CN110929014B - Information processing method, information processing device, electronic equipment and storage medium - Google Patents
Information processing method, information processing device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN110929014B CN110929014B CN201911249755.4A CN201911249755A CN110929014B CN 110929014 B CN110929014 B CN 110929014B CN 201911249755 A CN201911249755 A CN 201911249755A CN 110929014 B CN110929014 B CN 110929014B
- Authority
- CN
- China
- Prior art keywords
- information
- intention
- input
- background
- matched
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation or dialogue systems
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Human Computer Interaction (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- User Interface Of Digital Computer (AREA)
- Machine Translation (AREA)
Abstract
The application discloses an information processing method, an information processing device, electronic equipment and a storage medium, wherein through receiving input information of a user, intention analysis is carried out on the input information to obtain intention information, background information is acquired when the intention information meets specific conditions, and target feedback information matched with the input information is determined based on the background information. Therefore, the feedback information can be selected by acquiring the associated background information while the intention analysis is ambiguous or the intention information cannot be matched with the unique feedback information, and the problems of poor user experience effect and excessive occupation of processing resources caused by directly selecting the information can be avoided.
Description
Technical Field
The present invention relates to the field of information processing technologies, and in particular, to an information processing method, an apparatus, an electronic device, and a storage medium.
Background
With the increasing popularity of smart devices, a user may conduct a smart session with the smart device, e.g., the user may input a question to be consulted through the smart device, and then the smart device may push an answer to the user.
However, when the smart device confirms the feedback information from the input information of the user, the analyzed feedback information corresponding to the intention of the user may not be unique. At this time, if the intelligent device directly outputs the feedback information, the user needs to select to determine the unique feedback information, so that the intellectualization of the intelligent device cannot be highlighted, and the experience effect of the user is reduced. On the other hand, if the smart device determines a unique answer in this case, it is necessary to generate guidance information or analyze the user's intention again, which may cause excessive occupation of processing resources.
Disclosure of Invention
In view of this, the present application provides the following technical solutions:
an information processing method, comprising:
receiving input information of a user;
performing intention analysis on the input information to obtain intention information matched with the input information;
if the intention information meets a specific condition, acquiring background information, wherein the background information characterizes information associated with the intention information, and the specific condition characterizes that the intention information cannot be matched with unique feedback information corresponding to the input information;
and determining target feedback information matched with the input information based on the background information.
Optionally, the performing intent analysis on the input information to obtain intent information matched with the input information includes:
inputting the input information into a pre-constructed intention recognition model, and determining predicted intention information corresponding to the input information through the intention recognition model;
the intention recognition model has the capability of chemotactic intention information corresponding to the input information to the actual intention corresponding to the input information; the intention model is a model obtained by training each piece of obtained sample information as training input of the neural network, and the sample information is information matched with the input information.
Optionally, the feedback information of the specific condition characterization matched with the intention information includes at least two pieces of information, and the obtaining the background information includes:
and acquiring first input information of a user aiming at the feedback information, wherein the first input information comprises part of keywords of the feedback information, and the first input information is the background information.
Optionally, the determining, based on the background information, target feedback information matched with the input information includes:
and screening the feedback information based on the first input information to determine target feedback information.
Optionally, the acquiring the background information includes:
judging whether stored information matched with the user exists or not, and if so, analyzing the feedback information to obtain a target keyword;
and extracting information from the stored information according to the target keywords to obtain background information.
Optionally, the extracting information from the stored information according to the target keyword to obtain background information includes:
inputting the target keyword into a pre-constructed information identification model, and determining predicted background information corresponding to the target keyword through the information identification model;
the information identification model has the capability of enabling the background information corresponding to the target keyword to trend to the actual background information corresponding to the target keyword; the information identification model is a model obtained by training each piece of obtained sample information serving as training input of a neural network, and the sample information is keyword information.
Optionally, the determining, based on the background information, target feedback information matched with the input information includes:
and determining target feedback information matched with the input information based on the background information and the intention information.
An information processing apparatus comprising:
a receiving unit for receiving input information of a user;
the analysis unit is used for carrying out intention analysis on the input information to obtain intention information matched with the input information;
the acquisition unit is used for acquiring background information if the intention information meets a specific condition, wherein the background information characterizes information associated with the intention information, and the specific condition characterizes that the intention information cannot be matched with unique feedback information corresponding to the input information;
and the determining unit is used for determining target feedback information matched with the input information based on the background information.
An electronic device, comprising:
a memory for storing a program;
a processor, configured to execute the program, where the program is specifically configured to:
receiving input information of a user;
performing intention analysis on the input information to obtain intention information matched with the input information;
if the intention information meets a specific condition, acquiring background information, wherein the background information characterizes information associated with the intention information, and the specific condition characterizes that the intention information cannot be matched with unique feedback information corresponding to the input information;
and determining target feedback information matched with the input information based on the background information.
A storage medium storing computer program code which, when executed, implements the information processing method of any one of the above.
As can be seen from the above technical solution, the present application discloses an information processing method, an apparatus, an electronic device, and a storage medium, which perform intent analysis on input information by receiving the input information of a user, obtain intent information, acquire background information in the event that the intent information satisfies a specific condition, and determine target feedback information matched with the input information based on the background information. Therefore, the feedback information can be selected by acquiring the associated background information while the intention analysis is ambiguous or the intention information cannot be matched with the unique feedback information, and the problems of poor user experience effect and excessive occupation of processing resources caused by directly selecting the information can be avoided.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings that are needed in the description of the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only embodiments of the present application, and that other drawings can be obtained according to the provided drawings without inventive effort for a person skilled in the art.
Fig. 1 is a schematic flow chart of an information processing method according to an embodiment of the present application;
FIG. 2 is an interface diagram of a user interacting with a voice assistant according to the prior art;
FIG. 3 is an interface diagram of a user interacting with a voice assistant according to an embodiment of the present application;
FIG. 4 is a schematic diagram of another interactive interface according to an embodiment of the present application;
fig. 5 is a flowchart of a method for obtaining background information according to an embodiment of the present application;
fig. 6 is a flowchart of a method for obtaining answer information according to an embodiment of the present disclosure;
FIG. 7 is a space diagram of background information of a user according to an embodiment of the present application;
FIG. 8 is a spatial diagram of answers to intentions of users provided in embodiments of the present application;
fig. 9 is a schematic structural diagram of an information processing apparatus according to an embodiment of the present application.
Detailed Description
The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are only some, but not all, of the embodiments of the present application. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are within the scope of the present disclosure.
The information processing method described in the embodiments of the present application may be applied in the field of intelligent response, such as intelligent clients, voice assistants, intelligent robots, and other applications or entities. The method mainly aims at how to realize accurate feedback of information under the condition that the unique feedback information cannot be directly obtained in the prior art or a series of guide words are needed to be generated to obtain the unique feedback information, and can reduce excessive processing resources occupied in the process of screening a plurality of feedback information or intention of a processing system.
Referring to fig. 1, a flowchart of an information processing method provided in an embodiment of the present application is shown, where the method may include the following steps:
s101, receiving input information of a user.
Because the information processing method provided by the application is applied to the intelligent question-answering field, the input information of the user can be text information, audio information, image information and the like. The input information of the user may be received through a terminal used by the user, for example, when the user triggers a certain application of the terminal, the input information of the user may be received through the application. Or when the wake-up word input by the user to a certain function is detected, for example, after the user outputs the wake-up word of a certain voice assistant, the voice assistant wakes up the voice interaction function according to the wake-up word and receives input information sent by the user. At this time, the input information of the user is voice information of the user.
S102, carrying out intention analysis on the input information to obtain intention information matched with the input information.
S103, if the intention information meets the specific condition, acquiring background information.
After the input information of the user is obtained, the input information needs to be subjected to intention analysis to obtain intention information matched with the input information. The intention information represents the intention of the input information of the user, i.e. what feedback is desired to be obtained by the input information. If the input information of the user is "time", the corresponding intention information may be "acquire current time information"; the input information of the user is "next monday 15:30 see meeting in the area", and the corresponding intention information is "record memo information".
Accordingly, the input information of the user is different and the intention information matched with the input information is different. When the intention information of the user is acquired, the keyword of the input information can be acquired by extracting the characteristic information of the input information, and the corresponding intention information is acquired according to the matching of the keyword and the intention information. The neural network can also be utilized to train the information sample to obtain a corresponding model, input information is input into the trained model, and corresponding intention information is obtained through recognition and classification of the model. Specific procedures this application will be specifically described in the following examples.
After the intention information of the user is obtained in the existing scheme, feedback information matched with the intention information is directly output. In the embodiment of the present application, after the intention information is obtained, the process of generating the feedback information is not directly performed, but the intention information needs to be judged, and if the intention information meets a specific condition, the background information is obtained. The background information characterizes information associated with the intention information, and the specific condition characterizes the intention information and cannot be matched with unique feedback information corresponding to the input information.
The context information is information associated with the intention information, and since the intention information corresponds to input information of the user, that is, the intention information is information associated with the user, the context information represents relevant context information of the user, and may be identification information of the user or information associated with previous input information input by the user. For example, the context information may include information of the identity, region, age, gender, etc. of the user.
It should be noted that, in the embodiment of the present application, the background information is obtained only if the intention information satisfies a specific condition. Related resource calling can be reduced, and excessive processing resources are prevented from being occupied. That is, when the intention information cannot match the unique feedback information, the background information is acquired. It can be understood that if the intent analysis is performed according to the input information of the user, the obtained feedback information is not unique. If the information input by the user is a question, the background information is acquired when the corresponding feedback information is a plurality of answers. The background information may assist in obtaining a more accurate answer.
S104, determining target feedback information matched with the input information based on the background information.
After the background information is obtained, target feedback information is determined according to the background information and the intention information. Thus, the selection of a plurality of feedback information through the background information is realized, and the target feedback information is obtained.
Referring to FIG. 2, which shows an interface diagram of a user interacting with a voice assistant in the prior art, the content of the user's interaction with the voice assistant in FIG. 2 is as follows:
the user: your, i live in beijing.
A voice assistant: you are very happy to serve you.
The user: please give the address of me service center.
A voice assistant: please select what you need at the lower service center address.
Tokyo: * Zone x and way x
London: * Street number
Beijing: * First major street number of region
It can be found through the interaction process of the user and the voice assistant presented in fig. 2 that the interaction process cannot be intelligent, the service center address output by the voice assistant is not unique, the user is required to select, the user is not intelligently felt by the voice assistant, and even if the user provides information capable of determining to output an answer, the information is not utilized in the process, so that the user still needs to select the answer again. Therefore, the experience effect of the user is poor, and the output feedback information is not unique and occupies output resources.
Referring to fig. 3, an interface diagram of user interaction with a voice assistant according to an embodiment of the present application is shown, where in fig. 3, the user interaction process with the voice assistant is as follows:
the user: your, i live in beijing.
A voice assistant: you are very happy to serve you.
The user: please give the address of me service center.
A voice assistant: beijing: * First major street of region.
It can be seen that, when the input information of the user is processed in the embodiment of the present application, the background information input by the user, that is, the regional information "beijing" input by the user is utilized, then when the intention analysis is performed, the address of the service center that the user needs to obtain is obtained, and the address of the service center that obtains beijing is directly located, and then the address is output as the output answer. Therefore, on one hand, the answer output is unique and accurate, on the other hand, the answer selection by the user is avoided, and the experience effect of the user is improved.
The application discloses an information processing method, which is used for carrying out intention analysis on input information by receiving the input information of a user to obtain intention information, acquiring background information when the intention information meets specific conditions, and determining target feedback information matched with the input information based on the background information. Therefore, the feedback information can be selected by acquiring the associated background information while the intention analysis is ambiguous or the intention information cannot be matched with the unique feedback information, and the problems of poor user experience effect and excessive occupation of processing resources caused by directly selecting the information can be avoided.
The technical features of the above embodiments are described in detail below.
In another embodiment of the present application, the intention information corresponding to the user may be acquired by a method using an artificial neural network. That is, the process of acquiring the intention information includes:
s201, inputting the input information into a pre-constructed intention recognition model, and determining predicted intention information corresponding to the input information through the intention recognition model.
Wherein the intention recognition model has the capability of inputting the actual intention corresponding to the information in the intention information area corresponding to the input information; the intention model is a model obtained by training each piece of obtained sample information serving as training input of the neural network, and the sample information is information matched with the input information. I.e. the sample information may be the same application scenario as the input information of the user or the input information generated under the application environment, such as may be historical input information of the user.
In order to achieve accurate user intention analysis, the intention recognition model trained on the basis of big data is specifically adopted in the embodiment to analyze the intention information of the user from the input information of the user.
Specifically, in the model training stage, multiple pieces of historical input data need to be obtained first as a training sample set, and manual labeling is performed on each piece of historical input data in the training sample set, wherein user intention corresponding to each piece of historical input data needs to be accurately labeled during labeling, so that labeling information of each piece of historical input data may comprise zero or one/multiple pieces of user intention labeling information.
After the intention information of the history input data is marked, training of an intention recognition model can be performed based on each training sample marked with the intention information of the user, and when the model is trained, the intention recognition model can adopt a text classification model, such as an SVM (Support Vector Machine ), a CNN (Convolutional Neural Network, convolutional neural network) or an LSTM (Long Short-Term Memory network) and the like; information extraction models such as CRF (conditional random field algorithm ), LSTM+CRF, etc. can also be selected, or text classification models and information extraction models can be effectively fused.
On the basis of completing model training, the input information of the user can be used as the input information of the intention recognition model to be input into the model, and the intention information set of the user is output after the model is processed. For example, the user's input information is "i want to know where the nearest bus stop is," and the model may output the user intent information set { address of bus stop }.
In the embodiment of the present application, the background information is acquired only when the intention information satisfies a specific condition. The specific condition characterizes the intention information and cannot be matched with the unique feedback information corresponding to the input information. That is, when the input information of the user is "today day of week", the intention of the user is analyzed that the user wants to know that today is day of week, and the answer is unique, so that the corresponding answer, such as "thursday", is directly output without acquiring the background information to determine the answer. The specific condition characterizes the intention information and cannot be matched with the unique feedback information corresponding to the input information, and the method can comprise the steps of analyzing to obtain a plurality of intentions, namely, the plurality of intentions can be matched with the plurality of feedback information; analysis may also include obtaining an intent information, but the intent information may result in a plurality of feedback information.
If the feedback information of the feature condition representation matched with the intention information at least comprises two pieces of information, the obtaining the background information comprises the following steps:
s301, acquiring first input information of a user aiming at feedback information.
Wherein the first input information includes a part of the keywords of the feedback information, and the first input information is used as background information.
It should be noted that, part of the keywords of the feedback information may be obtained by directly extracting the feedback information, or may be obtained by performing semantic information on the feedback information. The background information may be acquired after the feedback information is output, or may be acquired when the feedback information is ready to be output. For example, if the feedback information is all characterized as information about a location, the first input information obtained is a location keyword. If the feedback information is obtained through voice analysis and all the pieces of feedback information contained in the feedback information are correlated with the age, the obtained first input information is an age keyword.
Correspondingly, the determining the target feedback information matched with the input information based on the background information comprises the following steps:
s302, screening the feedback information based on the first input information to determine target feedback information.
Because the first input information characterizes part of keywords of the feedback information, the first information can be utilized to screen a plurality of pieces of feedback information, and target feedback information with highest matching degree with the user background and the intention is obtained.
The first input information may also be used to determine an output mode of the target feedback information when determining the target feedback information. For example, determining the ordering of the feedback information using the first input information, ordering the target feedback information at the first position of the feedback information; or the target feedback information is displayed, other display information except the target feedback information is hidden and output, for example, feedback information which is beyond the target feedback information is output in a pull-down list mode.
Referring to fig. 4, a schematic diagram of another interactive interface according to an embodiment of the present application is shown.
An output mode in which feedback information is output first and then target feedback information is output is shown in fig. 4 (a); an output model for directly outputting feedback information is shown in fig. 4 (b).
In fig. 4 (a), the interaction process of the user with the voice assistant is:
the user: vaccination sites.
A voice assistant: 0-3 years old at epidemic prevention center No. 1 floor 2;
3-18 years old in floor 3 of epidemic prevention center No. 1;
floor 1 at center for epidemic prevention No. 2 over 18 years old.
The user: is 15 years old.
A voice assistant: * Epidemic prevention center No. 1 floor 3.
From the user interaction process with the voice assistant in fig. 4 (a), it is known that the user inputs the "vaccination place" and "15 years old" two times before and after each other, respectively. Although the user inputs a plurality of information in the process, the method is different from the prior art that the effective association of the two information can not be carried out, or the user needs to rely on the prompt information of the voice assistant to input again when inputting the information for the second time. In the embodiment of the application, the user can input information twice, and for the processor corresponding to the voice assistant, the user does not need to generate prompt or guide information to assist the user in inputting information again, so that the complexity of information processing and the waste of resources can be reduced.
The user interaction process with the voice assistant shown in fig. 4 (b) is:
the user: you are good, i last year 15.
A voice assistant: you are very happy to serve you.
The user: vaccination sites.
A voice assistant: floor 3 at center of epidemic prevention No. 1.
In fig. 4 (b), when a user interacts with the voice assistant, the processor corresponding to the voice assistant can record the input information of the user each time when processing the interaction process with the user, so that when the user inputs a question, the user can combine the background information input before to perform effective screening of the answer, and obtain the target feedback information.
In another embodiment of the present application, a method for obtaining background information is further provided, referring to fig. 5, which is a schematic flow chart of a method for obtaining background information according to an embodiment of the present application, where the method may include the following steps:
s401, judging whether stored information matched with a user exists or not, and if so, executing S402;
s402, analyzing the feedback information to obtain a target keyword;
s403, information extraction is carried out in the stored information according to the target keywords, and background information is obtained.
According to the method and the device for processing the user input information, daily input information of the user can be analyzed and stored, and basic information of some users is extracted by taking the users as centers to serve as background information of the users. Such context information may include the user's name, age, gender, birthday, native place, location, and terminal device information, etc. When the background information needs to be acquired, whether the stored information matched with the user exists in the existing database can be judged, and if so, the stored information still needs to be extracted or screened. Instead of directly applying all of the stored context information, this increases the amount of computation and wastes additional data resources. At this time, it is necessary to analyze feedback information matching with the intention information to obtain the target keyword. The target keyword may be obtained by summarizing each piece of feedback information, or may be part of the feedback information representing the semantics thereof. And then extracting the target keywords from the stored information to obtain background information. If the target keyword is age, information related to the age of the user is extracted from the stored information as background information.
Specifically, information extraction is performed in the stored information according to the target keyword, so as to obtain background information, which can be realized in the following manner:
the target keywords belong to a pre-constructed information identification model, and predicted background information corresponding to the target scissors is determined through the information identification model;
the information identification model has the capability of enabling background information corresponding to the target keywords to trend to actual background information corresponding to the target keywords; the information identification model is used for training the obtained sample information serving as training input of the neural network, and the obtained model is trained, and the sample information is keyword information.
Therefore, the background information acquisition process can be more intelligent, and the background information can be searched more quickly.
In the embodiment of the invention, when the feedback information corresponding to the intention information is not unique, the feedback information can not be directly output, the background information is acquired according to the feedback information, and then the target feedback information is determined according to the background information and the intention information. The output feedback information is unique and more accurate. For example, if the intention information of the user is "get bus stop" and the background information is "user current location information", the target feedback information "bus stop nearest to the user" will be obtained.
In another possible implementation, the input information of the user may be more specific using the background information and the input information, so that the input information corresponds to unique or accurate feedback information. For example, the user's input information is "cell phone repair shop address", and the user's corresponding background information is "residential city is a city D zone". The mobile phone maintenance store address in the A city D area can be obtained by inputting information and background information of a user, so that the feedback information is more accurate according to the accurate input information, namely, the mobile phone maintenance store address in the feedback information is positioned in the A city D area, and the range of the feedback information obtained according to the initial input information of the user is reduced.
The application will be described by taking an intelligent session scenario as an example, referring to fig. 6, which shows a flow chart of a method for obtaining answer information provided in an embodiment of the application, where the method includes the following steps:
s501, acquiring current input information of a user;
s502, extracting an intention space of a user according to current input information;
s503, extracting an answer space corresponding to the intention space of the user;
s504, extracting background information of a user;
s505, selecting answer information matched with the user background by utilizing the answer space and the background information.
Referring to fig. 7, a space diagram of background information of a user according to an embodiment of the present application is shown. I.e. to extract all background information present and before the user. The user context information extracted includes: name, age, gender, birthday, place of birth, place of departure, location, equipment, etc.
Referring to fig. 8, a spatial diagram of answers to intentions of a user provided in an embodiment of the present application is shown. All intentions of the input information of the user are analyzed to obtain an intent set composed of a plurality of intent information, reasonable intent information can be ordered in descending order according to weight, and the highest-order intent can be obtained according to requirements. Namely, when the input information of the user corresponds to a plurality of intentions, the input information is selected according to the weight of the intentions. And each intention corresponds to a plurality of answer information.
So that final answer information matching the user's background can be determined from the answer information and the background information. It should be noted that not all answers are related to the background information, and in this case, answer information is directly output.
For example, in a part of use cases of the intelligent customer service system for solving mobile phone products, the processing procedure is as follows:
example 1:
the user: HI, I come from China, I have a mobile phone A;
intelligent customer service: you are good and happy to serve you;
the user: and the service center.
The treatment process comprises the following steps:
extracting all background information { native-China; handset brand-a };
extracting the intention { service center } of the current input;
searching a corresponding answer space { American service center, chinese service center };
and obtaining a user background answer { Chinese service center-W city M street }, and then outputting the user background answer { Chinese service center-W city M street }, according to the answer space and the background information.
Example 2:
the user: HI, 80 years old in india;
intelligent customer service: you are good and happy to serve you;
the user: please give me agent center telephone number.
The processing flow is as follows:
extracting all background information of a user: { Address-India; age-80 };
extracting the intention of the current input: { telephone number of proxy center };
searching the corresponding answer space { the telephone number of the agency center in the United states }; telephone number of agency center in China; phone number of indian proxy center };
based on the answer space and the background information, the telephone number 234 x 909 of the proxy center in india of the user background answer is obtained and then output to the user answer.
There is also provided an information processing apparatus in an embodiment of the present application, referring to fig. 9, the apparatus includes:
a receiving unit 10 for receiving input information of a user;
an analysis unit 11, configured to perform intent analysis on the input information, so as to obtain intent information matched with the input information;
an obtaining unit 12, configured to obtain background information if the intention information meets a specific condition, where the background information characterizes information associated with the intention information, and the specific condition characterizes that the intention information cannot match unique feedback information corresponding to the input information;
a determining unit 13 for determining target feedback information matched with the input information based on the background information.
The application discloses an information processing device, which is characterized in that input information of a user is received through a receiving unit, an analysis unit is used for carrying out intention analysis on the input information to obtain intention information, an acquisition unit is used for acquiring background information when the intention information meets specific conditions, and a determination unit is used for determining target feedback information matched with the input information based on the background information. Therefore, the feedback information can be selected by acquiring the associated background information while the intention analysis is ambiguous or the intention information cannot be matched with the unique feedback information, and the problems of poor user experience effect and excessive occupation of processing resources caused by directly selecting the information can be avoided.
On the basis of the above embodiment, the analysis unit includes:
a model analysis subunit, configured to input the input information into a pre-constructed intent recognition model, and determine predicted intent information corresponding to the input information through the intent recognition model;
the intention recognition model has the capability of chemotactic intention information corresponding to the input information to the actual intention corresponding to the input information; the intention model is a model obtained by training each piece of obtained sample information as training input of the neural network, and the sample information is information matched with the input information.
On the basis of the above embodiment, the feedback information that the specific condition characterization matches with the intention information includes at least two pieces of information, and the obtaining unit is specifically configured to:
and acquiring first input information of a user aiming at the feedback information, wherein the first input information comprises part of keywords of the feedback information, and the first input information is the background information.
Correspondingly, the determining unit is specifically configured to:
and screening the feedback information based on the first input information to determine target feedback information.
Optionally, the acquiring unit includes:
the judging subunit is used for judging whether the storage information matched with the user exists or not, and if so, analyzing the feedback information to obtain a target keyword;
and the extraction subunit is used for extracting information from the stored information according to the target keywords to obtain background information.
Optionally, the extraction subunit is specifically configured to:
inputting the target keyword into a pre-constructed information identification model, and determining predicted background information corresponding to the target keyword through the information identification model;
the information identification model has the capability of enabling the background information corresponding to the target keyword to trend to the actual background information corresponding to the target keyword; the information identification model is a model obtained by training each piece of obtained sample information serving as training input of a neural network, and the sample information is keyword information.
On the basis of the above embodiments, the determining unit is specifically configured to:
and determining target feedback information matched with the input information based on the background information and the intention information.
The embodiment of the application also provides electronic equipment, which comprises:
a memory for storing a program;
a processor, configured to execute the program, where the program is specifically configured to:
receiving input information of a user;
performing intention analysis on the input information to obtain intention information matched with the input information;
if the intention information meets a specific condition, acquiring background information, wherein the background information characterizes information associated with the intention information, and the specific condition characterizes that the intention information cannot be matched with unique feedback information corresponding to the input information;
and determining target feedback information matched with the input information based on the background information.
Further, the program is for:
the performing intent analysis on the input information to obtain intent information matched with the input information includes:
inputting the input information into a pre-constructed intention recognition model, and determining predicted intention information corresponding to the input information through the intention recognition model;
the intention recognition model has the capability of chemotactic intention information corresponding to the input information to the actual intention corresponding to the input information; the intention model is a model obtained by training each piece of obtained sample information as training input of the neural network, and the sample information is information matched with the input information.
Further, the program is for:
the feedback information of the specific condition representation matched with the intention information at least comprises two pieces of information, and the background information acquisition comprises the following steps:
and acquiring first input information of a user aiming at the feedback information, wherein the first input information comprises part of keywords of the feedback information, and the first input information is the background information.
Further, the program is for:
the determining, based on the context information, target feedback information that matches the input information includes:
and screening the feedback information based on the first input information to determine target feedback information.
Optionally, the acquiring the background information includes:
judging whether stored information matched with the user exists or not, and if so, analyzing the feedback information to obtain a target keyword;
and extracting information from the stored information according to the target keywords to obtain background information.
Optionally, the extracting information from the stored information according to the target keyword to obtain background information includes:
inputting the target keyword into a pre-constructed information identification model, and determining predicted background information corresponding to the target keyword through the information identification model;
the information identification model has the capability of enabling the background information corresponding to the target keyword to trend to the actual background information corresponding to the target keyword; the information identification model is a model obtained by training each piece of obtained sample information serving as training input of a neural network, and the sample information is keyword information.
Optionally, the determining, based on the background information, target feedback information matched with the input information includes:
and determining target feedback information matched with the input information based on the background information and the intention information.
The embodiment of the application also provides a storage medium, wherein the storage medium stores computer program codes, and the computer program codes realize the information processing method according to any one of the above when being executed.
In this specification, each embodiment is mainly described in the specification as a difference from other embodiments, and the same similar parts between the embodiments are referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant points refer to the description of the method section.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (11)
1. An information processing method, comprising:
receiving input information of a user;
performing intention analysis on the input information to obtain intention information matched with the input information;
if the intention information meets a specific condition, acquiring background information, wherein the background information characterizes information associated with the intention information, and the specific condition characterizes that the intention information cannot be matched with unique feedback information corresponding to the input information;
determining target feedback information matched with the input information based on the background information and the intention information;
the feedback information of the specific condition representation matched with the intention information at least comprises two pieces of information, and the background information acquisition comprises the following steps:
and acquiring first input information of a user aiming at the feedback information, wherein the first input information comprises part of keywords of the feedback information, and the first input information is the background information.
2. The method of claim 1, wherein the performing intent analysis on the input information to obtain intent information matched with the input information comprises:
inputting the input information into a pre-constructed intention recognition model, and determining predicted intention information corresponding to the input information through the intention recognition model;
the intention recognition model has the capability of chemotactic intention information corresponding to the input information to the actual intention corresponding to the input information; the intention model is a model obtained by training each piece of obtained sample information as training input of the neural network, and the sample information is information matched with the input information.
3. The method of claim 1, the determining target feedback information that matches the input information based on the context information, comprising:
and screening the feedback information based on the first input information to determine target feedback information.
4. An information processing method, comprising:
receiving input information of a user;
performing intention analysis on the input information to obtain intention information matched with the input information;
if the intention information meets a specific condition, acquiring background information, wherein the background information characterizes information associated with the intention information, and the specific condition characterizes that the intention information cannot be matched with unique feedback information corresponding to the input information;
determining target feedback information matched with the input information based on the background information and the intention information;
the obtaining the background information comprises the following steps:
judging whether stored information matched with the user exists or not, and if so, analyzing the feedback information to obtain a target keyword;
and extracting information from the stored information according to the target keywords to obtain background information.
5. The method according to claim 4, wherein the extracting information from the stored information according to the target keyword, to obtain background information, includes:
inputting the target keyword into a pre-constructed information identification model, and determining predicted background information corresponding to the target keyword through the information identification model;
the information identification model has the capability of enabling the background information corresponding to the target keyword to trend to the actual background information corresponding to the target keyword; the information identification model is a model obtained by training each piece of obtained sample information serving as training input of a neural network, and the sample information is keyword information.
6. The method of claim 4, the determining target feedback information that matches the input information based on the context information, comprising:
and determining target feedback information matched with the input information based on the background information and the intention information.
7. An information processing apparatus comprising:
a receiving unit for receiving input information of a user;
the analysis unit is used for carrying out intention analysis on the input information to obtain intention information matched with the input information;
the acquisition unit is used for acquiring background information if the intention information meets a specific condition, wherein the background information characterizes information associated with the intention information, and the specific condition characterizes that the intention information cannot be matched with unique feedback information corresponding to the input information;
a determining unit configured to determine target feedback information that matches the input information based on the background information and the intention information;
the feedback information of the specific condition representation matched with the intention information at least comprises two pieces of information, and the background information acquisition comprises the following steps:
and acquiring first input information of a user aiming at the feedback information, wherein the first input information comprises part of keywords of the feedback information, and the first input information is the background information.
8. An information processing apparatus comprising:
a receiving unit for receiving input information of a user;
the analysis unit is used for carrying out intention analysis on the input information to obtain intention information matched with the input information;
the acquisition unit is used for acquiring background information if the intention information meets a specific condition, wherein the background information characterizes information associated with the intention information, and the specific condition characterizes that the intention information cannot be matched with unique feedback information corresponding to the input information;
a determining unit configured to determine target feedback information that matches the input information based on the background information and the intention information;
the obtaining the background information comprises the following steps:
judging whether stored information matched with the user exists or not, and if so, analyzing the feedback information to obtain a target keyword;
and extracting information from the stored information according to the target keywords to obtain background information.
9. An electronic device, comprising:
a memory for storing a program;
a processor, configured to execute the program, where the program is specifically configured to:
receiving input information of a user;
performing intention analysis on the input information to obtain intention information matched with the input information;
if the intention information meets a specific condition, acquiring background information, wherein the background information characterizes information associated with the intention information, and the specific condition characterizes that the intention information cannot be matched with unique feedback information corresponding to the input information;
determining target feedback information matched with the input information based on the background information and the intention information;
the feedback information of the specific condition representation matched with the intention information at least comprises two pieces of information, and the background information acquisition comprises the following steps:
and acquiring first input information of a user aiming at the feedback information, wherein the first input information comprises part of keywords of the feedback information, and the first input information is the background information.
10. An electronic device, comprising:
a memory for storing a program;
a processor, configured to execute the program, where the program is specifically configured to:
receiving input information of a user;
performing intention analysis on the input information to obtain intention information matched with the input information;
if the intention information meets a specific condition, acquiring background information, wherein the background information characterizes information associated with the intention information, and the specific condition characterizes that the intention information cannot be matched with unique feedback information corresponding to the input information;
determining target feedback information matched with the input information based on the background information and the intention information;
the obtaining the background information comprises the following steps:
judging whether stored information matched with the user exists or not, and if so, analyzing the feedback information to obtain a target keyword;
and extracting information from the stored information according to the target keywords to obtain background information.
11. A storage medium storing computer program code which, when executed, implements the information processing method of any one of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911249755.4A CN110929014B (en) | 2019-12-09 | 2019-12-09 | Information processing method, information processing device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911249755.4A CN110929014B (en) | 2019-12-09 | 2019-12-09 | Information processing method, information processing device, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110929014A CN110929014A (en) | 2020-03-27 |
| CN110929014B true CN110929014B (en) | 2023-05-23 |
Family
ID=69857617
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911249755.4A Active CN110929014B (en) | 2019-12-09 | 2019-12-09 | Information processing method, information processing device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110929014B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111930904A (en) * | 2020-07-08 | 2020-11-13 | 联想(北京)有限公司 | Information response method, device, equipment and storage medium |
| CN112131364B (en) * | 2020-09-22 | 2024-03-26 | 沈阳东软智能医疗科技研究院有限公司 | Question answering method and device, electronic equipment and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104951456A (en) * | 2014-03-26 | 2015-09-30 | 上海智臻网络科技有限公司 | Method, device and equipment used for obtaining answer information |
| CN106383875A (en) * | 2016-09-09 | 2017-02-08 | 北京百度网讯科技有限公司 | Artificial intelligence-based man-machine interaction method and device |
| CN109063100A (en) * | 2018-07-27 | 2018-12-21 | 联想(北京)有限公司 | A kind of data processing method, server and electronic equipment |
| CN109697230A (en) * | 2018-12-26 | 2019-04-30 | 联想(北京)有限公司 | Information processing method and electronic equipment |
| WO2019084810A1 (en) * | 2017-10-31 | 2019-05-09 | 腾讯科技（深圳）有限公司 | Information processing method and terminal, and computer storage medium |
| CN110334201A (en) * | 2019-07-18 | 2019-10-15 | 中国工商银行股份有限公司 | A kind of intension recognizing method, apparatus and system |
-
2019
- 2019-12-09 CN CN201911249755.4A patent/CN110929014B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104951456A (en) * | 2014-03-26 | 2015-09-30 | 上海智臻网络科技有限公司 | Method, device and equipment used for obtaining answer information |
| CN106383875A (en) * | 2016-09-09 | 2017-02-08 | 北京百度网讯科技有限公司 | Artificial intelligence-based man-machine interaction method and device |
| WO2019084810A1 (en) * | 2017-10-31 | 2019-05-09 | 腾讯科技（深圳）有限公司 | Information processing method and terminal, and computer storage medium |
| CN109063100A (en) * | 2018-07-27 | 2018-12-21 | 联想(北京)有限公司 | A kind of data processing method, server and electronic equipment |
| CN109697230A (en) * | 2018-12-26 | 2019-04-30 | 联想(北京)有限公司 | Information processing method and electronic equipment |
| CN110334201A (en) * | 2019-07-18 | 2019-10-15 | 中国工商银行股份有限公司 | A kind of intension recognizing method, apparatus and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110929014A (en) | 2020-03-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107832286B (en) | Intelligent interaction method, equipment and storage medium | |
| CN107818781B (en) | Intelligent interaction method, equipment and storage medium | |
| JP6440732B2 (en) | Automatic task classification based on machine learning | |
| CN110020424B (en) | Contract information extraction method and device and text information extraction method | |
| KR102288249B1 (en) | Information processing method, terminal, and computer storage medium | |
| CN109284399B (en) | Similarity prediction model training method and device and computer readable storage medium | |
| CN111428010B (en) | Man-machine intelligent question-answering method and device | |
| CN110597952A (en) | Information processing method, server, and computer storage medium | |
| CN109145213B (en) | Historical information based query recommendation method and device | |
| KR20190016653A (en) | System and method for providing intelligent counselling service | |
| CN110187780B (en) | Long text prediction method, long text prediction device, long text prediction equipment and storage medium | |
| CN113360622B (en) | User dialogue information processing method and device and computer equipment | |
| CN109597874B (en) | Information recommendation method, device and server | |
| CN108268450B (en) | Method and apparatus for generating information | |
| CN110597965B (en) | Emotion polarity analysis method and device for article, electronic equipment and storage medium | |
| CN111310440A (en) | Text error correction method, device and system | |
| CN109522397B (en) | Information processing method and device | |
| CN110399473B (en) | Method and device for determining answers to user questions | |
| CN110929014B (en) | Information processing method, information processing device, electronic equipment and storage medium | |
| CN113806503A (en) | Dialog fusion method, device and equipment | |
| CN113901837A (en) | Intention understanding method, device, equipment and storage medium | |
| CN114065057A (en) | Search result determining method, display method, device, equipment and medium | |
| CN110110143B (en) | Video classification method and device | |
| CN111510563A (en) | Intelligent outbound method and device, storage medium and electronic equipment | |
| CN110750626B (en) | Scene-based task-driven multi-turn dialogue method 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 |