial addition manner and machine learning mode mode come
It generates.
Preferably, the reply obtaining unit includes: similarity calculated, calculates the semantic processes result and each figure
Semantic similarity between piece description；And result return unit, it will be associated with the semantic similarity for meeting predetermined condition
Picture is as the picture for matching the input information.
Preferably, the reply obtaining unit further include: sequencing unit is ranked up calculated semantic similarity,
Wherein, the result return unit using with N before ranking and meet the associated picture of the semantic similarity of predetermined condition as
Picture with the input information, the N are positive integer.
Preferably, the return mechanism further include: pretreatment unit pre-processes the input information of user's input,
Wherein, the semantic processing unit carries out the processing based on semantic understanding to pretreated information, to generate input information
Semantic processes result.
Preferably, the pretreatment includes the removal of punctuate, the removal of meaningless symbol, the removal of messy code, capital and small letter turn
Change and simplified traditional font conversion at least one of.
Preferably, it is described it is semantic-based processing include participle, synonym replacement, the filtering of invalid word, mood analyze with
And at least one of contextual relation.
Preferably, the input information is the text information of user's input, by the voice messaging of conversion user's input
The text information of generation and the text information of user's input are composed with the text information for being converted into user speech information
One of text information.
The third aspect of the present invention provides a kind of electronic equipment for human-computer interaction comprising receives the input of user
The receiving unit of information, storage reply storage unit, return mechanism and the output unit of picture database, the return mechanism
Include: semantic processing unit, the processing based on semantic understanding is carried out to the input information received by receiving unit, it is defeated to generate
Enter the semantic processes result of information；And obtaining unit is replied, by carrying out the semantic processes result based on reply picture
The retrieval process of database, to obtain the picture for matching the input information, to be directed to input letter to user as to be output
The reply of breath, wherein the output unit will reply the reply for input information of obtaining unit acquisition, export to user.
The fourth aspect of the present invention provides a kind of computer readable storage medium, stores computer program, the calculating
Machine program is realized and is included the steps that in above-mentioned answering method when being executed by processor.
According to the present invention, by the analysis described to picture, the intelligent replying in human-computer interaction process is realized, it can be with
For solve the problems, such as can not intelligent replying picture.In addition, the present invention also may be implemented to remove from user under some interaction scenarios
It needs actively to go, for example, user wants to reply a picture expression in chat, to need to look into before by the cumbersome of figure tool is searched
Picture is looked for, appropriate figure is found and is replied again；And based on the present invention, user can directly input the text for thinking expression, machine
The picture met can be replied according to word content real-time prompting.In the prior art, such as in wechat some tables are loaded with down
Feelings can also recommend relevant picture in the text that input is intended by.In this respect, the difference of the application and the prior art
Point is: the picture in the application is downloaded in advance without user；The application can reply picture database by updating, and come real-time
Ground updates picture in preset time.
Detailed description of the invention
In order to more clearly explain the technical solutions in the embodiments of the present application, make required in being described below to embodiment
Attached drawing is briefly described, it should be apparent that, the accompanying drawings in the following description is only some implementations as described in this application
Example, without creative efforts, can also be according to these attached drawings for this field or those of ordinary skill
Obtain other attached drawings.
Fig. 1 is to illustrate the figure of the hardware construction of the return mechanism in the present invention.
Fig. 2 is to illustrate the flow chart of answering method according to a first embodiment of the present invention.
Fig. 3 is to illustrate the flow chart of the reply acquisition step of answering method according to a first embodiment of the present invention.
Fig. 4 is to illustrate the block diagram of the modular structure of return mechanism according to first embodiment.
Fig. 5 is to illustrate the flow chart of the reply acquisition step of answering method according to a second embodiment of the present invention.
Fig. 6 is to illustrate the block diagram of the modular structure of return mechanism according to the second embodiment.
Fig. 7 is to illustrate the flow chart of answering method according to a third embodiment of the present invention.
Fig. 8 is to illustrate the block diagram of the modular structure of return mechanism according to the third embodiment.
Fig. 9 A to Fig. 9 D is to illustrate the example display screen of answering method of the invention.
Specific embodiment
Hereinafter describe the embodiment of the present invention in detail with reference to the accompanying drawings.It should be appreciated that following embodiments and unawareness
The figure limitation present invention, also, about the means according to the present invention solved the problems, such as, it is not absolutely required to be retouched according to following embodiments
The whole combinations for the various aspects stated.For simplicity, to identical structure division or step, identical label or mark have been used
Number, and the description thereof will be omitted.
[hardware configuration of return mechanism]
Fig. 1 is the figure for showing the hardware construction of the return mechanism in the present invention.In the present embodiment, using smart phone as
The example of return mechanism provides description.Although it is noted that instantiating smart phone in the present embodiment as return mechanism
1000, but it is clear that without being limited thereto, return mechanism of the invention can be mobile terminal (smart phone, smartwatch, intelligent hand
Ring, music player devices), laptop, tablet computer, PDA (personal digital assistant), picture unit, printer or be
The various devices such as the internet device (such as digital camera, refrigerator, television set etc.) with recovery function.
Firstly, block diagram referring to Fig.1 describes the hardware configuration of return mechanism 1000 (2000,3000).In addition, in this implementation
Following construction is described as example in example, but return mechanism of the invention is not limited to construction shown in FIG. 1.
Return mechanism 1000 includes the input interface 101, CPU 102, ROM 103, RAM being connected to each other via system bus
105, storage device 106, output interface 104, communication unit 107 and short-range wireless communication unit 108 and display unit 109.
Input interface 101 is the interface executed instruction for receiving data and function that user is inputted, and be for via
The operating unit (not shown) of such as microphone, key, button or touch screen receives the data inputted from user and operational order
Interface.It note that the display unit 109 being described later on and operating unit can be at least partly integrated, also, for example, can
To be to carry out picture output in same picture and receive the construction of user's operation.
CPU 102 is system control unit, and generally comprehensively controls return mechanism 1000.In addition, for example, CPU
102 carry out the display control of the display unit 109 of return mechanism 1000.ROM 103 stores such as tables of data that CPU 102 is executed
With the fixed data of control program and operating system (OS) program etc..In the present embodiment, each control stored in ROM 103
Processing procedure sequence, for example, dispatched under the management of the OS stored in ROM 103, task switching and interrupt processing etc. it is soft
Part executes control.
RAM 105 is constructed such as SRAM (static random access memory), DRAM as needing backup power source.This
In the case of, RAM 105 can store the significant data of control variable of program etc. in a non-volatile manner.In addition, RAM 105
Working storage and main memory as CPU 102.
Storage device 106 stores model trained in advance (for example, word error correction mode, physical model, Rank model, semanteme
Model etc.), the reply picture database for being retrieved and the application journey for executing answering method according to the present invention
Sequence etc..It note that reply picture database here also can store in the external device (ED) of such as server.In addition, storage
The storage of device 106 such as send via communication unit 107 and communication device (not shown)/received information sends/
The various information that the various programs and these programs of receiving control program etc. use.In addition, storage device 106 is also stored back into
Setting information, management data of return mechanism 1000 of apparatus for coating 1000 etc..
Output interface 104 is the display picture for being controlled display unit 109 to show information and application program
The interface in face.Display unit 109 is for example constructed by LCD (liquid crystal display).Have such as by being arranged on display unit 109
The soft keyboard of the key of numerical value enter key, mode setting button, decision key, cancel key and power key etc. can receive single via display
The input from the user of member 109.
Return mechanism 100 is via communication unit 107 for example, by side wireless communications such as Wi-Fi (Wireless Fidelity) or bluetooth
Method executes data communication with external device (ED) (not shown).
In addition, return mechanism 1000 can also via short-range wireless communication unit 108, in short-range with outside
Device etc. is wirelessly connected and executes data communication.And short-range wireless communication unit 108 by with communication unit 107 not
Same communication means is communicated.It is, for example, possible to use its communication range bluetooth shorter than the communication means of communication unit 107 is low
Communication means of the power consumption (BLE) as short-range wireless communication unit 108.In addition, as short-range wireless communication unit 108
Communication means, for example, it is also possible to use NFC (near-field communication) or Wi-Fi perception (Wi-Fi Aware).
[first embodiment]
[answering method according to first embodiment]
Answering method according to the present invention can by the CPU 102 of return mechanism 1000 read be stored in ROM 103 or
Control program on storage device 106 or via communication unit 107 from the network being connect by network with return mechanism 1000
Server (not shown) and the control program downloaded are realized.
Before carrying out answering method according to the present invention, need first to prepare to reply picture database.Detailed process is as follows:
(1) it collects and arranges picture materials: can be crawled from public network, collect picture materials (including expression picture, openly
Chat record picture, network picture etc.).
(2) the picture description for reflecting the image content is generated for each picture materials: each picture needs one section to figure
(that is, " picture description ") is briefly described in piece content, is able to reflect the content and meaning that picture is included.Picture description can be with
It is generated by any one of artificial addition manner and machine learning mode mode.For example, for the picture for having text
Material extracts word content using OCR technique, to form picture description.In addition, marked in advance based on a batch
Picture can train deep neural network, obtain the model for automatically generating label character ability, add text for new picture
It describes (picture description).Wherein, verbal description can add in picture, the new picture of combination producing；It can also be retouched with association
Picture and picture description are associated by the mode stated.
(3) reply picture database is added in the picture description of picture and the content for reflecting the picture in a manner of associated
In.
As described above, the picture replied in picture database can be individual picture, it is also possible to verbal description
Picture.
In another implementation, a Duan Wenben can be inputted, is scanned for from network according to text to obtain and text
This corresponding picture, and when database is added, by text and the associated addition of picture, in this way without being generated again for picture
The step of picture of reaction image content describes.
Next, being illustrated in conjunction with Fig. 2 to Fig. 4 to answering method according to a first embodiment of the present invention.Wherein, Fig. 2
It is to illustrate the flow chart of answering method according to a first embodiment of the present invention；Fig. 3 is to illustrate reply party according to first embodiment
The reply of method obtains the flow chart of step；Fig. 4 is to illustrate the block diagram of the modular structure of return mechanism according to first embodiment.
As shown in Fig. 2, firstly, carrying out the input information of user's input based on semanteme in semantic processes step S101
The processing of understanding, to generate the semantic processes result of input information.Here the processing based on semantic understanding includes segmenting, being synonymous
Word replacement, the filtering of invalid word, mood analysis and contextual relation etc..Here, participle, which refers to, will in short be cut into language
The small segment of adopted meaning, for example, " I am especially happy today " can be cut into " I ", " today ", " special " and " happy ".
It is replaced, can be replaced with a series of synonyms (such as the synonym " happiness " and " happy " etc. of " happy ") by synonym
It changes.Invalid word filtering refers to the word for filtering some pairs of judgement contribution very littles, for example " special " in upper sentence can be removed, and leaves
Kernel keyword " I ", " today " and " happy ".Mood analysis, which refers to according to verbal description, to be judged, represented mood is just
Mood in face or negative, such as upper sentence is positive.Contextual relation refers in the presence of context, such as
" yesterday, the Lakers won ", then " Lakers " can also be used as keyword above, be considered in subsequent calculating.
Next, being obtained in step S102 replying, by carrying out the semantic processes result based on reply picture number
According to the retrieval process in library, to obtain the picture that can match input information, to be directed to input information to user as to be output
Reply.Preferably, as shown in figure 3, reply acquisition step S102 further comprises: similarity calculation step S1021 calculates institute
Semantic similarity between predicate justice processing result and the description of each picture；And result return step S1022, it will make a reservation for meeting
Picture of the associated picture of the semantic similarity of condition as the user to be replied that can match input information.
In the following, being illustrated by taking following situation as an example to above-mentioned steps.
For example, being directed to the input information " I am especially happy today " of user, the processing in S101 through the above steps is obtained
Semantic processes result (keyword " I ", " today " and " happy ").It is " non-for a certain picture description replied in picture database
Chang Xingfu " (keyword is " happiness "), in similarity calculation step S1021, all keywords all can be by the side word2vec
Method is converted to term vector, and according to the methods of the features such as unigram/bigram/trigram/ngram and word matching, calculates language
Adopted similarity, obtains the similarity score between one 0 to 1, and the more high then similarity of score is higher.In this example, " happy "
" happiness " can match, and calculate term vector similarity, and " I " and " today " the two words are not involved in similarity calculation.Optionally
Ground can also describe the picture to carry out identical based on semantic reason with for user's input information before calculating similarity
The processing of solution, such as the input information of user and the mood of picture description are judged by model trained in advance, and only in mood
Similarity calculation is carried out in the case where identical (for example, belonging to positive, negative or neutral mood), for example, in the examples described above,
Input information and the picture description of user belongs to positive mood.
Then, in result return step S1022, picture associated with the semantic similarity for meeting predetermined condition is made
For the picture that can match the user to be replied for inputting information.For example, can be by calculated semantic similarity and predetermined threshold
Be compared, and the semantic similarity be greater than or equal to predetermined threshold in the case where, will reply picture database in the language
The adopted associated picture of similarity as can match input information user to be replied picture (for example, display can be passed through
Picture is shown to user by unit 109).When the picture for being greater than predetermined threshold is more than or equal to 1, user semantic can be recommended
Similarity be more than or equal to predetermined threshold (such as 0.7) picture selected for user, in order to increase the experience of user, when be greater than etc.
When the picture of predetermined threshold is less than 1, at this moment recommends the higher picture of user semantic similarity and selected for user.This
In, using the relationship between semantic similarity and predetermined threshold as the example of predetermined condition, certainly, according to actual needs, make a reservation for
Condition can also use other forms.In addition, predetermined threshold here be according to model training, verifying and test the case where and
It is predetermined, to ensure that accuracy rate is higher.
In addition, when there are multiple results for meeting predetermined condition (being greater than or be equal to predetermined threshold), according to specific
Usage scenario, it is highest as a result, returning to one of result at random that similarity can be returned.
In addition, the information of user's input can be believed for the text information of user's input, by the voice of conversion user's input
Breath and generate text information and user input text information formed with the text combination for being converted into user speech information
One of text information.
Answering method according to a first embodiment of the present invention is carried out by the input information inputted to user based on semantic reason
The processing of solution is to generate the semantic processes of input information as a result, and by carrying out the semantic processes result based on reply picture
The retrieval process of database, to obtain the picture that can match input information, to be directed to input letter to user as to be output
The reply of breath can obtain following technical effect: the intelligent replying in human-computer interaction process be realized, so that solving can not intelligence
The problem of picture can be replied.It needs alternatively, it is also possible to removing user under some interaction scenarios actively to go by searching figure tool
It is cumbersome.
[modular structure of return mechanism according to first embodiment]
Fig. 4 is to illustrate the block diagram of the modular structure of return mechanism according to first embodiment.As shown in figure 4, return mechanism
1000 include semantic processing unit 1101 and reply obtaining unit 1102.
Specifically, the input information that semantic processing unit 1101 inputs user carries out the processing based on semantic understanding,
To generate the semantic processes result of input information；Obtaining unit 1102 is replied by being based on back to the semantic processes result
The retrieval process of multiple picture database, to obtain the picture that can match input information, using as being directed to user to be output
Input the reply of information.The reply obtaining unit further comprises: similarity calculated 11021, calculates at the semanteme
Manage the semantic similarity between result and the description of each picture；And result return unit 11022, by with the language that meets predetermined condition
Picture of the adopted associated picture of similarity as the user to be replied that can match input information.
[second embodiment]
[answering method according to the second embodiment]
Answering method according to a second embodiment of the present invention is illustrated below with reference to Fig. 5 and Fig. 6.Wherein, Fig. 5 is example
Show that the reply of answering method according to a second embodiment of the present invention obtains the flow chart of step.
As shown in figure 5, the difference of answering method according to the second embodiment and answering method according to first embodiment exists
In, reply obtain step in, increase sequence step S2023.
Specifically, being ranked up in sequence step S2023 to semantic similarity calculated in step S1021.
Next, in result return step S2022, will with (such as N before ranking, N are positive integer) in the top and meet predetermined item
Picture of the associated picture of the semantic similarity of part as the user to be replied that can match input information.
Answering method according to a second embodiment of the present invention, by being ranked up to semantic similarity, and by similarity compared with
High picture replies to user, can obtain following technical effect: allowing users to obtain optimal recovery result.
[modular structure of return mechanism according to the second embodiment]
Fig. 6 is to illustrate the block diagram of the modular structure of return mechanism according to the second embodiment.As shown in fig. 6, according to second
The return mechanism 2000 of embodiment and the difference of return mechanism 1000 according to first embodiment are, are replying obtaining unit
Sequencing unit 12023 is increased in 1102.
Specifically, sequencing unit 12023 is ranked up calculated semantic similarity, and result return unit
12022 using with N before ranking and meet the associated picture of the semantic similarity of predetermined condition as input information can be matched
The picture of user to be replied, the N are positive integer.
[3rd embodiment]
[answering method according to the third embodiment]
Answering method according to a third embodiment of the present invention is illustrated below with reference to Fig. 7 and Fig. 8.
Fig. 7 is to illustrate the flow chart of answering method according to a third embodiment of the present invention.
As shown in fig. 7, the difference of answering method according to the third embodiment and answering method according to first embodiment exists
In increasing pre-treatment step S300.
Specifically, being pre-processed in pre-treatment step S300 to the input information of user's input.Here, pre- place
Reason includes the removal of punctuate, the removal of meaningless symbol, the removal of messy code, capital and small letter conversion and the conversion of simplified traditional font etc..
Next, the processing based on semantic understanding is carried out to pretreated information in semantic processes step S301, with
Generative semantics processing result.
Answering method according to the third embodiment before carrying out based on the processing of semantic understanding by being pre-processed, energy
The enough information inputted to user is corrected, to improve the accuracy of later retrieval.
[modular structure of return mechanism according to the third embodiment]
Fig. 8 is to illustrate the block diagram of the modular structure of return mechanism according to the third embodiment.As shown in figure 8, according to third
The return mechanism 3000 of embodiment and the difference of return mechanism 1000 according to first embodiment are, increase pretreatment unit
1300。
Specifically, the input information that pretreatment unit 1300 inputs user pre-processes, and semantic processes list
First 1301 pairs of pretreated information carry out the processing based on semantic understanding, with generative semantics processing result.
In addition, Fig. 9 A- Fig. 9 D is to illustrate the example display screen of answering method of the invention.For example, as shown in Figure 9 A,
For the input information " very good " of user, picture is retrieved from reply image data and is returned to user, wherein picture
The text " I feels fine " of lower section is the text carried in original picture, and the text " sincerity feels fine " above picture indicates
Verbal description of the addition in picture when preparing to reply picture database.
In addition, the present invention also provides a kind of electronic equipments for human-computer interaction comprising receive the input letter of user
The receiving unit of breath stores the storage unit for replying picture database, the return mechanism in any of the above-described embodiment and will return
Output unit of the reply output for input information that apparatus for coating obtains to user.Wherein, receiving unit can be by Fig. 1
Input interface 101 realizes that storage unit can realize that output unit can be by defeated in Fig. 1 by the storage device 106 in Fig. 1
Outgoing interface 104 is realized.
Although exemplary embodiments describe the present invention for reference above, above-described embodiment is only to illustrate this hair
Bright technical concepts and features, it is not intended to limit the scope of the present invention.What all Spirit Essences according to the present invention were done
Any equivalent variations or modification, should be covered by the protection scope of the present invention.
Present invention further provide that
A1. the answering method in a kind of human-computer interaction, the answering method include:
Semantic processes step carries out the processing based on semantic understanding to the input information of user's input, to generate input letter
The semantic processes result of breath；And
It replys and obtains step, by carrying out the semantic processes result based on the retrieval process for replying picture database,
Obtain the picture for matching the input information, using as it is to be output to user for the reply of input information.
A2. answering method according to a1, wherein the reply picture database with by picture and reflection image content
Picture describe associated mode and manage picture.
A3. the answering method according to A2, wherein the picture description is by artificial addition manner and machine learning
Any one of mode mode generates.
A4. the answering method according to A2, wherein the reply obtains step and includes:
Similarity calculation step calculates the semantic similarity between the semantic processes result and the description of each picture；And
As a result return step, using picture associated with the semantic similarity for meeting predetermined condition as the matching input
The picture of information.
A5. answering method according to a4, wherein the reply obtains step further include:
Sequence step is ranked up calculated semantic similarity,
Wherein, in the result return step, by with N before ranking and to meet the semantic similarity of predetermined condition associated
Picture as match it is described input information picture, the N be positive integer.
A6. answering method according to a1, the answering method is before the semantic processes step further include:
Pre-treatment step pre-processes the input information of user's input,
Wherein, in the semantic processes step, the processing based on semantic understanding is carried out to pretreated information, with life
At the semantic processes result of input information.
A7. the answering method according to A6, wherein the pretreatment includes that the removal of punctuate, meaningless symbol are gone
Remove, the removal of messy code, capital and small letter conversion and simplified traditional font conversion at least one of.
A8. answering method according to a1, wherein the processing based on semantic understanding is replaced including participle, synonym
It changes, the filtering of invalid word, mood analysis and at least one of contextual relation.
A9. answering method according to a1, the input information is the text information of user's input, by conversion user
The voice messaging of input and the text information generated and the text information of user's input with user speech information is converted into
One of the text information that text information is composed.
B10. the return mechanism in a kind of human-computer interaction, the return mechanism include:
Semantic processing unit carries out the processing based on semantic understanding to the input information of user's input, to generate input letter
The semantic processes result of breath；And
Obtaining unit is replied, by carrying out the semantic processes result based on the retrieval process for replying picture database,
Obtain the picture for matching the input information, using as it is to be output to user for the reply of input information.
B11. return mechanism according to b10, wherein the reply picture database with will picture and reflection picture in
The picture of appearance describes associated mode and manages picture.
B12. the return mechanism according to B11, wherein the picture description is by artificial addition manner and engineering
Any one of habit mode mode generates.
B13. the return mechanism according to B11, wherein the reply obtaining unit includes:
Similarity calculated calculates the semantic similarity between the semantic processes result and the description of each picture；And
As a result return unit, using picture associated with the semantic similarity for meeting predetermined condition as the matching input
The picture of information.
B14. return mechanism according to b13, wherein the reply obtaining unit further include:
Sequencing unit is ranked up calculated semantic similarity,
Wherein, the result return unit by with N before ranking and meet the associated figure of the semantic similarity of predetermined condition
Piece is positive integer as the picture for matching the input information, the N.
B15. return mechanism according to b10, the return mechanism further include:
Pretreatment unit pre-processes the input information of user's input,
Wherein, the semantic processing unit carries out the processing based on semantic understanding to pretreated information, defeated to generate
Enter the semantic processes result of information.
B16. the return mechanism according to B15, wherein the pretreatment includes the removal of punctuate, meaningless symbol
At least one of removal, the removal of messy code, capital and small letter conversion and the conversion of simplified traditional font.
B17. return mechanism according to b10, wherein the semantic-based processing is including segmenting, synonym is replaced,
At least one of filtering, mood analysis and contextual relation of invalid word.
B18. return mechanism according to b10, the input information are the text information of user's input, pass through to convert and use
Family input voice messaging and the text information generated and user input text information be converted by user speech information
One of the text information that is composed of text information.
C19. a kind of electronic equipment for human-computer interaction comprising receive the receiving unit of the input information of user, deposit
Storage unit, return mechanism and the output unit of picture database are replied in storage, and the return mechanism includes:
Semantic processing unit carries out the processing based on semantic understanding to the input information received by receiving unit, with life
At the semantic processes result of input information；And
Obtaining unit is replied, by carrying out the semantic processes result based on the retrieval process for replying picture database,
Obtain the picture for matching the input information, using as it is to be output to user for the reply of input information,
Wherein, the output unit will reply the reply for input information of obtaining unit acquisition, export to user.
D20. a kind of computer readable storage medium stores computer program, and the computer program is by processor
When execution, realizes and include the steps that in the answering method according to any one of A1 to A9.
Claims (10)
1. the answering method in a kind of human-computer interaction, the answering method include:
Semantic processes step carries out the processing based on semantic understanding to the input information of user's input, to generate input information
Semantic processes result；And
It replys and obtains step, by carrying out the semantic processes result based on the retrieval process for replying picture database, to obtain
The picture that the input information must be matched, using as it is to be output to user for the reply of input information.
2. answering method according to claim 1, wherein the reply picture database with will picture and reflection picture in
The picture of appearance describes associated mode and manages picture.
3. answering method according to claim 2, wherein the picture description is by artificial addition manner and engineering
Any one of habit mode mode generates.
4. answering method according to claim 2, wherein the reply obtains step and includes:
Similarity calculation step calculates the semantic similarity between the semantic processes result and the description of each picture；And
As a result return step, using picture associated with the semantic similarity for meeting predetermined condition as the matching input information
Picture.
5. answering method according to claim 4, wherein the reply obtains step further include:
Sequence step is ranked up calculated semantic similarity,
Wherein, in the result return step, by with N before ranking and meet the associated figure of the semantic similarity of predetermined condition
Piece is positive integer as the picture for matching the input information, the N.
6. answering method according to claim 1, the answering method is before the semantic processes step further include:
Pre-treatment step pre-processes the input information of user's input,
Wherein, in the semantic processes step, the processing based on semantic understanding is carried out to pretreated information, it is defeated to generate
Enter the semantic processes result of information.
7. answering method according to claim 6, wherein the pretreatment includes the removal of punctuate, meaningless symbol
At least one of removal, the removal of messy code, capital and small letter conversion and the conversion of simplified traditional font.
8. the return mechanism in a kind of human-computer interaction, the return mechanism include:
Semantic processing unit carries out the processing based on semantic understanding to the input information of user's input, to generate input information
Semantic processes result；And
Obtaining unit is replied, by carrying out the semantic processes result based on the retrieval process for replying picture database, to obtain
The picture that the input information must be matched, using as it is to be output to user for the reply of input information.
9. a kind of electronic equipment for human-computer interaction comprising receive the receiving unit of the input information of user, storage is replied
Storage unit, return mechanism and the output unit of picture database, the return mechanism include:
Semantic processing unit carries out the processing based on semantic understanding to the input information received by receiving unit, defeated to generate
Enter the semantic processes result of information；And
Obtaining unit is replied, by carrying out the semantic processes result based on the retrieval process for replying picture database, to obtain
The picture that the input information must be matched, using as it is to be output to user for the reply of input information,
Wherein, the output unit will reply the reply for input information of obtaining unit acquisition, export to user.
10. a kind of computer readable storage medium, stores computer program, the computer program is being executed by processor
When, it realizes and includes the steps that in answering method according to any one of claim 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910052612.8A CN109783677A (en) | 2019-01-21 | 2019-01-21 | Answering method, return mechanism, electronic equipment and computer readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910052612.8A CN109783677A (en) | 2019-01-21 | 2019-01-21 | Answering method, return mechanism, electronic equipment and computer readable storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109783677A true CN109783677A (en) | 2019-05-21 |
Family
ID=66501667
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910052612.8A Pending CN109783677A (en) | 2019-01-21 | 2019-01-21 | Answering method, return mechanism, electronic equipment and computer readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109783677A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111147356A (en) * | 2019-12-30 | 2020-05-12 | 上海连尚网络科技有限公司 | Information processing method and device |
| CN112612877A (en) * | 2020-12-16 | 2021-04-06 | 平安普惠企业管理有限公司 | Multi-type message intelligent reply method, device, computer equipment and storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102567509A (en) * | 2011-12-26 | 2012-07-11 | 中国科学院自动化研究所 | Method and system for instant messaging with visual messaging assistance |
| CN104866116A (en) * | 2015-03-25 | 2015-08-26 | 百度在线网络技术（北京）有限公司 | Method and device for outputting expression information |
| WO2016068455A1 (en) * | 2014-10-30 | 2016-05-06 | 주식회사 플런티코리아 | Method and system for providing adaptive keyboard interface and response input method using adaptive keyboard linked with conversation content |
| CN106021273A (en) * | 2016-04-25 | 2016-10-12 | 北京光年无限科技有限公司 | Method and system for processing information facing question answering robot |
| CN106570106A (en) * | 2016-11-01 | 2017-04-19 | 北京百度网讯科技有限公司 | Method and device for converting voice information into expression in input process |
| US20170169822A1 (en) * | 2015-12-14 | 2017-06-15 | Hitachi, Ltd. | Dialog text summarization device and method |
| CN108681541A (en) * | 2018-01-17 | 2018-10-19 | 百度在线网络技术（北京）有限公司 | Image searching method, device and computer equipment |
-
2019
- 2019-01-21 CN CN201910052612.8A patent/CN109783677A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102567509A (en) * | 2011-12-26 | 2012-07-11 | 中国科学院自动化研究所 | Method and system for instant messaging with visual messaging assistance |
| WO2016068455A1 (en) * | 2014-10-30 | 2016-05-06 | 주식회사 플런티코리아 | Method and system for providing adaptive keyboard interface and response input method using adaptive keyboard linked with conversation content |
| CN104866116A (en) * | 2015-03-25 | 2015-08-26 | 百度在线网络技术（北京）有限公司 | Method and device for outputting expression information |
| US20170169822A1 (en) * | 2015-12-14 | 2017-06-15 | Hitachi, Ltd. | Dialog text summarization device and method |
| CN106021273A (en) * | 2016-04-25 | 2016-10-12 | 北京光年无限科技有限公司 | Method and system for processing information facing question answering robot |
| CN106570106A (en) * | 2016-11-01 | 2017-04-19 | 北京百度网讯科技有限公司 | Method and device for converting voice information into expression in input process |
| CN108681541A (en) * | 2018-01-17 | 2018-10-19 | 百度在线网络技术（北京）有限公司 | Image searching method, device and computer equipment |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111147356A (en) * | 2019-12-30 | 2020-05-12 | 上海连尚网络科技有限公司 | Information processing method and device |
| CN112612877A (en) * | 2020-12-16 | 2021-04-06 | 平安普惠企业管理有限公司 | Multi-type message intelligent reply method, device, computer equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11062270B2 (en) | Generating enriched action items | |
| US20190272269A1 (en) | Method and system of classification in a natural language user interface | |
| JP6971853B2 (en) | Automatic extraction of commitments and requests from communication and content | |
| US11482212B2 (en) | Electronic device for analyzing meaning of speech, and operation method therefor | |
| US9081411B2 (en) | Rapid development of virtual personal assistant applications | |
| US9489625B2 (en) | Rapid development of virtual personal assistant applications | |
| US11538468B2 (en) | Using semantic frames for intent classification | |
| US20190251417A1 (en) | Artificial Intelligence System for Inferring Grounded Intent | |
| US20160306800A1 (en) | Reply recommendation apparatus and system and method for text construction | |
| US20140257794A1 (en) | Semantic Re-Ranking of NLU Results in Conversational Dialogue Applications | |
| WO2014047270A1 (en) | Method and apparatus for predicting intent in ivr using natural language queries | |
| US20140257793A1 (en) | Communicating Context Across Different Components of Multi-Modal Dialog Applications | |
| CN112154465A (en) | Method, device and equipment for learning intention recognition model | |
| CN108228808B (en) | Method and device for determining hot event, storage medium and electronic equipment | |
| CN112287085B (en) | Semantic matching method, system, equipment and storage medium | |
| CN112579733A (en) | Rule matching method, rule matching device, storage medium and electronic equipment | |
| CN116615727A (en) | Keyword data augmentation tool for natural language processing | |
| EP4060971A1 (en) | Generating action items during a conferencing session | |
| CN114444462B (en) | Model training method and man-machine interaction method and device | |
| CN109783677A (en) | Answering method, return mechanism, electronic equipment and computer readable storage medium | |
| US20230350929A1 (en) | Method and system for generating intent responses through virtual agents | |
| CN109002498B (en) | Man-machine conversation method, device, equipment and storage medium | |
| CN114546326A (en) | Virtual human sign language generation method and system | |
| CN109829157B (en) | Text emotion presenting method, text emotion presenting device and storage medium | |
| CN115427993A (en) | Updating constraints of computerized assistant operations |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20200804 Address after: 518000 Nanshan District science and technology zone, Guangdong, Zhejiang Province, science and technology in the Tencent Building on the 1st floor of the 35 layer Applicant after: TENCENT TECHNOLOGY (SHENZHEN) Co.,Ltd. Address before: 100029, Beijing, Chaoyang District new East Street, building No. 2, -3 to 25, 101, 8, 804 rooms Applicant before: Tricorn (Beijing) Technology Co.,Ltd. |
|
| TA01 | Transfer of patent application right | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190521 |
|
| RJ01 | Rejection of invention patent application after publication |