CN108024005B

CN108024005B - Information processing method and device, intelligent terminal, server and system

Info

Publication number: CN108024005B
Application number: CN201610974695.2A
Authority: CN
Inventors: 涂畅; 张扬; 王砚峰
Original assignee: Beijing Sogou Technology Development Co Ltd
Current assignee: Beijing Sogou Technology Development Co Ltd
Priority date: 2016-11-04
Filing date: 2016-11-04
Publication date: 2020-08-21
Anticipated expiration: 2036-11-04
Also published as: CN108024005A

Abstract

The embodiment of the invention provides an information processing method, an information processing device, an intelligent terminal, a server and an information processing system, wherein the method comprises the following steps: converting the received information through a pre-configured word embedding model to generate a corresponding vector sequence, wherein the word embedding model is obtained by training together with a training model configured in a server; sending the vector sequence to the server so that the server determines candidate information corresponding to the vector sequence through the training model; and displaying according to the candidate information fed back by the server. According to the embodiment of the invention, the intelligent terminal uploads the converted vector sequence to the server, so that the server cannot restore the original information received by the intelligent terminal according to the converted vector sequence, and the problem of information leakage caused by directly uploading the received information is further avoided.

Description

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

Technical Field

The present invention relates to the field of input methods, and in particular, to an information processing method, an information processing apparatus, an intelligent terminal, a server, and an information processing system.

Background

With the development of mobile communication technology, intelligent terminals such as mobile phones are more and more popular, and great convenience is brought to life, study and work of people.

These smart terminals typically install various Applications (APPs) to implement their functions. For example, the intelligent terminal can reply to the received information through the APP with the intelligent reply function, namely, the intelligent reply service is provided, and the service requirement of the user is met.

Specifically, when the user uses the APP to have a conversation with another user, the APP can receive information sent by the other party, such as a short message, an email, chat content, and the like, and can send information to the other party, such as a reply to the information sent by the other party. When using the APP with the intelligent reply function, the intelligent terminal needs to upload information received by the user to the server to generate reply information corresponding to the information through the server, and then reply can be performed according to the reply information fed back by the server. Users typically refuse to upload received information to the server, i.e., refuse such services, in view of privacy concerns.

Disclosure of Invention

The technical problem to be solved by the embodiment of the invention is to provide an information processing method to solve the problem of information leakage in the process of uploading information to a server by the existing intelligent terminal, so that the information safety is ensured.

Correspondingly, the embodiment of the invention also provides an information processing device, an intelligent terminal, a server and an information processing system, which are used for ensuring the realization and the application of the method.

In order to solve the above problem, an embodiment of the present invention discloses an information processing method, including:

converting the received information through a pre-configured word embedding model to generate a corresponding vector sequence, wherein the word embedding model is obtained by training together with a training model configured in a server;

sending the vector sequence to the server so that the server determines candidate information corresponding to the vector sequence through the training model;

and displaying according to the candidate information fed back by the server.

Optionally, the converting the received information through the pre-configured word embedding model to generate a corresponding vector sequence includes:

when the received information is text information, performing word segmentation on the text information to obtain a word sequence after word segmentation;

converting the sequence of words into the sequence of vectors by the word embedding model.

Optionally, converting the word sequence into the vector sequence by using the word embedding model includes:

determining a first word vector corresponding to each word in the word sequence according to a word list;

converting a first word vector corresponding to each word into a second word vector through the word embedding model, wherein the dimension of the first word vector is larger than that of the second word vector;

and generating the vector sequence by adopting the second word vector corresponding to each word.

Optionally, the converting the received information to generate a corresponding vector sequence further includes:

when the received information is voice information, identifying the voice information to determine corresponding text information;

after the text information is determined, the step of segmenting the text information is performed.

Optionally, the candidate information includes: the server determines at least one candidate by an encoder and a decoder in the training model,

the displaying according to the candidate information fed back by the server includes:

receiving candidate information fed back by the server for the vector sequence;

and displaying the candidate items in the candidate information based on a preset display rule.

Optionally, the method further includes:

determining a selected candidate item according to the operation instruction;

inputting the selected candidate item into a dialog box of a communication application to reply to the information, wherein the communication application comprises at least one of the following types: short message application, instant messaging application, mail application, and telecommunication application.

The embodiment of the invention also discloses another information processing method, which comprises the following steps:

receiving a vector sequence sent by an intelligent terminal, wherein the vector sequence is generated by converting received information through a pre-configured word embedding model by the intelligent terminal;

determining candidate information corresponding to the vector sequence through a pre-configured training model, wherein the training model and the word embedding model are obtained through training together;

and feeding back the candidate information to the intelligent terminal so that the intelligent terminal can display according to the candidate information.

Optionally, the determining candidate information corresponding to the vector sequence through a pre-configured training model includes:

inputting the vector sequence into an encoder in the training model to obtain a corresponding encoding vector sequence;

inputting the coding vector sequence into a decoder in the training model to obtain at least one candidate item;

and determining the candidate information based on the obtained candidate.

Optionally, the candidate item is used for the intelligent terminal to reply in a dialog box of a communication application according to the information;

wherein the communication application comprises at least one of the following types: short message application, instant messaging application, mail application, and telecommunication application.

The embodiment of the invention also discloses an information processing device, which comprises:

the vector sequence generation module is used for converting the received information through a pre-configured word embedding model to generate a corresponding vector sequence, wherein the word embedding model is obtained by training together with a training model configured in the server;

the vector sequence sending module is used for sending the vector sequence to the server so that the server can determine candidate information corresponding to the vector sequence through the training model;

and the candidate information display module is used for displaying according to the candidate information fed back by the server.

Optionally, the vector sequence generating module includes the following sub-modules: the word segmentation sub-module is used for segmenting the text information to obtain a word sequence after word segmentation when the received information is the text information; and the conversion sub-module is used for converting the word sequence into the vector sequence through a pre-configured word embedding model.

Optionally, the conversion sub-module is specifically configured to determine, according to a word list, a first word vector corresponding to each word in the word sequence; converting a first word vector corresponding to each word into a second word vector through the word embedding model, wherein the dimension of the first word vector is larger than that of the second word vector; and generating the vector sequence by adopting the second word vector corresponding to each word.

Optionally, the vector sequence generating module further includes: a sub-module is identified. The recognition sub-module is used for recognizing the voice information to determine corresponding text information when the received information is the voice information; and after the text information is determined, triggering the word segmentation submodule to perform word segmentation on the text information.

Optionally, the candidate information includes: the server determines at least one candidate by an encoder and a decoder in the training model. The candidate information display module comprises the following sub-modules: a receiving submodule, configured to receive candidate information fed back by the server for the vector sequence; and the display sub-module is used for displaying each candidate item in the candidate information based on a preset display rule.

Optionally, the information processing apparatus in this embodiment of the present application may further include the following modules: the candidate item selecting module is used for determining a selected candidate item according to the operation instruction; and the information reply module is used for inputting the selected candidate item into a dialog box of the communication application so as to reply the information. Wherein the communication application comprises at least one of the following types: short message application, instant messaging application, mail application, and telecommunication application.

The embodiment of the invention also discloses another information processing device, which comprises:

the intelligent terminal comprises a vector sequence receiving module, a word embedding module and a word embedding module, wherein the vector sequence receiving module is used for receiving a vector sequence sent by the intelligent terminal, and the vector sequence is generated by converting received information through a pre-configured word embedding model by the intelligent terminal;

the candidate information determining module is used for determining candidate information corresponding to the vector sequence through a pre-configured training model, and the training model and the word embedding model are obtained through training together;

and the candidate information sending module is used for feeding back the candidate information to the intelligent terminal so that the intelligent terminal can display according to the candidate information.

Optionally, the candidate information determining module includes the following sub-modules: the coding submodule is used for inputting the vector sequence into a coder in the training model to obtain a corresponding coding vector sequence; a decoding submodule, configured to input the coded vector sequence to a decoder in the training model, so as to obtain at least one candidate; and the determining submodule is used for determining the candidate information based on the obtained candidate.

Optionally, the candidate is specifically used for the intelligent terminal to reply in a dialog box of the communication application according to the information. The communication application specifically comprises at least one of the following types: short message application, instant messaging application, mail application, and telecommunication application.

The embodiment of the invention also discloses an intelligent terminal, which comprises a memory and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by one or more processors, and the one or more programs comprise instructions for:

and displaying according to the candidate information fed back by the server.

The embodiment of the invention also discloses a server, which comprises a memory and one or more programs, wherein the one or more programs are stored in the memory, and the one or more programs are configured to be executed by one or more processors and comprise instructions for:

The embodiment of the invention also discloses an information processing system, which comprises: a server and an intelligent terminal; wherein, the server comprises the server according to the embodiment of the invention; the intelligent terminal comprises the intelligent terminal according to the embodiment of the invention.

Compared with the prior art, the embodiment of the invention has the following advantages:

in the embodiment of the invention, a word embedding model and a training model obtained by co-training are respectively configured in an intelligent terminal and a server in advance, so that the intelligent terminal can convert received information through the pre-configured word embedding model, generate a converted vector sequence, upload the vector sequence to the server, and trigger the server to determine candidate information corresponding to the vector sequence through the training model, so that the intelligent terminal can acquire the candidate information fed back by the server; and the converted vector sequence can not be restored by the server, namely the server can not restore the original information according to the vector sequence, so that the server can not check the original information received by the intelligent terminal, the problem of information leakage caused by directly uploading the received information is avoided, and the privacy of a user is effectively protected.

Drawings

FIG. 1 is a flow chart of the steps of an embodiment of the invention at the intelligent terminal side in an information processing method;

FIG. 2 is a flow chart of the steps of a server-side embodiment of an information processing method of the present invention;

FIG. 3 is a schematic diagram of the interaction between a server and an intelligent terminal according to an embodiment of the present invention;

FIG. 4 is a block diagram of an embodiment of an information processing apparatus according to the present invention;

FIG. 5 is a block diagram of another embodiment of an information processing apparatus according to the present invention;

FIG. 6 is a block diagram illustrating an architecture of an intelligent terminal for information processing, according to an exemplary embodiment;

fig. 7 is a schematic structural diagram of a server in an embodiment of the present invention.

Detailed Description

In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.

When using an APP with an intelligent reply function, an intelligent terminal serving as a client needs to upload information received by a user, such as a short message, an email, chat content, and the like sent by an opposite party, to a server serving as a server, so as to generate candidate information corresponding to the information through the server. The server returns the generated candidate information to the intelligent terminal, so that the intelligent terminal can reply to the received information according to the returned candidate information. However, the intelligent terminal directly uploads the information received by the user to the server, so that the risk of information leakage exists, and even the privacy of the user may be leaked, that is, the problem of information security exists.

One of the core ideas of the embodiments of the present invention is to pre-configure a word embedding model and a training model obtained by co-training in an intelligent terminal and a server, respectively, so that the intelligent terminal can convert received information into a vector sequence through the pre-configured word embedding model, and send the converted vector sequence to the server, so that the server determines candidate information corresponding to the converted vector sequence through the training model, and thus the intelligent terminal can obtain the candidate information fed back by the server, and the converted vector sequence cannot be restored by the server, i.e., the server cannot restore original information according to the vector sequence, thereby avoiding the problem of information leakage caused by directly uploading the received information to the server, and effectively protecting the privacy of a user.

Referring to fig. 1, a flowchart illustrating steps of an embodiment of an intelligent terminal side in an information processing method according to the present invention is shown, which may specifically include the following steps:

and 102, converting the received information through a pre-configured word embedding model to generate a corresponding vector sequence, wherein the word embedding model is obtained by training together with a training model configured in a server.

The application program installed in the intelligent terminal can be called a client, for example, a communication application for establishing a conversation by a user can be used as the client. The communication application may specifically include: short message application, instant messaging application, mail application, telecommunication application, etc., which are not specifically limited in the embodiments of the present application. The short message application can be used for receiving and/or sending short messages; the instant messaging application can be used for instantly sending and receiving information transmitted by a network, such as internet information; the mail application may be used to receive and/or send mail; the telecommunication application is an application for making a call in a telecommunication manner, such as an application for making a call in a terminal.

After the user logs in the client, the intelligent terminal can receive the information sent to the user through the client and can display the received information, so that the user can check the information, and the user can reply to the received information. After receiving the information, the intelligent terminal can convert the received information into a corresponding vector sequence through a pre-configured word embedding model, and uploads the information to the server in a vector form.

And 104, sending the vector sequence to a server so that the server determines candidate information corresponding to the vector sequence through the training model.

After the intelligent terminal generates the converted vector sequence, the converted vector sequence can be uploaded to a server through a network, so that candidate information corresponding to the converted vector sequence is generated through the server. After receiving the vector sequence, the server can determine candidate information corresponding to the converted vector sequence through a pre-configured training model, and feed back the candidate information to the intelligent terminal.

And 106, displaying according to the candidate information fed back by the server.

The candidate information may specifically include: and the server determines each candidate item corresponding to the information according to the vector sequence. Specifically, after receiving the candidate information fed back by the server, the intelligent terminal can display the candidate information according to a preset display rule, so that the user can reply to the received information by using the candidate information or a certain candidate item in the candidate information, that is, the user can reply to the information conveniently, and the operation simplicity is improved.

In the embodiment of the invention, the word embedding model configured in the intelligent terminal and the training model configured in the server are obtained by training together. The word embedding model is configured in the intelligent terminal, so that the intelligent terminal can convert the received information into a vector sequence through the word embedding model, and upload the converted vector sequence to the server. Obviously, the vector sequence is obtained by converting the word embedding model of the intelligent terminal, so that the vector sequence cannot be restored by the server, namely, the server cannot restore the original information according to the vector sequence, and the server cannot check the original information received by the intelligent terminal, thereby avoiding the problem of information leakage caused by directly uploading the received information, and effectively protecting the privacy of a user. In addition, the training model pre-configured in the server is obtained by training together with the word embedding model in the intelligent terminal, so that when the server receives the vector sequence uploaded by the intelligent terminal, candidate information corresponding to the vector sequence can be determined through the pre-configured training model, and the candidate information is fed back to the intelligent terminal, so that the intelligent terminal can display according to the candidate information fed back by the server.

Referring to fig. 2, a flowchart illustrating steps of a server-side embodiment in an information processing method according to the present invention is shown, and specifically may include the following steps:

step 202, receiving a vector sequence sent by the intelligent terminal.

The vector sequence is generated by converting the received information through a pre-configured word embedding model by the intelligent terminal.

And 204, determining candidate information corresponding to the vector sequence through a pre-configured training model.

The training model configured by the server and the word embedding model configured by the intelligent terminal are obtained through training together. The server can take the vector sequence uploaded by the intelligent terminal as input, and candidate information corresponding to the vector sequence is determined through the training model. Preferably, the candidate information may specifically include each candidate corresponding to the information. Specifically, after receiving the vector sequence, the server may use the vector sequence as an input to generate corresponding candidate items, such as generating one or several candidate items intelligently replied, by inputting the vector sequence to a pre-configured training model; and the generated candidate can be used as candidate information to feed back the candidate information to the intelligent terminal.

And step 206, feeding back the candidate information to the intelligent terminal so that the intelligent terminal can display according to the candidate information.

The server feeds the candidate information back to the intelligent terminal, namely the candidate information is sent to the client operated by the intelligent terminal, so that the intelligent terminal can display according to the candidate information, and further, each candidate item in the candidate information can be recommended to the user. The user can select any one of the candidate items from the candidate items displayed by the intelligent terminal and reply to the received information. The intelligent terminal can send the candidate item selected by the user to the intelligent terminal used by the other party through the client so as to reply the information sent by the other party, and the user does not need to input each word in the reply information, so that the user can reply the information conveniently, and the convenience and the efficiency of replying the information by the user are improved.

In the embodiment of the invention, the server receives the converted vector sequence, and the originally received information of the intelligent terminal cannot be restored according to the converted vector sequence, so that the original information cannot be checked, the problem of information leakage in the process of uploading information to the server by the existing intelligent terminal is solved, and the privacy of a user can be effectively protected.

As a specific application of the invention, an intelligent reply model can be trained with a large amount of data based on a deep learning model algorithm, such as a Long Short-Term Memory artificial neural network (LSTM) algorithm, in combination with a word embedding technique, so that the intelligent terminal can reply to the received information through the intelligent reply model. In essence, the intelligent model may specifically include a word embedding model and a training model; the word embedding model can be configured on an intelligent terminal serving as a client, so that the intelligent terminal can convert received information into a vector sequence through the word embedding model; the training model may be configured in the server, so that the server may determine, through the training model, candidate information corresponding to the converted vector sequence uploaded by the intelligent terminal. Therefore, the intelligent terminal can determine the candidate information through the server and can show each candidate item in the candidate information to the user so as to recommend each candidate item in the candidate information to the user, and therefore the user can select any one of the candidate items to reply to the received information; or the candidate item can be selected according to a preset reply strategy so as to automatically reply the received information, thereby greatly improving the operation simplicity of replying the information by the user, and meanwhile, on the premise of effectively protecting the privacy of the user, the situation that the intelligent reply service is limited by the computing power of an intelligent terminal such as a mobile phone can be avoided, and the user experience is improved.

It should be noted that, in the embodiment of the present application, other algorithms may also be used to implement an intelligent reply model, such as a Recurrent Neural Network (RNN), or a deep learning algorithm derived from the RNN, such as a bidirectional LSTM, a bidirectional RNN, an LSTM with an Attention mechanism, a CopyNet, and the like, which is not limited in this embodiment of the present application.

Referring to fig. 3, a schematic diagram of interaction between a server and an intelligent terminal in an embodiment of the present invention is shown. The information processing method provided by the embodiment of the application specifically comprises the following steps:

step 302, the intelligent terminal converts the received information through a pre-configured word embedding model to generate a corresponding vector sequence.

In the embodiment of the application, the intelligent terminal can receive various information through communication application; the information may include various data, such as picture data, text data, video data, audio data, and the like, which is not limited by the embodiment of the present application. For example, the intelligent terminal may receive text information and/or voice information through the communication application, where the text information may specifically include text data, and the voice information may specifically include audio data.

In an optional embodiment of the present application, the converting the received information through the pre-configured word embedding model to generate a corresponding vector sequence may specifically include: when the received information is text information, performing word segmentation on the text information to obtain a word sequence after word segmentation; and converting the word sequence into the vector sequence through a pre-configured word embedding model. Specifically, after receiving the text information, the intelligent terminal may perform word segmentation on the text information to obtain each word and/or word in the text information, and may combine the obtained words into a corresponding word sequence, and may input a word vector corresponding to each word in the word sequence before conversion into a pre-configured word embedding model according to an arrangement order of each word in the word sequence, so as to generate a word vector corresponding to each word after conversion, that is, the word sequence is converted into a vector sequence by a pre-trained word embedding model.

As a specific example of the present application, after receiving text information, an intelligent terminal may obtain a word sequence after word segmentation by performing word segmentation on the text information. Each word in the sequence of words may be represented by a vector, e.g., a word is represented by a vector an; n may be used to indicate the sequence of words in the text message, and may be an integer. The intelligent terminal can represent each word in the word sequence by adopting a vector, so that a first vector sequence A (a1, a2, a3 … …, an) corresponding to the word segmentation can be obtained; and the first vector sequence a (a1, a2, a3 … …, an) can be converted into the second vector sequence B (B1, B2, B3 … …, bn) by word embedding techniques. For example, the smart terminal may divide a sentence into a plurality of words by word segmentation, and may be represented by a first vector sequence a (a1, a 2.., an). After the word embedding model, the first vector sequence a (a1, a 2.., an) may be converted into another vector sequence B (B1, B2, B3 … …, bn), that is, into a second vector sequence B (B1, B2, B3 … …, bn).

It should be noted that, in order to distinguish between word vectors before and after conversion, the present application may refer to a word vector before conversion as a first word vector, and refer to a word vector after conversion as a second word vector, for example, each word vector in the first vector sequence a is taken as a first word vector corresponding to each word, and each word vector in the second vector sequence B is taken as a second word vector corresponding to each word. The dimensionality of the first word vector is not equal to the dimensionality of the second word vector, namely, the intelligent terminal converts the reducible common first word vector into the unreducible second word vector through the word embedding model, so that the server cannot reduce each word in the word sequence through the second word vector, namely, the server cannot obtain the original information received by the intelligent terminal.

In an optional embodiment of the present application, the intelligent terminal converts the word sequence into the vector sequence through a pre-trained word embedding model, which may specifically include: determining a first word vector corresponding to each word in the word sequence according to a word list; converting a first word vector corresponding to each word into a second word vector through the word embedding model, wherein the dimension of the first word vector is larger than that of the second word vector; and generating the vector sequence by adopting the second word vector corresponding to each word.

As a specific example of the present application, the intelligent terminal may determine a first word vector an (0, 0, 1, 0, 0, 0, … …, 0) corresponding to each word after word segmentation by performing word segmentation on the received text information and according to a word list, and may convert the first word vector an (0, 0, 1, 0, 0, 0, … …, 0) into a second word vector bn (0.01, 0.03, 0.13, … …, 0.2) through a preset word embedding model M, for example, convert the first vector an into the second vector bn through a calculation formula bn ═ M ×. The word embedding model M may be a matrix with x × y dimensions, for example, if x is 100 and y is 50000, the entire matrix M may include 5000000 elements. x may be used as a converted dimension, and may be specifically used to characterize a dimension of the second word vector in the second vector sequence B, such as a dimension that may be equal to B1; y may be used as a dimension before conversion, and may be specifically used to characterize a dimension of the first word vector in the first vector sequence a, such as a dimension that may be equal to a 1.

It should be noted that if the dimension y before conversion is too large, the word embedding model is very large, which results in more parameters of the whole model and affects the training speed and the calculation speed. Optionally, the dimension y before conversion may be configured according to the size of the vocabulary, and for example, the dimension y before conversion may be equal to the size of the vocabulary, specifically, may be tens of thousands; the dimension of the second word vector can be configured according to a specific algorithm adopted in the word embedding model, and can be usually smaller than the dimension of the first word vector, preferably, the dimension x after conversion can be usually smaller than the dimension y before conversion, such as thousands or hundreds, or even dozens, so that useful information can be represented in a more compact form, the processing at the server end is convenient, the server cannot restore the original information received by the intelligent terminal through the converted second vector sequence, and the problem of information leakage can be avoided.

In an optional embodiment of the present application, the converting, by the intelligent terminal, the received information to generate a corresponding vector sequence may further include: when the received information is voice information, identifying the voice information to determine corresponding text information; after the text information is determined, the step of segmenting the text information is performed. In the embodiment of the application, when the intelligent terminal receives the voice information, the voice information can be converted into the text information, so that a first vector sequence corresponding to word segmentation can be obtained by executing the step of word segmentation on the text information, and the first vector sequence can be converted into a second vector sequence through a word embedding model.

And step 304, the intelligent terminal sends the vector sequence to a server.

After the intelligent terminal obtains the converted second vector sequence, the second vector sequence can be used as a vector sequence and is sent to the server, so that the server can receive the vector sequence.

And step 306, the server receives the vector sequence sent by the intelligent terminal.

Specifically, the server receives the vector sequence converted by the intelligent terminal, and the vector sequence cannot be restored to original information, so that the server cannot acquire the information received by the intelligent terminal, the risk of revealing user information can be avoided, and the privacy safety of a user is effectively protected.

And 308, the server determines candidate information corresponding to the vector sequence through a pre-configured training model.

In this embodiment, the training model configured in the server may include a decoder and an encoder, so as to encode the received vector sequence through the decoder to obtain an encoded vector sequence, that is, an encoded vector sequence, and then decode the encoded vector sequence through the decoder to obtain the corresponding candidate. The number of the candidate items can be configured by adopting parameters in the training model, and the number of the candidate items is not particularly limited in the implementation of the application. Therefore, the candidate information determined by the server may specifically include: at least one candidate determined by an encoder and a decoder in the training model. The candidate item can be specifically used for the intelligent terminal to reply in a dialog box of the communication application according to the information. The communication application may specifically include: short message application, instant messaging application, mail application, telecommunication application, etc., which are not limited in the embodiments of the present invention.

In an optional embodiment of the present application, the step of determining, by a server, candidate information corresponding to the vector sequence through a pre-configured training model may specifically include: inputting the vector sequence into an encoder in the training model to obtain an encoded vector sequence; the coded vector sequence is then input to a decoder in the training model to derive at least one candidate, so that candidate information may be determined based on the derived candidate. As a specific example of the present invention, the training model configured by the server generates a model, such as a Sequence to Sequence (Sequence to Sequence) model, for the Sequence in deep learning. In the seq2seq model, the server may employ one LSTM model as an encoder (encoder) and another LSTM model as a decoder (decoder). After the intelligent terminal uploads the converted second vector sequence B (B1, B2, B3 … …, bn) to the server, the server inputs the second vector sequence B (B1, B2, B3 … …, bn) to an encoder in the training model to be encoded by the encoder to obtain an encoded encoding vector sequence C (C1, C2, C3 … …, cn), then inputs the encoding vector sequence C (C1, C2, C3 … …, cn) to a decoder in the training model to be decoded by the decoder, and generates a certain number of candidates by using an algorithm such as beam search (beam search), namely at least one candidate; and the obtained candidate item can be used as candidate information to be sent to the intelligent terminal so as to feed back the second vector sequence uploaded by the intelligent terminal.

And step 310, the server feeds the candidate information back to the intelligent terminal.

The server can return the generated candidate information to the intelligent terminal, so that the intelligent terminal can display according to the candidate information.

And step 312, the intelligent terminal displays the candidate information fed back by the server.

Wherein the candidate information includes: and the server determines each candidate item corresponding to the information according to the vector sequence.

In an optional embodiment of the present application, the displaying, by the intelligent terminal, according to the candidate information fed back by the server may specifically include: receiving candidate information fed back by the server for the vector sequence; and displaying each candidate item in the candidate information based on a preset display rule.

After receiving the candidate information fed back by the server, the intelligent terminal may extract each candidate item from the candidate information, and may determine a display order of each candidate item based on a preset display rule, for example, according to a priority order of each candidate item, so as to display each candidate item in a designated area according to the display order, for example, display each candidate item in a candidate area of an input method, or display each candidate item in a display area associated with a dialog box in a communication application, which is not limited in the implementation of the present application.

And step 314, the intelligent terminal determines the selected candidate item according to the operation instruction.

The intelligent terminal displays the candidate items, so that the user can acquire the candidate items recommended by the server for the received information. The user can input an operation instruction to the intelligent terminal to trigger the intelligent terminal to select a certain candidate item according to the operation instruction, namely, the candidate item selected by the user is determined.

And step 316, the intelligent terminal inputs the selected candidate item into a dialog box of the communication application so as to reply the information.

Wherein the communication application may include, but is not limited to, at least one of the following types: short message application, instant messaging application, mail application, and telecommunication application.

After determining the candidate item selected by the user, the intelligent terminal can input the selected candidate item into a dialog box in the communication application, so that the user can reply to the received information according to the selected candidate item, the convenience of reply operation is improved, and the reply efficiency of the user is improved.

Of course, the intelligent terminal can modify the candidate item according to the editing operation or the modifying operation input by the user for the selected candidate item, so as to reply the received information according to the modified candidate item, thereby improving the user experience.

It should be noted that, for simplicity of description, the method embodiments are described as a series of acts or combination of acts, but those skilled in the art will recognize that the present invention is not limited by the illustrated order of acts, as some steps may occur in other orders or concurrently in accordance with the embodiments of the present invention. Further, those skilled in the art will appreciate that the embodiments described in the specification are presently preferred and that no particular act is required to implement the invention.

Referring to fig. 4, a block diagram of an embodiment of an information processing apparatus according to the present invention is shown, and may specifically include the following modules:

a vector sequence generating module 402, configured to convert the received information through a pre-configured word embedding model, and generate a corresponding vector sequence, where the word embedding model is obtained by training together with a training model configured in the server.

A vector sequence sending module 404, configured to send the vector sequence to the server, so that the server determines, through the training model, candidate information corresponding to the vector sequence.

And a candidate information display module 406, configured to display according to the candidate information fed back by the server.

In an optional embodiment of the present application, the vector sequence generating module 402 may include the following sub-modules:

and the word segmentation sub-module is used for segmenting the text information to obtain a word sequence after word segmentation when the received information is the text information.

And the conversion sub-module is used for converting the word sequence into the vector sequence through a pre-configured word embedding model.

In an optional embodiment of the present application, the conversion sub-module may be specifically configured to determine, according to a word list, a first word vector corresponding to each word in the word sequence; converting a first word vector corresponding to each word into a second word vector through the word embedding model, wherein the dimension of the first word vector is not equal to the dimension of the second word vector; and generating the vector sequence by adopting the second word vector corresponding to each word. Optionally, the dimension of the first word vector is larger than the dimension of the second word vector.

In an optional embodiment of the present application, the vector sequence generating module 402 may further include: a sub-module is identified. The recognition sub-module may be configured to, when the received information is voice information, determine corresponding text information by recognizing the voice information; and after the text information is determined, triggering the word segmentation sub-module to execute the step of segmenting the text information.

In an optional embodiment of the present application, the candidate information may include: the server determines at least one candidate by an encoder and a decoder in the training model. The candidate information presentation module 406 may include the following sub-modules:

and the receiving submodule is used for receiving the candidate information fed back by the server aiming at the vector sequence.

And the display sub-module is used for displaying each candidate item in the candidate information based on a preset display rule.

Optionally, the information processing apparatus in this embodiment of the present application may further include the following modules:

the candidate item selecting module is used for determining a selected candidate item according to the operation instruction;

and the information reply module is used for inputting the selected candidate item into a dialog box of the communication application so as to reply the information.

The information processing device in the embodiment of the application can be particularly applied to the intelligent terminal in the embodiment of the method, so that the received information can be converted into a vector sequence to be uploaded to the server, the server cannot restore the information received by the intelligent terminal according to the converted vector sequence, and the problem of information leakage caused by uploading the received information to the server is further avoided; and the candidate information fed back by the server can be received, and the candidate items in the candidate information can be recommended to the user, so that the user can reply to the received information by adopting the candidate items generated by the server, and the efficiency of replying the information by the user can be improved.

Referring to fig. 5, a block diagram of another information processing apparatus according to another embodiment of the present invention is shown, and may specifically include the following modules:

a vector sequence receiving module 502, configured to receive a vector sequence sent by an intelligent terminal, where the vector sequence is generated by converting received information through a pre-configured word embedding model by the intelligent terminal;

a candidate information determining module 504, configured to determine candidate information corresponding to the vector sequence through a pre-configured training model, where the training model is obtained by training together with the word embedding model;

and a candidate information sending module 506, configured to feed back the candidate information to the intelligent terminal, so that the intelligent terminal performs display according to the candidate information.

In an optional embodiment of the present application, the candidate information determining module 504 may specifically include the following sub-modules:

the coding submodule is used for inputting the vector sequence into a coder in the training model to obtain a corresponding coding vector sequence;

a decoding submodule, configured to input the coded vector sequence to a decoder in the training model, so as to obtain at least one candidate item;

and the determining submodule is used for determining the candidate information based on the obtained candidate items.

In an optional embodiment of the present application, the candidate may be specifically used for the intelligent terminal to reply to the information in a dialog box of a communication application. The communication application may specifically include, but is not limited to, at least one of the following types: short message application, instant messaging application, mail application, and telecommunication application.

The information processing apparatus in the embodiment of the application is specifically applicable to the server in the above method embodiment, and the converted vector sequence uploaded by the intelligent terminal may be input into a pre-configured training model to obtain corresponding candidate information, and the candidate information is fed back to the intelligent terminal, so that the intelligent terminal may display according to the candidate information, and thus, a user may reply to the received information by using the candidate item generated by the server, and the efficiency of replying information by the user is improved.

For the device embodiment, since it is basically similar to the method embodiment, the description is simple, and for the relevant points, refer to the partial description of the method embodiment.

Fig. 6 is a block diagram illustrating a structure of an intelligent terminal 600 for information processing according to an exemplary embodiment. For example, the smart terminal 600 may be a mobile phone, a computer, a digital broadcast terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, and the like.

Referring to fig. 6, the smart terminal 600 may include one or more of the following components: processing component 602, memory 604, power component 606, multimedia component 608, audio component 610, input/output (I/O) interface 612, sensor component 614, and communication component 616.

The processing component 602 generally controls the overall operation of the smart terminal 600, such as operations associated with display, telephone calls, data communications, camera operations, and recording operations. The processing elements 602 may include one or more processors 620 to execute instructions to perform all or a portion of the steps of the methods described above. Further, the processing component 602 can include one or more modules that facilitate interaction between the processing component 602 and other components. For example, the processing component 602 can include a multimedia module to facilitate interaction between the multimedia component 608 and the processing component 602.

The memory 604 is configured to store various types of data to support operation at the device 600. Examples of such data include instructions for any application or method operating on the smart terminal 600, contact data, phonebook data, messages, pictures, videos, and the like. The memory 604 may be implemented by any type or combination of volatile or non-volatile memory devices such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disks.

The power component 606 provides power to the various components of the smart terminal 600. The power components 606 may include a power management system, one or more power sources, and other components associated with generating, managing, and distributing power for the smart terminal 600.

The multimedia component 608 includes a screen providing an output interface between the intelligent terminal 600 and the user. In some embodiments, the screen may include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive an input signal from a user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundary of a touch or slide action, but also detect the duration and pressure associated with the touch or slide operation. In some embodiments, the multimedia component 608 includes a front facing camera and/or a rear facing camera. When the smart terminal 600 is in an operation mode, such as a shooting mode or a video mode, the front camera and/or the rear camera may receive external multimedia data. Each front camera and rear camera may be a fixed optical lens system or have a focal length and optical zoom capability.

The audio component 610 is configured to output and/or input audio signals. For example, the audio component 610 includes a Microphone (MIC) configured to receive external audio signals when the smart terminal 600 is in an operating mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signal may further be stored in the memory 604 or transmitted via the communication component 616. In some embodiments, audio component 610 further includes a speaker for outputting audio signals.

The I/O interface 612 provides an interface between the processing component 602 and peripheral interface modules, which may be keyboards, click wheels, buttons, etc. These buttons may include, but are not limited to: a home button, a volume button, a start button, and a lock button.

The sensor component 614 includes one or more sensors for providing various aspects of status assessment for the smart terminal 600. For example, the sensor component 614 may detect an open/closed state of the device 600, the relative positioning of components, such as a display and keypad of the smart terminal 600, the sensor component 614 may also detect a change in the position of the smart terminal 600 or a component of the smart terminal 600, the presence or absence of user contact with the smart terminal 600, orientation or acceleration/deceleration of the smart terminal 600, and a change in the temperature of the smart terminal 600. The sensor assembly 614 may include a proximity sensor configured to detect the presence of a nearby object without any physical contact. The sensor assembly 614 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 614 may also include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

The communication component 616 is configured to facilitate communications between the smart terminal 600 and other devices in a wired or wireless manner. The smart terminal 600 may access a wireless network based on a communication standard, such as WiFi, 2G or 3G, or a combination thereof. In an exemplary embodiment, the communication component 616 receives broadcast signals or broadcast related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 616 further includes a Near Field Communication (NFC) module to facilitate short-range communications. For example, the NFC module may be implemented based on Radio Frequency Identification (RFID) technology, infrared data association (IrDA) technology, Ultra Wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

In an exemplary embodiment, the smart terminal 600 may be implemented by one or more Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), Programmable Logic Devices (PLDs), Field Programmable Gate Arrays (FPGAs), controllers, micro-controllers, microprocessors or other electronic components for performing the above-described methods.

In an exemplary embodiment, a non-transitory computer readable storage medium including instructions, such as the memory 604 including instructions, executable by the processor 620 of the smart terminal 600 to perform the above-described method is also provided. For example, the non-transitory computer readable storage medium may be a ROM, a Random Access Memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like.

A non-transitory computer-readable storage medium in which instructions, when executed by a processor of a smart terminal, enable the smart terminal to perform an information processing method, the method comprising: converting the received information through a pre-configured word embedding model to generate a corresponding vector sequence, wherein the word embedding model is obtained by training together with a training model configured in a server; sending the vector sequence to the server so that the server determines candidate information corresponding to the vector sequence through the training model; and displaying according to the candidate information fed back by the server.

Fig. 7 is a schematic structural diagram of a server in an embodiment of the present invention. The server 700 may vary widely in configuration or performance and may include one or more Central Processing Units (CPUs) 722 (e.g., one or more processors) and memory 732, one or more storage media 730 (e.g., one or more mass storage devices) storing applications 742 or data 744. Memory 732 and storage medium 730 may be, among other things, transient storage or persistent storage. The program stored in the storage medium 730 may include one or more modules (not shown), each of which may include a series of instruction operations for the server. Further, the central processor 722 may be configured to communicate with the storage medium 730, and execute a series of instruction operations in the storage medium 730 on the server 700.

The server 700 may also include one or more power supplies 726, one or more wired or wireless network interfaces 750, one or more input-output interfaces 758, one or more keyboards 756, and/or one or more operating systems 741, such as Windows Server, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM, etc.

A non-transitory computer-readable storage medium in which instructions, when executed by a processor of a server, enable the server to perform a method of information processing, the method comprising: receiving a vector sequence sent by an intelligent terminal, wherein the vector sequence is generated by converting received information through a pre-configured word embedding model by the intelligent terminal; determining candidate information corresponding to the vector sequence through a pre-configured training model, wherein the training model and the word embedding model are obtained through training together; and feeding back the candidate information to the intelligent terminal so that the intelligent terminal can display according to the candidate information.

On the basis of the foregoing embodiments, an embodiment of the present application further provides an information processing system, which may specifically include: a server and an intelligent terminal; the server is basically the same as the server described in the above embodiment, and the intelligent terminal is basically the same as the intelligent terminal described in the above embodiment, so the description is omitted.

In the information processing system of the embodiment of the application, the word embedding model configured by the intelligent terminal and the training model configured by the server are obtained through training together. The intelligent terminal can convert the received information into a corresponding vector sequence through a pre-configured word embedding model, and upload the converted vector sequence to the server, and the trigger server determines the candidate information corresponding to the converted vector sequence through a pre-configured training model, so that the candidate information fed back by the server can be obtained, and the server cannot restore the original information received by the intelligent terminal according to the converted vector sequence, namely cannot check the original information received by the intelligent terminal, so that the problem of information leakage caused by directly uploading the received information is avoided, and the privacy of a user is effectively protected.

The embodiments in the present specification are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other.

As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, apparatus, or computer program product. Accordingly, embodiments of the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, embodiments of the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

Embodiments of the present invention are described with reference to flowchart illustrations and/or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing terminal to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing terminal to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing terminal to cause a series of operational steps to be performed on the computer or other programmable terminal to produce a computer implemented process such that the instructions which execute on the computer or other programmable terminal provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

While preferred embodiments of the present invention have been described, additional variations and modifications of these embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all such alterations and modifications as fall within the scope of the embodiments of the invention.

Finally, it should also be noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or terminal. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or terminal that comprises the element.

The information processing method, the information processing apparatus, the intelligent terminal, the server and the information processing system provided by the present invention are introduced in detail, and a specific example is applied in the text to explain the principle and the implementation of the present invention, and the description of the above embodiment is only used to help understanding the method and the core idea of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.

Claims

1. An information processing method characterized by comprising:

and displaying according to the candidate information fed back by the server.

2. The method of claim 1, wherein transforming the received information through a pre-configured word embedding model to generate a corresponding sequence of vectors comprises:

3. The method of claim 2, wherein converting the sequence of words into the sequence of vectors via the word embedding model comprises:

4. The method of claim 2 or 3, wherein converting the received information to generate a corresponding sequence of vectors further comprises:

5. The method according to any one of claims 1 to 3, wherein the candidate information comprises: the server determines at least one candidate by an encoder and a decoder in the training model,

receiving candidate information fed back by the server for the vector sequence;

6. The method of claim 5, further comprising:

determining a selected candidate item according to the operation instruction;

7. An information processing method characterized by comprising:

determining candidate information corresponding to the vector sequence through a pre-configured training model, wherein the training model and the word embedding model are obtained through co-training;

8. The method according to claim 7, wherein the determining candidate information corresponding to the vector sequence through a pre-configured training model comprises:

and determining the candidate information based on the obtained candidate.

9. The method of claim 8, wherein the candidate is used for the intelligent terminal to reply to the message in a dialog box of a messaging application;

10. An information processing apparatus characterized by comprising:

11. The apparatus of claim 10, wherein the vector sequence generation module comprises:

the word segmentation sub-module is used for segmenting the text information to obtain a word sequence after word segmentation when the received information is the text information;

12. The apparatus of claim 11,

the conversion submodule is specifically used for determining a first word vector corresponding to each word in the word sequence according to a word list; converting a first word vector corresponding to each word into a second word vector through the word embedding model, wherein the dimension of the first word vector is larger than that of the second word vector; and generating the vector sequence by adopting the second word vector corresponding to each word.

13. The apparatus of claim 11 or 12, wherein the vector sequence generation module further comprises:

the recognition sub-module is used for recognizing the voice information to determine corresponding text information when the received information is the voice information; and after the text information is determined, triggering the word segmentation submodule to perform word segmentation on the text information.

14. The apparatus according to any one of claims 10 to 12, wherein the candidate information comprises: at least one candidate determined by the server by an encoder and a decoder in the training model;

the candidate information presentation module comprises:

a receiving submodule, configured to receive candidate information fed back by the server for the vector sequence;

15. The apparatus of claim 14, further comprising the following modules:

the information reply module is used for inputting the selected candidate item into a dialog box of the communication application so as to reply the information;

16. An information processing apparatus characterized by comprising:

the candidate information determining module is used for determining candidate information corresponding to the vector sequence through a pre-configured training model, wherein the training model and the word embedding model are obtained through co-training;

17. The apparatus of claim 16, wherein the candidate information determining module comprises:

a decoding submodule, configured to input the coded vector sequence to a decoder in the training model, so as to obtain at least one candidate;

and the determining submodule is used for determining the candidate information based on the obtained candidate.

18. The apparatus according to claim 17, wherein the candidate is specifically used for the intelligent terminal to reply to the message in a dialog box of a communication application;

the communication application specifically comprises at least one of the following types: short message application, instant messaging application, mail application, and telecommunication application.

19. An intelligent terminal comprising a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by one or more processors, the one or more programs comprising instructions for:

and displaying according to the candidate information fed back by the server.

20. The intelligent terminal according to claim 19, wherein the converting the received information through the pre-configured word embedding model to generate the corresponding vector sequence comprises:

21. The intelligent terminal of claim 20, wherein converting the sequence of words into the sequence of vectors through the word embedding model comprises:

22. The intelligent terminal of claim 20 or 21, wherein the converting the received information to generate a corresponding sequence of vectors further comprises instructions for:

23. The intelligent terminal according to any of claims 19 to 21, wherein the candidate information comprises: the server determines at least one candidate by an encoder and a decoder in the training model,

receiving candidate information fed back by the server for the vector sequence;

24. The intelligent terminal of claim 23, further comprising instructions for:

determining a selected candidate item according to the operation instruction;

25. A server comprising a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured for execution by one or more processors the one or more programs including instructions for:

26. The server according to claim 25, wherein the determining candidate information corresponding to the vector sequence through a pre-configured training model comprises:

and determining the candidate information based on the obtained candidate.

27. The server according to claim 26, wherein the candidate is used for the intelligent terminal to reply to the message in a dialog box of a messaging application;

28. An information processing system, the system comprising: a server and an intelligent terminal;

wherein the server comprises the server of claim 25;

the intelligent terminal comprises the intelligent terminal according to claim 19.

29. A readable storage medium, characterized in that instructions in the storage medium, when executed by a processor of a smart terminal, enable the smart terminal to perform the information processing method according to any one of method claims 1-6.

30. A readable storage medium, characterized in that instructions in the storage medium, when executed by a processor of a server, enable the server to perform the information processing method according to any one of method claims 7-9.