CN108024005A - Information processing method, device, intelligent terminal, server and system - Google Patents

Information processing method, device, intelligent terminal, server and system Download PDF

Info

Publication number
CN108024005A
CN108024005A CN201610974695.2A CN201610974695A CN108024005A CN 108024005 A CN108024005 A CN 108024005A CN 201610974695 A CN201610974695 A CN 201610974695A CN 108024005 A CN108024005 A CN 108024005A
Authority
CN
China
Prior art keywords
information
server
vector
intelligent terminal
sequence
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.)
Granted
Application number
CN201610974695.2A
Other languages
Chinese (zh)
Other versions
CN108024005B (en
Inventor
涂畅
张扬
王砚峰
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Sogou Technology Development Co Ltd
Original Assignee
Beijing Sogou Technology Development Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Beijing Sogou Technology Development Co Ltd filed Critical Beijing Sogou Technology Development Co Ltd
Priority to CN201610974695.2A priority Critical patent/CN108024005B/en
Publication of CN108024005A publication Critical patent/CN108024005A/en
Application granted granted Critical
Publication of CN108024005B publication Critical patent/CN108024005B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04M—TELEPHONIC COMMUNICATION
    • H04M1/00—Substation equipment, e.g. for use by subscribers
    • H04M1/72—Mobile telephones; Cordless telephones, i.e. devices for establishing wireless links to base stations without route selection
    • H04M1/724—User interfaces specially adapted for cordless or mobile telephones
    • H04M1/72403—User interfaces specially adapted for cordless or mobile telephones with means for local support of applications that increase the functionality
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
    • G06F21/60—Protecting data
    • G06F21/62—Protecting access to data via a platform, e.g. using keys or access control rules
    • G06F21/6218—Protecting access to data via a platform, e.g. using keys or access control rules to a system of files or objects, e.g. local or distributed file system or database
    • G06F21/6245—Protecting personal data, e.g. for financial or medical purposes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/10—Text processing

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Bioethics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Computational Linguistics (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Signal Processing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Human Computer Interaction (AREA)
  • Medical Informatics (AREA)
  • Databases & Information Systems (AREA)
  • Computer Hardware Design (AREA)
  • Computer Security & Cryptography (AREA)
  • Software Systems (AREA)
  • Information Transfer Between Computers (AREA)
  • Telephonic Communication Services (AREA)

Abstract

An embodiment of the present invention provides a kind of information processing method, device, intelligent terminal, server and system, this method includes：Received information is docked by preconfigured word incorporation model to be changed, and generates corresponding sequence vector, wherein, institute's predicate incorporation model is obtained with the common training of the training pattern being configured in server；The sequence vector is sent to the server, so that the server determines the corresponding candidate information of the sequence vector by the training pattern；Candidate information according to the server feedback is shown.Pass through the embodiment of the present invention, transformed sequence vector is uploaded to server by intelligent terminal, so that server can not be according to the raw information received by the transformed sequence vector reduction intelligent terminal, and then the problem of avoid the information leakage caused by the information for directly uploading and receiving.

Description

Information processing method, device, intelligent terminal, server and system
Technical field
The present invention relates to input method technique field, more particularly to a kind of information processing method, a kind of information processor, A kind of intelligent terminal, a kind of server and a kind of information processing system.
Background technology
As the intelligent terminals such as the development of mobile communication technology, mobile phone are increasingly popularized, life, to people Practise, work brings great convenience.
These intelligent terminals are commonly installed various application programs (Application, APP), to realize its function. For example, intelligent terminal can dock received information and replied, that is, provide intelligence by having the function of the APP of intelligent replying Service is replied, meets the demand for services of user.
Specifically, when user is engaged in the dialogue using APP and other users, can be received by APP transmitted by other side Information, such as short message, mail, chat content, and information can be sent to other side, such as returned for the information that other side sends It is multiple.When using having the function of the APP of intelligent replying, intelligent terminal needs the information received by user being uploaded to clothes Business device, to generate the corresponding return information of the information by server, and then can be according to the return information that server is fed back Replied.User is generally considered that privacy concern, and the information that can refuse to receive is uploaded to server, that is, refuse this The service of sample.
The content of the invention
The embodiment of the present invention is the technical problem to be solved is that a kind of information processing method is provided, to solve existing intelligence eventually Upload information is held to the information leakage problem in the presence of server, so as to ensure information security.
Correspondingly, the embodiment of the present invention additionally provides a kind of information processor, intelligent terminal, server and a kind of information Processing system, to ensure the realization of the above method and application.
To solve the above-mentioned problems, the embodiment of the invention discloses a kind of information processing method, including：
Received information is docked by preconfigured word incorporation model to be changed, and generates corresponding sequence vector, its In, institute's predicate incorporation model is obtained with the common training of the training pattern being configured in server；
The sequence vector is sent to the server so that the server determined by the training pattern it is described The corresponding candidate information of sequence vector；
Candidate information according to the server feedback is shown.
Optionally, described to be changed by the received information of preconfigured word incorporation model docking, generation is corresponding Sequence vector, including：
When the information received is text message, the text message is segmented, the word sequence after being segmented；
By institute's predicate incorporation model, the word sequence is converted into the sequence vector.
Optionally, by institute's predicate incorporation model, the word sequence is converted into the sequence vector, including：
Corresponding first term vector of each word in the word sequence is determined according to vocabulary,；
Corresponding first term vector of each word is converted to by the second term vector by institute's predicate incorporation model, wherein, described The dimension of one vector is more than the dimension of second term vector；
The sequence vector is generated using corresponding second term vector of each word.
Optionally, the received information of docking is changed, and is generated corresponding sequence vector, is further included：
When the information received is voice messaging, by the way that the voice messaging is identified, corresponding text is determined Information；
After text message is determined, the step of being segmented to text message is performed.
Optionally, the candidate information includes：The server passes through the encoder and decoder in the training pattern Definite at least one candidate item,
The candidate information according to the server feedback is shown, including：
Receive the server and be directed to the candidate information that the sequence vector is fed back；
Based on default display rule, the candidate item in the candidate information is shown.
Optionally, the method further includes：
The candidate item for determining to choose according to operation instruction；
The candidate item chosen is inputted into the dialog box of communication applications, to reply described information, wherein, it is described Communication applications include a following at least type：Short message application, instant messaging application, mail applications, telecommunications talk application.
The embodiment of the invention also discloses another information processing method, including：
The sequence vector that intelligent terminal is sent is received, the sequence vector passes through preconfigured word for the intelligent terminal Incorporation model docks received information and carries out conversion generation；
By preconfigured training pattern, determine the corresponding candidate information of the sequence vector, the training pattern with Training obtains institute's predicate incorporation model jointly；
The candidate information is fed back into the intelligent terminal, so that the intelligent terminal is carried out according to the candidate information Displaying.
Optionally, it is described by preconfigured training pattern, determine the corresponding candidate information of the sequence vector, wrap Include：
The sequence vector is inputted into the encoder in the training pattern, obtains corresponding coding vector sequence；
By the decoder in training pattern described in the coding vector sequence inputting, at least one candidate item is obtained；
Based on obtained candidate item, the candidate information is determined.
Optionally, the candidate item is carried out for the intelligent terminal for described information in the dialog box of communication applications Reply；
Wherein, the communication applications include a following at least type：Short message application, instant messaging application, mail applications, electricity Believe talk application.
The embodiment of the invention also discloses a kind of information processor, including：
Sequence vector generation module, is changed for docking received information by preconfigured word incorporation model, Corresponding sequence vector is generated, wherein, institute's predicate incorporation model is obtained with the common training of the training pattern being configured in server；
Sequence vector sending module, for sending the sequence vector to the server, so that the server leads to Cross the training pattern and determine the corresponding candidate information of the sequence vector；
Candidate information display module, for being shown according to the candidate information of the server feedback.
Optionally, sequence vector generation module includes following submodule：Submodule is segmented, for being when the information received During text message, the text message is segmented, the word sequence after being segmented；Transform subblock, for by advance The word incorporation model of configuration, the sequence vector is converted into by the word sequence.
Optionally, which is specifically used for determining corresponding first word of each word in the word sequence according to vocabulary Vector；Corresponding first term vector of each word is converted to by the second term vector by institute's predicate incorporation model, wherein, first word The dimension of vector is more than the dimension of the second term vector；And the sequence vector is generated using corresponding second term vector of each word.
Optionally, sequence vector generation module further includes：Identify submodule.The identification submodule is used for when the letter received Cease for voice messaging when, by the way that the voice messaging is identified, determine corresponding text message；And in definite text envelope After breath, triggering participle submodule performs the step of being segmented to text message.
Optionally, candidate information includes：The server is determined by the encoder and decoder in the training pattern At least one candidate item.Candidate information display module includes following submodule：Receiving submodule, for receiving the server The candidate information fed back for the sequence vector；Submodule is shown, for regular based on default display, to the candidate Each candidate item in information is shown.
Optionally, the information processor in the embodiment of the present application can also include following module：Candidate item chooses module, For the candidate item for determining to choose according to operation instruction；Information-reply module, should to communication for inputting the candidate item chosen In dialog box, to reply described information.Wherein, the communication applications include a following at least type：Short message should With, instant messaging application, mail applications, telecommunications talk application.
The embodiment of the invention also discloses another information processor, including：
Sequence vector receiving module, for receiving the sequence vector of intelligent terminal transmission, the sequence vector is the intelligence Energy terminal docks received information by preconfigured word incorporation model and carries out conversion generation；
Candidate information determining module, for by preconfigured training pattern, determining the corresponding time of the sequence vector Select information, training obtains the training pattern jointly with institute predicate incorporation model；
Candidate information sending module, for the candidate information to be fed back to the intelligent terminal, so that the intelligence is eventually End is shown according to the candidate information.
Optionally, candidate information determining module includes following submodule：Encoding submodule, for the sequence vector is defeated Enter the encoder in the training pattern, obtain corresponding coding vector sequence；Decoding sub-module, for by the coding vector Decoder in training pattern described in sequence inputting, obtains at least one candidate item；Determination sub-module, for based on obtained time Option, determines the candidate information.
Optionally, candidate item is carried out specifically for the intelligent terminal for described information in the dialog box of communication applications Reply.Wherein, communication applications specifically include a following at least type：Short message application, instant messaging application, mail applications, telecommunications Talk application.
The embodiment of the invention also discloses a kind of intelligent terminal, include memory, and one or more than one Program, either more than one program storage in memory and is configured to by one or more than one processing for one of them Device performs the one or more programs and includes the instruction for being used for being operated below：
Received information is docked by preconfigured word incorporation model to be changed, and generates corresponding sequence vector, its In, institute's predicate incorporation model is obtained with the common training of the training pattern being configured in server；
The sequence vector is sent to the server so that the server determined by the training pattern it is described The corresponding candidate information of sequence vector；
Candidate information according to the server feedback is shown.
The embodiment of the invention also discloses a kind of server, includes memory, and one or more than one journey Sequence, either more than one program storage in memory and is configured to by one or more than one processor for one of them Perform the one or more programs and include the instruction for being used for being operated below：
The sequence vector that intelligent terminal is sent is received, the sequence vector passes through preconfigured word for the intelligent terminal Incorporation model docks received information and carries out conversion generation；
By preconfigured training pattern, determine the corresponding candidate information of the sequence vector, the training pattern with Training obtains institute's predicate incorporation model jointly；
The candidate information is fed back into the intelligent terminal, so that the intelligent terminal is carried out according to the candidate information Displaying.
The embodiment of the invention also discloses a kind of information processing system, including：Server and intelligent terminal；Wherein, it is described Server includes server as described in the present embodiment of the invention；The intelligence that intelligent terminal is included as described in the present embodiment of the invention is whole End.
Compared with prior art, the embodiment of the present invention includes advantages below：
In the embodiment of the present invention, the word incorporation model and training pattern that in advance obtain common training are respectively arranged at intelligence In terminal and server, turned in this way, intelligent terminal can dock received information by preconfigured word incorporation model Change, generate transformed sequence vector, and upload the sequence vector and determined to server, trigger the server by training pattern The corresponding candidate information of the sequence vector, so that intelligent terminal can get the candidate information that server is fed back；And turn Sequence vector after changing cannot be reduced by server, i.e., server can not restore raw information according to the sequence vector, therefore Server can not view the raw information received by intelligent terminal, avoid caused by the information for directly uploading and receiving The problem of information leakage, effectively protect the privacy of user.
Brief description of the drawings
Fig. 1 is the step flow chart of intelligent terminal side embodiment in a kind of information processing method of the invention；
Fig. 2 is the step flow chart of server-side embodiments in a kind of information processing method of the invention；
Fig. 3 is to interact schematic diagram between server and intelligent terminal in one embodiment of the invention；
Fig. 4 is a kind of structure diagram of information processor embodiment of the present invention；
Fig. 5 is the structure diagram of another information processor embodiment of the invention；
Fig. 6 is a kind of structure diagram of intelligent terminal for information processing according to an exemplary embodiment；
Fig. 7 is a kind of structure diagram of server in the embodiment of the present invention.
Embodiment
In order to make the foregoing objectives, features and advantages of the present invention clearer and more comprehensible, it is below in conjunction with the accompanying drawings and specific real Applying mode, the present invention is described in further detail.
When using having the function of the APP of intelligent replying, the intelligent terminal as client is needed received by user Information, short message that such as other side sends, mail, chat content, are uploaded to the server as server-side, to pass through server Generate the corresponding candidate information of the information.The candidate information of generation is back to intelligent terminal by server, so that intelligence is eventually End can dock received information according to the candidate information of return and be replied.But intelligent terminal directly receives user Information upload onto the server, there is the risk of leakage information, in some instances it may even be possible to reveal the privacy of user, that is, existence information safety The problem of.
One of the core concepts of the embodiments of the present invention is, in advance by the word incorporation model that common training obtains and training mould Type is respectively arranged in intelligent terminal and server, so that intelligent terminal can be incited somebody to action by preconfigured word incorporation model The information received is converted to sequence vector, and transformed sequence vector is sent to server, so that server passes through instruction Practice model and determine the corresponding candidate information of transformed sequence vector, so that intelligent terminal can get what server was fed back Candidate information, and transformed sequence vector cannot be reduced by server, i.e., and server can not be reduced according to the sequence vector Go out raw information, avoid caused information leakage problem that receive information directly uploads onto the server, effectively protect user Privacy.
Reference Fig. 1, shows the step flow chart of intelligent terminal side embodiment in a kind of information processing method of the present invention, Specifically it may include steps of：
Step 102, dock received information by preconfigured word incorporation model to be changed, generate corresponding vector Sequence, wherein, institute's predicate incorporation model is obtained with the common training of the training pattern being configured in server.
The application implements that client can will be known as installed in the application program of intelligent terminal, such as can be by user's foundation pair The communication applications of words are as client.Communication applications can specifically include：Short message application, instant messaging application, mail applications, electricity Believe talk application etc., the embodiment of the present application is not specifically limited this.Wherein, short message application can be used for receiving and/or send Short message；Instant messaging application can be used for sending and receiving the information that network is transmitted immediately, such as internet information；Mail applications It can be used for reception/or send mail；Telecommunications talk application is the application conversed by telecommunication modes, as dialed in terminal The application of phone.
After user logs in client, intelligent terminal can receive the information for being sent to the user by client, and And received information can be docked and be shown so that user can check the information, so that user can dock received letter Breath is replied.After information is received, intelligent terminal can be by preconfigured word incorporation model, the information that will be received Corresponding sequence vector is converted to, information is uploaded to server in the form of by vector.
Step 104, the sequence vector is sent to server, so that the server is determined by the training pattern The corresponding candidate information of the sequence vector.
Intelligent terminal can be uploaded the transformed sequence vector by network after transformed sequence vector is generated To server, to generate the corresponding candidate information of the transformed sequence vector by server.Server is receiving vector After sequence, the corresponding candidate information of the transformed sequence vector can be determined by preconfigured training pattern, and will The candidate information feeds back to intelligent terminal.
Step 106, the candidate information according to the server feedback is shown.
Wherein, candidate information can specifically include：Server determines that described information is corresponding each according to the sequence vector Candidate item.Specifically, intelligent terminal after the candidate information that server is fed back is received, can show rule according to default Candidate information is shown so that user can use a certain candidate item in the candidate information or the candidate information to receiving Replied to information, that is, facilitate user's return information, so as to improve ease-to-operate.
In the embodiment of the present invention, the training pattern that is configured in the word incorporation model and server that are configured in intelligent terminal It is trained together to obtain.Wherein, word incorporation model is configured in intelligent terminal, so that intelligent terminal can pass through the word The information received is converted to sequence vector by incorporation model, which is uploaded to server.Obviously, The sequence vector is converted to by the word incorporation model of intelligent terminal, can thus prevent the sequence vector from by Server reduces, i.e., server can not restore raw information according to the sequence vector, and server can not view intelligent terminal Received raw information, effectively the problem of avoiding the information leakage caused by the information for directly uploading and receiving, protection The privacy of user.In addition, being pre-configured with training pattern in the server is instructed jointly with the word incorporation model in intelligent terminal Get, therefore server can pass through preconfigured training when receiving the sequence vector that intelligent terminal is uploaded Model determines the corresponding candidate information of the sequence vector, and the candidate information is fed back to intelligent terminal, so that intelligence is eventually End can be shown according to the candidate information that server is fed back.
With reference to Fig. 2, show the step flow chart of server-side embodiments in a kind of information processing method of the present invention, have Body may include steps of：
Step 202, the sequence vector that intelligent terminal is sent is received.
Wherein, sequence vector docks received information by preconfigured word incorporation model for intelligent terminal and is changed Generation.
Step 204, by preconfigured training pattern, the corresponding candidate information of the sequence vector is determined.
Wherein, the training pattern of server configuration and the word incorporation model of intelligent terminal configuration train what is obtained jointly.Clothes Being engaged in device can be using the sequence vector that intelligent terminal is uploaded as input, and by the training pattern, determines the vector sequence Arrange corresponding candidate information.Preferably, which can specifically include the corresponding each candidate item of information.Specifically, take Device be engaged in after sequence vector is received, can be using the sequence vector as input, with by inputting the sequence vector to advance The training pattern of configuration, generates corresponding candidate item, such as generates the candidate item of one or several intelligent replyings；And it can incite somebody to action The candidate information is fed back to intelligent terminal by the candidate item of generation as candidate information.
Step 206, the candidate information is fed back into the intelligent terminal, so that the intelligent terminal is according to the candidate Information is shown.
Candidate information is fed back to intelligent terminal by server, also will the candidate information be sent to the intelligent terminal and run Client so that intelligent terminal can be shown according to the candidate information, and then can be by the candidate information Each candidate item recommends user.User can choose wherein any one candidate item from the candidate item that intelligent terminal is shown, Received information is docked to be replied.Intelligent terminal can be by client, and the candidate item that user is chosen is sent to other side institute The intelligent terminal used, to reply the information transmitted by other side, each word in return information, side are inputted without user Just user's return information, improves the simplicity and efficiency of user's return information.
In the embodiment of the present invention, what server received is transformed sequence vector, according to the transformed vectorial sequence Row can not reduce the information received by intelligent terminal script, therefore can not view the information of script, solve existing intelligence Terminal upload information is to the information leakage problem in the presence of server, so as to effectively protect the privacy of user.
As the concrete application of the present invention, deep learning model algorithm can be based on, as shot and long term remembers manually god Through network (Long Short-Term Memory, LSTM) algorithm, bluebeard compound embedded technology, an intelligence is trained with mass data Model is replied, is replied so that intelligent terminal can dock received information by the intelligent replying model.Substantially, The model of mind can specifically include word incorporation model and training pattern；Word incorporation model can be configured at the intelligence as client Energy terminal, so that the information received can be converted to sequence vector by intelligent terminal by the word incorporation model；Training pattern It can be configured in server, so that server can be by the training pattern, after determining the conversion that intelligent terminal is uploaded The corresponding candidate information of sequence vector.In this way, intelligent terminal can determine candidate information by server, and can be with Each candidate item in the candidate information is shown to user, each candidate item in candidate information is recommended into user, so that User can select the wherein received information of any one candidate item docking to be replied；Or can be according to default reply plan Candidate item is slightly chosen, to automatically reply the information received, greatly improves the ease-to-operate of user's return information, while having , can be to avoid calculating of the intelligent replying limited-service in intelligent terminals such as mobile phones in itself on the premise of effect protection privacy of user The situation of ability, improves user experience.
It should be noted that the embodiment of the present application can also realize intelligent replying model using other algorithms, nerve is such as circulated Network (Recurrent neural Network, RNN), or the deep learning algorithm being derived by RNN, it is such as two-way LSTM, two-way RNN, LSTM, CopyNet scheduling algorithm with Attention mechanism, the embodiment of the present application are not restricted this.
With reference to Fig. 3, show in one embodiment of the invention and interact schematic diagram between server and intelligent terminal.This Shen Please the information processing method that is provided of embodiment specifically may include steps of：
Step 302, intelligent terminal docks received information by preconfigured word incorporation model and is changed, generation pair The sequence vector answered.
In the embodiment of the present application, intelligent terminal can receive various information by communication applications；The information can include Various data, such as image data, text data, video data and voice data, the embodiment of the present application do not limit this.Example Such as, intelligent terminal can receive text message and/or voice messaging by communication applications, and text information can specifically include Text data, the voice messaging can specifically include voice data.
It is described that received information is docked by preconfigured word incorporation model in the alternative embodiment of the application Changed, generate corresponding sequence vector, can specifically be included：When the information received is text message, to the text This information is segmented, the word sequence after being segmented；By preconfigured word incorporation model, the word sequence is converted into The sequence vector.Specifically, intelligent terminal can segment text message, obtain after text message is received Each word and/or word in text information, and obtained words can be formed to corresponding word sequence, and can be according to Each words putting in order in word sequence, by each words in word sequence, corresponding term vector is inputted to matching somebody with somebody in advance before switching In the word incorporation model put, to generate each words corresponding term vector after conversion, that is, embedding by word trained in advance Enter model, the word sequence is converted into sequence vector.
As the specific example of the application, intelligent terminal, can be by text envelope after text message is received Breath is segmented, the word sequence after being segmented.Each word in the word sequence can use a vector representation, such as a word Represented with vectorial an；N can be used to indicate that each word putting in order in word sequence in text message, such as can be integer.Intelligence Can terminal can use each word in vector representation word sequence, so as to the primary vector sequence A corresponding to after segment (a1, a2, a3 ..., an)；And primary vector sequence A (a1, a2, a3 ..., an) can be changed by word embedded technology For secondary vector sequence B (b1, b2, b3 ..., bn).For example, a sentence can be divided into more by intelligent terminal by participle A word, and can be indicated with primary vector sequence A (a1, a2 ..., an).Primary vector sequence A (a1, a2 ..., An) pass through after word incorporation model, another sequence vector B (b1, b2, b3 ..., bn) can be converted to, that is, be converted to the Two sequence vector B (b1, b2, b3 ..., bn).
It should be noted that in order to distinguish the term vector of pre- and post-conversion, the application is implemented can be by before conversion Term vector is known as the first term vector, and transformed term vector is known as the second term vector, such as by each word in primary vector sequence A Vector is as the first term vector corresponding to each word, and using each term vector in secondary vector sequence B as corresponding to each word The second term vector.Wherein, the dimension of the first term vector is not equal to the dimension of the second term vector, i.e. intelligent terminal is embedded in by word The the first common term vector that can be reduced, is converted to the second term vector that cannot be reduced by model so that server without Method restores each word in word sequence by the second term vector, i.e. server cannot get the original letter received by intelligent terminal Breath.
In the alternative embodiment of the application, intelligent terminal is by word incorporation model trained in advance, by institute's predicate Sequence is converted into the sequence vector, can specifically include：Each word corresponding first in the word sequence is determined according to vocabulary Term vector；Corresponding first term vector of each word is converted to by the second term vector by institute's predicate incorporation model, wherein, described first The dimension of term vector is more than the dimension of second term vector；Using the corresponding second term vector generation of each word vectorial sequence Row.
As the specific example of the application, after intelligent terminal is segmented by docking received text message, with And the corresponding first term vector an (0,0,1,0,0,0 ... ..., 0) of each word after participle can be determined according to vocabulary, and can lead to Pre-set word incorporation model M is crossed, the first term vector an (0,0,1,0,0,0 ... ..., 0) is converted into the second term vector bn (0.01,0.03,0.13 ... ..., 0.2), is such as converted to secondary vector bn by calculation formula bn=M*an by primary vector an. Wherein, word incorporation model M can be the matrix of an x*y dimension, if x=100, y=50000, then whole matrix M can be wrapped Containing 5000000 elements.X can be used as conversion after dimension, specifically can be used for characterize secondary vector sequence B in the second word to The dimension of amount, can such as be equal to the dimension of b1；Y can be used as dimension before conversion, specifically can be used for characterizing primary vector sequence A In the first term vector dimension, such as can be equal to a1 dimension.
It should be noted that if dimension y is too big before conversion, then word incorporation model can be very huge, can cause whole model Parameter it is relatively more, influence training speed and calculating speed.Optionally, the embodiment of the present application can be configured according to the size of vocabulary Dimension y before conversion, as dimension y can be specifically as follows tens of thousands of equal to the size of vocabulary before changing；And the dimension of the second term vector It can be configured according to the specific algorithm employed in word incorporation model, may be typically less than the dimension of the first term vector, it is excellent Choosing, dimension x may be typically less than dimension y before conversion after conversion, such as can be thousands of or hundreds of, it might even be possible to it is tens, So that useful information can be represented with overall compact form, easy to server-side processes, and server can not lead to The raw information received by transformed secondary vector sequence reduction intelligent terminal is crossed, and then can asking to avoid information leakage Topic.
In the alternative embodiment of the application, intelligent terminal, which docks received information, to be changed, and generation is corresponding Sequence vector, can also include：When the information received is voice messaging, by the way that the voice messaging is identified, really Fixed corresponding text message；After text message is determined, the step of being segmented to text message is performed.The embodiment of the present application In, which can be converted to text message by intelligent terminal when receiving voice messaging, be performed so as to pass through The step of being segmented to text message, corresponding primary vector sequence after being segmented, and mould can be embedded in by word Primary vector sequence is converted to secondary vector sequence by type.
Step 304, intelligent terminal sends the sequence vector to server.
After intelligent terminal obtains transformed secondary vector sequence, can using the secondary vector sequence as sequence vector, And server is sent to, so that server can receive the sequence vector.
Step 306, server receives the sequence vector that intelligent terminal is sent.
Specifically, server receive be by the transformed sequence vector of intelligent terminal, can not be by the sequence vector The information of script is reduced to, therefore server can not get the information received by intelligent terminal, and then can be to avoid leakage The risk of user information, effectively protects the personal secrets of user.
Step 308, server determines the corresponding candidate information of the sequence vector by preconfigured training pattern.
In the embodiment of the present application, the training pattern configured in server can include decoder and encoder, to pass through solution Code device docks received sequence vector and is encoded, and obtains the sequence vector after a coding, i.e. coding vector sequence, Ran Houtong Cross decoder to decode coding vector sequence, obtain corresponding candidate item.Wherein, instruction can be used to the quantity of candidate item The parameter practiced in model is configured, and the application implements to be not specifically limited the quantity of candidate item.Therefore, server determines Candidate information can specifically include：At least one candidate item determined by the encoder and decoder in the training pattern. The candidate item specifically can be used for intelligent terminal and be replied for described information in the dialog box of communication applications.Wherein, lead to News application can specifically include：Short message application, instant messaging application, mail applications, telecommunications talk application etc., the embodiment of the present invention This is not restricted.
In the alternative embodiment of the application, server determines the vector by preconfigured training pattern The step of sequence corresponding candidate information, it can specifically include：The sequence vector is inputted into the coding in the training pattern Device, obtains coding vector sequence；Then by the decoder in training pattern described in the coding vector sequence inputting, obtain at least One candidate item, so as to based on obtained candidate item, determine candidate information.As the specific example of the present invention, clothes The training pattern that business device is configured is that the sequence in deep learning generates model, such as seq2seq (Sequence to Sequence) model.In seq2seq models, server can use LSTM model as encoder (encoder), Another LSTM model is as decoder (decoder).Intelligent terminal by transformed secondary vector sequence B (b1, b2, B3 ..., bn) be uploaded to server after, server inputs secondary vector sequence B (b1, b2, b3 ..., bn) to training pattern In encoder, to be encoded by the encoder, obtain one coding after coding vector sequence C (c1, c2, c3 ..., Cn), then coding vector sequence C (c1, c2, c3 ..., cn) is inputted to the decoder in training pattern, to pass through the decoding Device is decoded, and uses such as beam search (beam search) scheduling algorithm, generates a certain number of candidate items, to obtain the final product To at least one candidate item；And obtained candidate item can be sent to intelligent terminal as candidate information, with to intelligent end The uploaded secondary vector sequence in end is fed back.
Step 310, the candidate information is fed back to the intelligent terminal by server.
The candidate information of generation can be returned to intelligent terminal by server, so that intelligent terminal can be according to described Candidate information is shown.
Step 312, intelligent terminal is shown according to the candidate information of the server feedback.
Wherein, candidate information includes：Server determines the corresponding each candidate item of described information according to the sequence vector.
In the alternative embodiment of the application, intelligent terminal is opened up according to the candidate information of the server feedback Show, can specifically include：Receive the server and be directed to the candidate information that the sequence vector is fed back；Based on default display Rule, is shown each candidate item in the candidate information.
Intelligent terminal can extract each candidate after the candidate information that server is fed back is received from the candidate information , and default display rule can be based on, such as determine that the display of each candidate item is suitable according to the priority orders of each candidate item Sequence, to be shown according to display order in specified region to each candidate item, such as shows each candidate in the candidate region of input method , or display area in communication applications associated by dialog box is shown each candidate item, the application is implemented to this not It is restricted.
Step 314, the candidate item that intelligent terminal determines to choose according to operation instruction.
Intelligent terminal is shown each candidate item, so that user, which can get server, is directed to the letter received The recommended candidate item of breath.User can input operation instruction to intelligent terminal, be referred to the whole terminal of triggering intelligence according to the operation Show and choose some candidate item, that is, determine the candidate item that user chooses.
Step 316, intelligent terminal inputs the candidate item chosen into the dialog box of communication applications, with to described information into Row is replied.
Wherein, the communication applications can include but is not limited to a following at least type：Short message application, instant messaging should With, mail applications, telecommunications talk application.
After the candidate item that definite user chooses, the candidate item that can choose this is inputted into communication applications intelligent terminal Dialog box, is replied so that user can dock received information according to the candidate item that this is chosen, improves and reply operation Simplicity, and improve the reply efficiency of user.
Certainly, intelligent terminal can be directed to the edit operation or modification operation that the candidate item chosen is inputted according to user, Modify, replied with docking received information according to amended candidate item, so as to improve user to candidate item Experience.
It should be noted that for embodiment of the method, in order to be briefly described, therefore it is all expressed as to a series of action group Close, but those skilled in the art should know, the embodiment of the present invention and from the limitation of described sequence of movement, because according to According to the embodiment of the present invention, some steps can use other orders or be carried out at the same time.Secondly, those skilled in the art also should Know, embodiment described in this description belongs to preferred embodiment, and the involved action not necessarily present invention is implemented Necessary to example.
With reference to Fig. 4, show a kind of structure diagram of information processor embodiment of the present invention, can specifically include as follows Module：
Sequence vector generation module 402, is turned for docking received information by preconfigured word incorporation model Change, generate corresponding sequence vector, wherein, institute's predicate incorporation model and the training pattern that is configured in server are common trained Arrive.
Sequence vector sending module 404, for sending the sequence vector to the server, so that the server The corresponding candidate information of the sequence vector is determined by the training pattern.
Candidate information display module 406, for being shown according to the candidate information of the server feedback.
In the alternative embodiment of the application, sequence vector generation module 402, can include following submodule：
Submodule is segmented, for when the information received is text message, segmenting, obtaining to the text message Word sequence after participle.
Transform subblock, for by preconfigured word incorporation model, the word sequence to be converted into the vectorial sequence Row.
In the alternative embodiment of the application, transform subblock, specifically can be used for determining institute's predicate according to vocabulary Corresponding first term vector of each word in sequence；Corresponding first term vector of each word is converted to by institute's predicate incorporation model Two term vectors, wherein, the dimension of first term vector is not equal to the dimension of second term vector；And corresponded to using each word The second term vector generate the sequence vector.Optionally, the dimension of the first term vector is more than the dimension of the second term vector.
In the alternative embodiment of the application, sequence vector generation module 402, can also include：Identify submodule. The identification submodule can be used for when the information received is voice messaging, by the way that the voice messaging is identified, really Fixed corresponding text message；And after text message is determined, participle submodule can be triggered and performed text message is divided The step of word.
In the alternative embodiment of the application, candidate information can include：The server passes through the trained mould At least one candidate item that encoder and decoder in type determine.Candidate information display module 406 can include following submodule Block：
Receiving submodule, the candidate information for being directed to the sequence vector for receiving the server and being fed back.
Submodule is shown, for based on default display rule, being shown to each candidate item in the candidate information.
Optionally, the information processor in the embodiment of the present application can also include following module：
Candidate item chooses module, for the candidate item for determining to choose according to operation instruction；
Information-reply module, for inputting the candidate item chosen into the dialog box of communication applications, with to described information Replied.
Wherein, the communication applications can include but is not limited to a following at least type：Short message application, instant messaging should With, mail applications, telecommunications talk application.
Information processor in the embodiment of the present application specifically can be applicable in the intelligent terminal of above method embodiment, from And the information received can be converted to sequence vector, transformed sequence vector is uploaded to server so that service Information of the device according to received by the transformed sequence vector can not reduce intelligent terminal, and then avoid the letter that will be received Breath is uploaded to the information leakage problem caused by server；And the candidate information that server is fed back can be received, and then It can recommend each candidate item in the candidate information to user so that user can use the candidate item docking that server is generated Received information is replied, so as to improve the efficiency of user's return information.
With reference to Fig. 5, show the structure diagram of another information processor embodiment of the invention, can specifically include such as Lower module：
Sequence vector receiving module 502, for receiving the sequence vector of intelligent terminal transmission, the sequence vector is described Intelligent terminal docks received information by preconfigured word incorporation model and carries out conversion generation；
Candidate information determining module 504, for by preconfigured training pattern, determining that the sequence vector is corresponding Candidate information, training obtains the training pattern jointly with institute predicate incorporation model；
Candidate information sending module 506, for the candidate information to be fed back to the intelligent terminal, so that the intelligence Terminal is shown according to the candidate information.
In the alternative embodiment of the application, candidate information determining module 504, can specifically include following submodule Block：
Encoding submodule, for the sequence vector to be inputted to the encoder in the training pattern, obtains corresponding volume Code vector sequence；
Decoding sub-module, for by the decoder in training pattern described in the coding vector sequence inputting, obtaining at least One each candidate item；
Determination sub-module, for based on obtained each candidate item, determining the candidate information.
In the alternative embodiment of the application, candidate item specifically can be used for the intelligent terminal and be directed to described information Replied in the dialog box of communication applications.Wherein, communication applications specifically can include but is not limited to a following at least type： Short message application, instant messaging application, mail applications, telecommunications talk application.
Information processor in the embodiment of the present application specifically can be applicable in the server of above method embodiment, can be with The transformed sequence vector that intelligent terminal is uploaded is input in preconfigured training pattern, obtains corresponding candidate's letter Breath, and the candidate information is fed back into intelligent terminal so that intelligent terminal can be shown according to the candidate information, so that The received information of candidate item docking that obtaining user can use server to be generated be replied, and improve the effect of user's return information Rate.
For device embodiment, since it is substantially similar to embodiment of the method, so description is fairly simple, it is related Part illustrates referring to the part of embodiment of the method.
Fig. 6 is a kind of structure diagram of intelligent terminal 600 for information processing according to an exemplary embodiment. For example, intelligent terminal 600 can be mobile phone, computer, digital broadcast terminal, messaging devices, game console, puts down Board device, Medical Devices, body-building equipment, personal digital assistant etc..
With reference to Fig. 6, intelligent terminal 600 can include following one or more assemblies：Processing component 602, memory 604, Power supply module 606, multimedia component 608, audio component 610, the interface 612 of input/output (I/O), sensor component 614, And communication component 616.
Processing component 602 usually controls the integrated operation of intelligent terminal 600, such as leads to display, call, data The operation that letter, camera operation and record operation are associated.Treatment element 602 can be held including one or more processors 620 Row instruction, to complete all or part of step of above-mentioned method.In addition, processing component 602 can include one or more moulds Block, easy to the interaction between processing component 602 and other assemblies.For example, processing component 602 can include multi-media module, with Facilitate the interaction between multimedia component 608 and processing component 602.
Memory 604 is configured as storing various types of data to support the operation in equipment 600.These data are shown Example includes the instruction of any application program or method for being operated on intelligent terminal 600, contact data, telephone directory number According to, message, picture, video etc..Memory 604 can by any kind of volatibility or non-volatile memory device or they Combination realize, as static RAM (SRAM), electrically erasable programmable read-only memory (EEPROM) are erasable Programmable read only memory (EPROM), programmable read only memory (PROM), read-only storage (ROM), magnetic memory, quick flashing Memory, disk or CD.
Electric power assembly 606 provides electric power for the various assemblies of intelligent terminal 600.Electric power assembly 606 can include power supply pipe Reason system, one or more power supplys, and other components associated with generating, managing and distributing electric power for intelligent terminal 600.
Multimedia component 608 is included in the screen of one output interface of offer between the intelligent terminal 600 and user. In certain embodiments, screen can include liquid crystal display (LCD) and touch panel (TP).If screen includes touch surface Plate, screen may be implemented as touch-screen, to receive input signal from the user.Touch panel includes one or more touch Sensor is to sense the gesture on touch, slip and touch panel.The touch sensor can not only sense touch or slip The border of action, but also detection and the duration and pressure associated with the touch or slide operation.In certain embodiments, Multimedia component 608 includes a front camera and/or rear camera.When intelligent terminal 600 is in operator scheme, such as clap When taking the photograph pattern or video mode, front camera and/or rear camera can receive exterior multi-medium data.It is each preposition Camera and rear camera can be a fixed optical lens system or have focusing and optical zoom capabilities.
Audio component 610 is configured as output and/or input audio signal.For example, audio component 610 includes a Mike Wind (MIC), when intelligent terminal 600 is in operator scheme, during such as call model, logging mode and speech recognition mode, microphone It is configured as receiving external audio signal.The received audio signal can be further stored in memory 604 or via logical Letter component 616 is sent.In certain embodiments, audio component 610 further includes a loudspeaker, for exports audio signal.
I/O interfaces 612 provide interface between processing component 602 and peripheral interface module, and above-mentioned peripheral interface module can To be keyboard, click wheel, button etc..These buttons may include but be not limited to：Home button, volume button, start button and lock Determine button.
Sensor component 614 includes one or more sensors, for providing the state of various aspects for intelligent terminal 600 Assessment.For example, sensor component 614 can detect opening/closed mode of equipment 600, the relative positioning of component, such as institute The display and keypad that component is intelligent terminal 600 are stated, sensor component 614 can also detect intelligent terminal 600 or intelligence The position of 600 1 components of terminal changes, the existence or non-existence that user contacts with intelligent terminal 600,600 orientation of intelligent terminal Or acceleration/deceleration and the temperature change of intelligent terminal 600.Sensor component 614 can include proximity sensor, be configured to Presence of nearby objects is detected without any physical contact.Sensor component 614 can also include optical sensor, such as CMOS or ccd image sensor, for being used in imaging applications.In certain embodiments, which can be with Including acceleration transducer, gyro sensor, Magnetic Sensor, pressure sensor or temperature sensor.
Communication component 616 is configured to facilitate the communication of wired or wireless way between intelligent terminal 600 and other equipment. Intelligent terminal 600 can access the wireless network based on communication standard, such as WiFi, 2G or 3G, or combinations thereof.Show at one In example property embodiment, communication component 616 receives broadcast singal or broadcast from external broadcasting management system via broadcast channel Relevant information.In one exemplary embodiment, the communication component 616 further includes near-field communication (NFC) module, short to promote Cheng Tongxin.For example, radio frequency identification (RFID) technology, Infrared Data Association (IrDA) technology, ultra wide band can be based in NFC module (UWB) technology, bluetooth (BT) technology and other technologies are realized.
In the exemplary embodiment, intelligent terminal 600 can be by one or more application application-specific integrated circuit (ASIC), number Word signal processor (DSP), digital signal processing appts (DSPD), programmable logic device (PLD), field programmable gate array (FPGA), controller, microcontroller, microprocessor or other electronic components are realized, for performing the above method.
In the exemplary embodiment, a kind of non-transitorycomputer readable storage medium including instructing, example are additionally provided Such as include the memory 604 of instruction, above-metioned instruction can be performed to complete the above method by the processor 620 of intelligent terminal 600.Example Such as, the non-transitorycomputer readable storage medium can be ROM, it is random access memory (RAM), CD-ROM, tape, soft Disk and optical data storage devices etc..
A kind of non-transitorycomputer readable storage medium, when the instruction in the storage medium is by the processing of intelligent terminal When device performs so that intelligent terminal is able to carry out a kind of information processing method, the described method includes：It is embedding by preconfigured word Enter the received information of model docking to be changed, generate corresponding sequence vector, wherein, institute's predicate incorporation model is with being configured at clothes Training obtains training pattern in business device jointly；The sequence vector is sent to the server, so that the server leads to Cross the training pattern and determine the corresponding candidate information of the sequence vector；Candidate information according to the server feedback carries out Displaying.
Fig. 7 is a kind of structure diagram of server in the embodiment of the present invention.The server 700 can because configuration or performance not Bigger difference is produced together, one or more central processing units (central processing can be included Units, CPU) 722 (for example, one or more processors) and memory 732, one or more storages apply journey The storage medium 730 of sequence 742 or data 744 (such as one or more mass memory units).Wherein, 732 He of memory Storage medium 730 can be of short duration storage or persistently storage.Can be included by being stored in the program of storage medium 730 by one or one With upper module (diagram does not mark), each module can include operating the series of instructions in server.Further, in Central processor 722 could be provided as communicating with storage medium 730, be performed on server 700 a series of in storage medium 730 Command operating.
Server 700 can also include one or more power supplys 726, one or more wired or wireless networks Interface 750, one or more input/output interfaces 758, one or more keyboards 756, and/or, one or one Above operating system 741, such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM etc..
A kind of non-transitorycomputer readable storage medium, when the instruction in the storage medium is by the processor of server During execution so that server is able to carry out a kind of information processing method, the described method includes：Receive the vector that intelligent terminal is sent Sequence, the sequence vector docks received information for the intelligent terminal by preconfigured word incorporation model to be changed Generation；By preconfigured training pattern, the corresponding candidate information of the sequence vector, the training pattern and institute are determined Training obtains predicate incorporation model jointly；The candidate information is fed back into the intelligent terminal so that the intelligent terminal according to It is shown according to the candidate information.
On the basis of above-described embodiment, the embodiment of the present application additionally provides a kind of information processing system, the information processing System can specifically include：Server and intelligent terminal；Wherein, server and the server basic one described in above-described embodiment Cause, intelligent terminal and the intelligent terminal described in above-described embodiment are basically identical, therefore repeat no more.
In the information processing system of the embodiment of the present application, the word incorporation model that intelligent terminal is configured is matched somebody with somebody with server The training pattern put trains what is obtained jointly.The information that intelligent terminal be able to will be received by preconfigured word incorporation model Corresponding sequence vector is converted to, and uploads transformed sequence vector to server, trigger the server is by being pre-configured with Training pattern determine the corresponding candidate information of the transformed sequence vector, so as to get the time that server is fed back Select information, and raw information of the server according to received by the transformed sequence vector can not reduce intelligent terminal, i.e., without Method views the raw information received by intelligent terminal, and then avoids the information caused by the information for directly uploading and receiving The problem of leakage, effectively protect the privacy of user.
Each embodiment in this specification is described by the way of progressive, what each embodiment stressed be with The difference of other embodiment, between each embodiment identical similar part mutually referring to.
It should be understood by those skilled in the art that, the embodiment of the embodiment of the present invention can be provided as method, apparatus or calculate Machine program product.Therefore, the embodiment of the present invention can use complete hardware embodiment, complete software embodiment or combine software and The form of the embodiment of hardware aspect.Moreover, the embodiment of the present invention can use one or more wherein include computer can With in the computer-usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.) of program code The form of the computer program product of implementation.
The embodiment of the present invention be with reference to according to the method for the embodiment of the present invention, terminal device (system) and computer program The flowchart and/or the block diagram of product describes.It should be understood that it can realize flowchart and/or the block diagram by computer program instructions In each flow and/or block and flowchart and/or the block diagram in flow and/or square frame combination.These can be provided Computer program instructions are set to all-purpose computer, special purpose computer, Embedded Processor or other programmable data processing terminals Standby processor is to produce a machine so that is held by the processor of computer or other programmable data processing terminal equipments Capable instruction is produced and is used for realization in one flow of flow chart or multiple flows and/or one square frame of block diagram or multiple square frames The device for the function of specifying.
These computer program instructions, which may also be stored in, can guide computer or other programmable data processing terminal equipments In the computer-readable memory to work in a specific way so that the instruction being stored in the computer-readable memory produces bag The manufacture of command device is included, which realizes in one flow of flow chart or multiple flows and/or one side of block diagram The function of being specified in frame or multiple square frames.
These computer program instructions can be also loaded into computer or other programmable data processing terminal equipments so that Series of operation steps is performed on computer or other programmable terminal equipments to produce computer implemented processing, so that The instruction performed on computer or other programmable terminal equipments is provided and is used for realization in one flow of flow chart or multiple flows And/or specified in one square frame of block diagram or multiple square frames function the step of.
Although having been described for the preferred embodiment of the embodiment of the present invention, those skilled in the art once know base This creative concept, then can make these embodiments other change and modification.So appended claims are intended to be construed to Including preferred embodiment and fall into all change and modification of range of embodiment of the invention.
Finally, it is to be noted that, herein, relational terms such as first and second and the like be used merely to by One entity or operation are distinguished with another entity or operation, without necessarily requiring or implying these entities or operation Between there are any actual relationship or order.Moreover, term " comprising ", "comprising" or its any other variant meaning Covering non-exclusive inclusion, so that process, method, article or terminal device including a series of elements are not only wrapped Those key elements are included, but also including other elements that are not explicitly listed, or further include as this process, method, article Or the key element that terminal device is intrinsic.In the absence of more restrictions, wanted by what sentence "including a ..." limited Element, it is not excluded that also there are other identical element in the process including the key element, method, article or terminal device.
Above to information processing method provided by the present invention, information processor, intelligent terminal, server and information at Reason system, is described in detail, and specific case used herein is set forth the principle of the present invention and embodiment, The explanation of above example is only intended to help to understand method and its core concept of the invention；Meanwhile for the one of this area As technical staff, according to the thought of the present invention, there will be changes in specific embodiments and applications, to sum up institute State, this specification content should not be construed as limiting the invention.

Claims (14)

  1. A kind of 1. information processing method, it is characterised in that including：
    Received information is docked by preconfigured word incorporation model to be changed, and generates corresponding sequence vector, wherein, institute Predicate incorporation model is obtained with the common training of the training pattern being configured in server；
    The sequence vector is sent to the server, so that the server determines the vector by the training pattern The corresponding candidate information of sequence；
    Candidate information according to the server feedback is shown.
  2. 2. according to the method described in claim 1, it is characterized in that, it is described by preconfigured word incorporation model to receiving Information changed, generate corresponding sequence vector, including：
    When the information received is text message, the text message is segmented, the word sequence after being segmented；
    By institute's predicate incorporation model, the word sequence is converted into the sequence vector.
  3. 3. according to the method described in claim 2, it is characterized in that, by institute's predicate incorporation model, the word sequence is changed Into the sequence vector, including：
    Corresponding first term vector of each word in the word sequence is determined according to vocabulary；
    Corresponding first term vector of each word is converted to by the second term vector by institute's predicate incorporation model, wherein, first word The dimension of vector is more than the dimension of second term vector；
    The sequence vector is generated using corresponding second term vector of each word.
  4. 4. according to the method in claim 2 or 3, it is characterised in that the received information of docking is changed, generation pair The sequence vector answered, further includes：
    When the information received is voice messaging, by the way that the voice messaging is identified, corresponding text message is determined；
    After text message is determined, the step of being segmented to text message is performed.
  5. 5. method according to any one of claims 1 to 3, it is characterised in that the candidate information includes：The server leads to At least one candidate item that the encoder and decoder crossed in the training pattern determine,
    The candidate information according to the server feedback is shown, including：
    Receive the server and be directed to the candidate information that the sequence vector is fed back；
    Based on default display rule, the candidate item in the candidate information is shown.
  6. 6. according to the method described in claim 5, it is characterized in that, the method further includes：
    The candidate item for determining to choose according to operation instruction；
    The candidate item chosen is inputted into the dialog box of communication applications, to reply described information, wherein, the communication Using including a following at least type：Short message application, instant messaging application, mail applications, telecommunications talk application.
  7. A kind of 7. information processing method, it is characterised in that including：
    The sequence vector that intelligent terminal is sent is received, the sequence vector is embedded in for the intelligent terminal by preconfigured word Model docks received information and carries out conversion generation；
    By preconfigured training pattern, determine the corresponding candidate information of the sequence vector, the training pattern with it is described Training obtains word incorporation model jointly；
    The candidate information is fed back into the intelligent terminal, so that the intelligent terminal is opened up according to the candidate information Show.
  8. 8. the method according to the description of claim 7 is characterized in that described by preconfigured training pattern, determine described The corresponding candidate information of sequence vector, including：
    The sequence vector is inputted into the encoder in the training pattern, obtains corresponding coding vector sequence；
    By the decoder in training pattern described in the coding vector sequence inputting, at least one candidate item is obtained；
    Based on obtained candidate item, the candidate information is determined.
  9. 9. according to the method described in claim 8, it is characterized in that, the candidate item is directed to the letter for the intelligent terminal Breath is replied in the dialog box of communication applications；
    Wherein, the communication applications include a following at least type：Short message application, instant messaging application, mail applications, telecommunications are led to Words application.
  10. A kind of 10. information processor, it is characterised in that including：
    Sequence vector generation module, is changed for docking received information by preconfigured word incorporation model, is generated Corresponding sequence vector, wherein, institute's predicate incorporation model is obtained with the common training of the training pattern being configured in server；
    Sequence vector sending module, for sending the sequence vector to the server, so that the server passes through institute State training pattern and determine the corresponding candidate information of the sequence vector；
    Candidate information display module, for being shown according to the candidate information of the server feedback.
  11. A kind of 11. information processor, it is characterised in that including：
    Sequence vector receiving module, for receiving the sequence vector of intelligent terminal transmission, the sequence vector is whole for the intelligence Received information is docked by preconfigured word incorporation model and carries out conversion generation in end；
    Candidate information determining module, for by preconfigured training pattern, determining the corresponding candidate's letter of the sequence vector Breath, the training pattern are trained and obtained jointly with institute predicate incorporation model；
    Candidate information sending module, for the candidate information to be fed back to the intelligent terminal so that the intelligent terminal according to It is shown according to the candidate information.
  12. 12. a kind of intelligent terminal, it is characterised in that include memory, and one or more than one program, wherein one A either more than one program storage is in memory and is configured to perform described one by one or more than one processor A or more than one program includes the instruction for being used for being operated below：
    Received information is docked by preconfigured word incorporation model to be changed, and generates corresponding sequence vector, wherein, institute Predicate incorporation model is obtained with the common training of the training pattern being configured in server；
    The sequence vector is sent to the server, so that the server determines the vector by the training pattern The corresponding candidate information of sequence；
    Candidate information according to the server feedback is shown.
  13. A kind of 13. server, it is characterised in that include memory, and one or more than one program, one of them Either more than one program storage in memory and is configured to be performed by one or more than one processor one Or more than one program includes the instruction for being used for being operated below：
    The sequence vector that intelligent terminal is sent is received, the sequence vector is embedded in for the intelligent terminal by preconfigured word Model docks received information and carries out conversion generation；
    By preconfigured training pattern, determine the corresponding candidate information of the sequence vector, the training pattern with it is described Training obtains word incorporation model jointly；
    The candidate information is fed back into the intelligent terminal, so that the intelligent terminal is opened up according to the candidate information Show.
  14. A kind of 14. information processing system, it is characterised in that the system comprises：Server and intelligent terminal；
    Wherein, the server includes server as claimed in claim 13；
    Intelligent terminal includes intelligent terminal as claimed in claim 12.
CN201610974695.2A 2016-11-04 2016-11-04 Information processing method and device, intelligent terminal, server and system Active CN108024005B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610974695.2A CN108024005B (en) 2016-11-04 2016-11-04 Information processing method and device, intelligent terminal, server and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610974695.2A CN108024005B (en) 2016-11-04 2016-11-04 Information processing method and device, intelligent terminal, server and system

Publications (2)

Publication Number Publication Date
CN108024005A true CN108024005A (en) 2018-05-11
CN108024005B CN108024005B (en) 2020-08-21

Family

ID=62083652

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610974695.2A Active CN108024005B (en) 2016-11-04 2016-11-04 Information processing method and device, intelligent terminal, server and system

Country Status (1)

Country Link
CN (1) CN108024005B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108959421A (en) * 2018-06-08 2018-12-07 三角兽（北京）科技有限公司 Candidate replys evaluating apparatus and inquiry reverting equipment and its method, storage medium
CN111209725A (en) * 2018-11-19 2020-05-29 阿里巴巴集团控股有限公司 Text information generation method and device and computing equipment

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103765493A (en) * 2011-09-06 2014-04-30 英特尔公司 Number squaring computer-implemented method and apparatus
KR20140129443A (en) * 2013-04-29 2014-11-07 고려대학교 산학협력단 Adaptive music recommendation system and method
CN104506545A (en) * 2014-12-30 2015-04-08 北京奇虎科技有限公司 Data leakage prevention method and data leakage prevention device
CN104917669A (en) * 2015-05-29 2015-09-16 北京搜狗科技发展有限公司 Information processing method and device
CN104933113A (en) * 2014-06-06 2015-09-23 北京搜狗科技发展有限公司 Expression input method and device based on semantic understanding
CN105426354A (en) * 2015-10-29 2016-03-23 杭州九言科技股份有限公司 Sentence vector fusion method and apparatus
CN105637542A (en) * 2014-01-07 2016-06-01 富士施乐株式会社 Information processing device, information processing program, storage medium, and information processing method
CN105786782A (en) * 2016-03-25 2016-07-20 北京搜狗科技发展有限公司 Word vector training method and device
CN105915520A (en) * 2016-04-18 2016-08-31 深圳大学 File storage and searching method based on public key searchable encryption, and storage system
CN105989040A (en) * 2015-02-03 2016-10-05 阿里巴巴集团控股有限公司 Intelligent question-answer method, device and system

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103765493A (en) * 2011-09-06 2014-04-30 英特尔公司 Number squaring computer-implemented method and apparatus
KR20140129443A (en) * 2013-04-29 2014-11-07 고려대학교 산학협력단 Adaptive music recommendation system and method
CN105637542A (en) * 2014-01-07 2016-06-01 富士施乐株式会社 Information processing device, information processing program, storage medium, and information processing method
CN104933113A (en) * 2014-06-06 2015-09-23 北京搜狗科技发展有限公司 Expression input method and device based on semantic understanding
CN104506545A (en) * 2014-12-30 2015-04-08 北京奇虎科技有限公司 Data leakage prevention method and data leakage prevention device
CN105989040A (en) * 2015-02-03 2016-10-05 阿里巴巴集团控股有限公司 Intelligent question-answer method, device and system
CN104917669A (en) * 2015-05-29 2015-09-16 北京搜狗科技发展有限公司 Information processing method and device
CN105426354A (en) * 2015-10-29 2016-03-23 杭州九言科技股份有限公司 Sentence vector fusion method and apparatus
CN105786782A (en) * 2016-03-25 2016-07-20 北京搜狗科技发展有限公司 Word vector training method and device
CN105915520A (en) * 2016-04-18 2016-08-31 深圳大学 File storage and searching method based on public key searchable encryption, and storage system

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108959421A (en) * 2018-06-08 2018-12-07 三角兽（北京）科技有限公司 Candidate replys evaluating apparatus and inquiry reverting equipment and its method, storage medium
CN111209725A (en) * 2018-11-19 2020-05-29 阿里巴巴集团控股有限公司 Text information generation method and device and computing equipment
CN111209725B (en) * 2018-11-19 2023-04-25 阿里巴巴集团控股有限公司 Text information generation method and device and computing equipment

Also Published As

Publication number Publication date
CN108024005B (en) 2020-08-21

Similar Documents

Publication Publication Date Title
CN105162693B (en) message display method and device
CN110650370B (en) Video coding parameter determination method and device, electronic equipment and storage medium
CN105120122A (en) Alarm method and device
CN105049269B (en) information feedback method and device
CN105095366B (en) Word message treating method and apparatus
CN107845062A (en) image generating method and device
CN106775224B (en) Remark information setting method and device
CN105335062B (en) Information editing's operation processing method and device
CN108156026A (en) Intercom configuration method and device
EP2676426A1 (en) Identification of an alternate contact for use in reaching a mobile device user
CN106331293A (en) Incoming call information processing method and device
CN105049219B (en) Flow booking method and system, mobile terminal and server
CN106209604A (en) Add the method and device of good friend
CN106101433B (en) Notification message display methods and device
CN110113251A (en) Message coalescing method and device
CN110490164A (en) Generate the method, apparatus, equipment and medium of virtual expression
CN108710791A (en) The method and device of voice control
CN104822129B (en) Data transmission method and device
CN105101078A (en) Information prompting method, terminal and server
CN108024005A (en) Information processing method, device, intelligent terminal, server and system
CN107404429A (en) Circle of friends answering method, device and computer-readable recording medium
CN104363166B (en) Instant communicating method, device and intelligent terminal
CN104539497B (en) Method for connecting network and device
CN109450894A (en) Information interacting method, device, system, server user's terminal and storage medium
CN109120929A (en) A kind of Video coding, coding/decoding method, device, electronic equipment and system

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