CN112347770A - Interactive processing method and device - Google Patents
Interactive processing method and device Download PDFInfo
- Publication number
- CN112347770A CN112347770A CN202011282527.XA CN202011282527A CN112347770A CN 112347770 A CN112347770 A CN 112347770A CN 202011282527 A CN202011282527 A CN 202011282527A CN 112347770 A CN112347770 A CN 112347770A
- Authority
- CN
- China
- Prior art keywords
- interaction
- information
- intention
- history information
- determining
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000002452 interceptive effect Effects 0.000 title claims abstract description 105
- 238000003672 processing method Methods 0.000 title claims abstract description 15
- 230000003993 interaction Effects 0.000 claims abstract description 360
- 230000004044 response Effects 0.000 claims abstract description 112
- 238000012545 processing Methods 0.000 claims abstract description 61
- 239000013598 vector Substances 0.000 claims description 53
- 238000000034 method Methods 0.000 claims description 41
- 238000012549 training Methods 0.000 claims description 14
- 230000015654 memory Effects 0.000 claims description 12
- 238000010586 diagram Methods 0.000 description 11
- 238000004590 computer program Methods 0.000 description 10
- 230000008569 process Effects 0.000 description 8
- 230000006870 function Effects 0.000 description 7
- 230000000694 effects Effects 0.000 description 6
- 239000000470 constituent Substances 0.000 description 5
- 238000004891 communication Methods 0.000 description 4
- 230000002349 favourable effect Effects 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000012935 Averaging Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 230000004931 aggregating effect Effects 0.000 description 1
- 230000002457 bidirectional effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000004806 packaging method and process Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 239000000758 substrate Substances 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
-
- 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
-
- 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
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3344—Query execution using natural language analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Databases & Information Systems (AREA)
- Mathematical Physics (AREA)
- Probability & Statistics with Applications (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- Human Computer Interaction (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- General Health & Medical Sciences (AREA)
- User Interface Of Digital Computer (AREA)
Abstract
The present disclosure provides an interactive processing method, which includes: acquiring interaction history information, wherein the interaction history information comprises input history information and/or response history information; determining an interaction intention of the interaction object indicated by the interaction history information by using a preset intention recognition model; and determining and outputting interaction response information for feeding back to the interaction object based on the interaction history information and the interaction intention. The disclosure also provides an interaction processing device, an electronic device and a computer readable storage medium.
Description
Technical Field
The present disclosure relates to the field of artificial intelligence technologies, and in particular, to an interaction processing method, an interaction processing apparatus, an electronic device, and a computer-readable storage medium.
Background
The rapid development of the artificial intelligence technology, the intelligent interaction can be widely permeated in different application fields, and the intelligent interaction can provide efficient interaction service for users, so that the service operation efficiency based on the intelligent interaction is favorably improved, and the labor cost of the platform for interacting with the users is favorably saved.
In the process of implementing the technical concept of the present disclosure, the inventor finds that in the intelligent interaction of the related art, the interactive response information for feedback to the user is determined based on the interactive history data interacted with the user, which has a problem that the determined interactive response information may not be matched with the interactive intention of the user, and not only affects the intelligent interaction efficiency, but also affects the intelligent interaction effect.
Disclosure of Invention
In view of this, the present disclosure provides an interaction processing method and apparatus with high interaction matching performance and effectively improved interaction efficiency.
One aspect of the present disclosure provides an interaction processing method, including: acquiring interaction history information, wherein the interaction history information comprises input history information and/or response history information; determining the interaction intention of the interaction object indicated by the interaction history information by using a preset intention recognition model; and determining and outputting interaction response information for feeding back to the interaction object based on the interaction history information and the interaction intention.
Optionally, the determining interactive response information for feeding back to the interactive object based on the interactive history information and the interactive intention includes: determining at least one intention descriptor associated with the interaction history information according to a preset association relationship between the interaction intention and the intention descriptors, wherein the association relationship is obtained by training based on interaction sample data by using an intention recognition model; and determining the interactive response information based on the interactive history information and the at least one intention descriptor.
Optionally, the determining the interaction response information based on the interaction history information and the at least one intention descriptor includes: determining an interaction candidate information base associated with the interaction intention, wherein the interaction candidate information base comprises at least one input candidate information and at least one response candidate information; determining target input candidate information with the highest matching degree with the interaction history information in the interaction candidate information base based on the interaction history information and the at least one intention descriptor; and using the response candidate information associated with the target input candidate information as the interactive response information.
Optionally, the determining, in the interaction candidate information base, target input candidate information that has a highest degree of matching with the interaction history information based on the interaction history information and the at least one intention descriptor includes: aiming at any input candidate information, determining a first matching degree between the interaction history information and the input candidate information; determining a second degree of matching between the at least one intent descriptor and the input candidate information; determining a third matching degree for the input candidate information based on the first matching degree and the second matching degree; repeating the steps until a third matching degree aiming at each input candidate information is determined; the target input candidate information is determined based on at least a third matching degree for each of the input candidate information.
Optionally, the determining the target input candidate information based on at least a third matching degree for each of the input candidate information includes: calculating a similarity between the average word vector of the at least one intention descriptor and the average word vector of the input candidate information as a fourth matching degree for the input candidate information, for any one of the input candidate information; calculating the number of the intention descriptors included in the input candidate information as a fifth matching degree for the input candidate information; calculating an average similarity between all the constituent words in the input candidate information and the at least one intention descriptor as a sixth matching degree for the input candidate information; and determining the target input candidate information based on the third matching degree, the fourth matching degree, the fifth matching degree and the sixth matching degree.
Optionally, the determining the interaction response information based on the interaction history information and the at least one intention descriptor includes: generating a model by using preset interaction information, and performing coding operation based on the interaction history information to obtain an intermediate feature vector; and performing a decoding operation based on the intermediate feature vector and the at least one intention descriptor to generate the interactive response information.
Optionally, in the training process of the interaction information generation model, the method includes: performing coding operation based on interactive sample information by using the interactive information generation model to obtain a sample intermediate feature vector, wherein the interactive sample information comprises sample input information and sample response information, and the interactive sample information has a sample label indicating an interactive intention; performing decoding operation based on the sample intermediate characteristic vector to obtain prediction response information; calculating a generation loss of the interaction information generation model based on the predicted response information and the sample response information; calculating a classification loss of the interaction information generation model based on the sample label and the interaction intention indicated by the prediction response information; and determining the total loss of the mutual information generation model based on the generation loss and the classification loss so as to perform model training based on the total loss.
Another aspect of the present disclosure provides an interaction processing apparatus, including: the acquisition module is used for acquiring interaction history information, wherein the interaction history information comprises input history information and/or response history information; the first processing module is used for determining the interaction intention of the interaction object indicated by the interaction history information by using a preset intention recognition model; and the second processing module is used for determining and outputting interaction response information fed back to the interaction object based on the interaction history information and the interaction intention.
Optionally, the second processing module includes: the first processing submodule is used for determining at least one intention descriptor associated with the interaction history information according to a preset association relationship between the interaction intention and the intention descriptor, wherein the association relationship is obtained by training based on interaction sample data by using an intention recognition model; and the second processing submodule is used for determining the interactive response information based on the interactive history information and the at least one intention descriptor.
Optionally, the second processing sub-module includes: a first processing unit, configured to determine an interaction candidate information base associated with the interaction intention, where the interaction candidate information base includes at least one input candidate information and at least one response candidate information; a second processing unit, configured to determine, in the interaction candidate information base, target input candidate information that is most highly matched with the interaction history information based on the interaction history information and the at least one intention descriptor; and a third processing unit, configured to use response candidate information associated with the target input candidate information as the interaction response information.
Optionally, the second processing unit includes: the first processing subunit is used for determining a first matching degree between the interaction history information and any input candidate information; a second processing subunit, configured to determine a second matching degree between the at least one intention descriptor and the input candidate information; a third processing subunit configured to determine a third matching degree for the input candidate information based on the first matching degree and the second matching degree; repeating the steps until a third matching degree aiming at each input candidate information is determined; a fourth processing subunit, configured to determine the target input candidate information based on at least a third matching degree for each of the input candidate information.
Optionally, the fourth processing subunit is specifically configured to: calculating a similarity between the average word vector of the at least one intention descriptor and the average word vector of the input candidate information as a fourth matching degree for the input candidate information, for any one of the input candidate information; calculating the number of the intention descriptors included in the input candidate information as a fifth matching degree for the input candidate information; calculating an average similarity between all the constituent words in the input candidate information and the at least one intention descriptor as a sixth matching degree for the input candidate information; and determining the target input candidate information based on the third matching degree, the fourth matching degree, the fifth matching degree and the sixth matching degree.
Optionally, the second processing sub-module includes: the fourth processing unit is used for generating a model by utilizing preset interaction information, and performing coding operation based on the interaction history information to obtain an intermediate feature vector; a fifth processing unit, configured to perform a decoding operation based on the intermediate feature vector and the at least one intention descriptor to generate the interactive response information.
Optionally, in the training process of the interaction information generation model, the method includes: performing coding operation based on interactive sample information by using the interactive information generation model to obtain a sample intermediate feature vector, wherein the interactive sample information comprises sample input information and sample response information, and the interactive sample information has a sample label indicating an interactive intention; performing decoding operation based on the sample intermediate characteristic vector to obtain prediction response information; calculating a generation loss of the interaction information generation model based on the predicted response information and the sample response information; calculating a classification loss of the interaction information generation model based on the sample label and the interaction intention indicated by the prediction response information; and determining the total loss of the mutual information generation model based on the generation loss and the classification loss so as to perform model training based on the total loss.
Another aspect of the present disclosure provides an electronic device. The electronic device includes at least one processor and a memory communicatively coupled to the at least one processor. The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to implement the method of the embodiment of the disclosure.
Another aspect of the present disclosure provides a computer-readable storage medium storing computer-executable instructions that, when executed, implement the method of embodiments of the present disclosure.
Another aspect of the present disclosure provides a computer program comprising computer executable instructions that when executed perform the method of embodiments of the present disclosure.
According to the embodiment of the disclosure, the interactive history information is acquired, and the interactive history information comprises the input history information and/or the response history information; determining the interaction intention of the interaction object indicated by the interaction history information by using a preset intention recognition model; the technical scheme for determining and outputting the interactive response information fed back to the interactive object based on the interactive historical information and the interactive intention at least partially overcomes the technical problems of poor interactive matching and poor interactive efficiency in the related technology, and further achieves the technical effect of effectively improving the interactive matching and the interactive efficiency.
Drawings
For a more complete understanding of the present disclosure and the advantages thereof, reference is now made to the following descriptions taken in conjunction with the accompanying drawings, in which:
FIG. 1 schematically illustrates an interaction processing system architecture according to an embodiment of the present disclosure;
FIG. 2 schematically shows a flow diagram of a method of interaction processing according to an embodiment of the present disclosure;
FIG. 3 schematically shows a flow chart of another interaction processing method according to an embodiment of the present disclosure;
FIG. 4 schematically illustrates a diagram of a method of determining a third degree of match for a certain candidate input information according to an embodiment of the present disclosure;
FIG. 5 schematically illustrates a flow chart of yet another interaction processing method according to an embodiment of the present disclosure;
FIG. 6 schematically shows a block diagram of an interaction processing device according to an embodiment of the present disclosure;
fig. 7 schematically illustrates a block diagram of an electronic device suitable for implementing the interaction processing method and apparatus according to an embodiment of the present disclosure.
Detailed Description
Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. It is to be understood that such description is merely illustrative and not intended to limit the scope of the present disclosure. In the following detailed description, for purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of the embodiments of the disclosure. It may be evident, however, that one or more embodiments may be practiced without these specific details. Moreover, in the following description, descriptions of well-known structures and techniques are omitted so as to not unnecessarily obscure the concepts of the present disclosure.
The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. The terms "comprises," "comprising," and the like, as used herein, specify the presence of stated features, steps, operations, and/or components, but do not preclude the presence or addition of one or more other features, operations steps, operations, or components.
All terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art unless otherwise defined. It is noted that the terms used herein should be interpreted as having a meaning that is consistent with the context of this specification and should not be interpreted in an idealized or overly formal sense.
Where a convention analogous to "at least one of A, B and C, etc." is used, in general such a construction is intended in the sense one having skill in the art would understand the convention (e.g., "a system having at least one of A, B and C" would include but not be limited to systems that have a alone, B alone, C alone, a and B together, a and C together, B and C together, and/or A, B, C together, etc.).
Various embodiments of the present disclosure provide an interaction processing method and an interaction processing apparatus to which the method can be applied. The method comprises the steps of obtaining interaction history information, wherein the interaction history information comprises input history information and/or response history information, determining an interaction intention of an interaction object indicated by the interaction history information by using a preset intention recognition model, and then determining interaction response information for feeding back the interaction object based on the interaction history information and the interaction intention and outputting the interaction response information.
As shown in fig. 1, the system architecture 100 includes at least one terminal (a plurality of which are shown, e.g., terminals 101, 102, 103) and a server 104. In the system architecture 100, a server 104 obtains interaction history information generated by interacting with an interaction object, wherein the interaction history information includes input history information and/or response history information, then determines an interaction intention of the interaction object indicated by the interaction history information by using a preset intention recognition model, and finally determines interaction response information for feeding back to the interaction object and sends the interaction response information to terminals (such as the terminals 101, 102, 103) based on the interaction history information and the interaction intention.
The present disclosure will be described in detail below with reference to the drawings and specific embodiments.
Fig. 2 schematically shows a flow chart of an interaction processing method according to an embodiment of the present disclosure.
As shown in fig. 2, the method may include operations S210 to S230, for example.
In operation S210, interaction history information including input history information and/or response history information is acquired.
In the embodiment of the present disclosure, specifically, the user interacts with the service platform through the terminal to generate the interaction data, for example, the user interacts with an intelligent assistant provided by the service platform through the terminal to generate the interaction data. The terminal may include, for example, a smart phone, a tablet, a laptop, a desktop, a smart watch, a smart speaker, a smart home, and so on. The business platforms may include, for example, shopping platforms, takeaway platforms, taxi platforms, machine conversation platforms, and the like. The intelligent assistants provided by the service platform may include, for example, virtual customer service, intelligent assistants, conversation robots, and the like.
The server for interactive processing acquires interactive history information between the user and the service platform, wherein the interactive history information comprises input history information from the user and response history information from the service platform. The server may be an independent physical server, or may be a server cluster or a distributed system formed by a plurality of physical servers. The form of the interactive history information can be text form information, voice form information, and other interactive form information.
When the interaction history information between the user and the service platform is obtained, the interaction history information in a preset time period can be obtained by using the interaction input information to be responded of the user as a starting point according to the time stamp sequence, or a preset number of interaction history information can be obtained. The interactive history information comprises input history information from the user and response history information from the service platform, and the interactive response information for feedback to the user is determined based on the interactive history information, so that the matching degree of the determined interactive response information is improved, the intelligent interactive experience is improved, and the intelligent interactive efficiency is improved.
Optionally, before determining the interaction response information for feedback to the user based on the interaction history information, the acquired interaction history information may be preprocessed in the forms of deduplication, filtering, text replacement, text conversion, and the like, so as to convert the non-standard information into the standard information. Illustratively, information that does not include text content in the session text information is filtered out, for example, session information that includes only punctuation marks or emoticons is filtered out. By preprocessing the acquired interaction history information, on one hand, the calculation amount of subsequent data processing is favorably reduced, and on the other hand, the accuracy of a data processing result is favorably improved.
Next, in operation S220, an interaction intention of the interaction object indicated by the interaction history information is determined using a preset intention recognition model.
In the embodiment of the disclosure, specifically, the acquired interaction history information is input into a trained intention recognition model, and the intention recognition model is used to perform recognition processing on the interaction history information to obtain the interaction intention of the interaction object indicated by the interaction history information, that is, the interaction intention of the user indicated by the interaction history information. Aiming at different application scenes and actual needs, the interaction intention has fine-grained division. For example, in an e-commerce application, for a payment scenario, the interactive intention may include payment means confirmation, preferential information consultation, postage inquiry, and the like; for an after-sales scenario, the interaction intent may include logistics information query, return consultation, merchandise question feedback, and the like.
And when the interaction intention of the interaction object indicated by the interaction history information is determined, encoding each interaction history information by using an intention identification model to obtain a feature vector associated with each interaction history information. The encoding method may include encoding based on a transform model (translation model), or encoding based on a Bi-GRU (Bidirectional Gated recursive Unit) model, for example.
Aiming at a large number of preset fine-grained divided interaction intents, each interaction intention has a corresponding intention label. And when the interaction intention indicated by the interaction history information is determined, vectorizing the preset intention label of the interaction intention to obtain the feature vectors of different intention labels. Then, determining the similarity between the feature vector of each interaction history information and the feature vector of the intention label, and determining a preset interaction intention with the similarity higher than a preset threshold value as the interaction intention indicated by the interaction history information. The similarity may include cosine similarity, inverse euclidean distance, and the like, and the specific type of the similarity is not limited in this embodiment.
The intention recognition model is trained by using a large amount of interaction sample data with a classification result indicating an interaction intention associated with the interaction sample data, wherein the intention recognition model may be, for example, an HAN (hierarchical Attention Network) model.
When a large amount of interaction sample data with classification results are used for model training, aiming at any interaction intention, an intention recognition model is used for marking n vocabularies with the highest attention scores in each interaction sample data, then the n vocabularies with the highest attention scores in each interaction sample data are collected, and the preset number of vocabularies with the highest occurrence frequency are determined as intention descriptors associated with the interaction intention. The preset number of words with the highest occurrence frequency form an intention description word bank associated with the interaction intention.
Illustratively, for the interaction intention of "express delivery information consultation", there are 20 ten thousand pieces of interaction sample data associated with the interaction intention. And labeling the 5 vocabularies with the highest attention scores in each interactive sample data by using an intention recognition model, then summarizing the 5 vocabularies with the highest attention scores in each interactive sample data in 20 ten thousand interactive sample data, and determining the 100 vocabularies with the highest occurrence frequency as the intention descriptors associated with the interactive intention. The 100 vocabularies with the highest frequency of occurrence form an intention description word bank associated with the interactive intention of 'express information consultation'.
Next, in operation S230, interaction response information for feedback to the interaction object is determined and output based on the interaction history information and the interaction intention.
In the embodiment of the present disclosure, specifically, at least one intention descriptor associated with the interaction history information is determined according to a preset association relationship between the interaction intention and the intention descriptor, where the association relationship is trained based on the interaction sample data by using an intention recognition model, and then the interaction response information is determined based on the interaction history information and the at least one intention descriptor.
After the interaction intention indicated by the interaction history information is determined by the intention recognition model, a preset number of intention descriptors are determined as the intention descriptors associated with the interaction history information according to an intention description word bank associated with the interaction intention. Illustratively, in the intention descriptor library associated with the interaction intention, 50 of the intention descriptors are determined as the intention descriptors associated with the interaction history information. The intention description lexicon associated with the interaction intention is obtained during the training process of the intention recognition model, and is specifically set forth in operation S220, and is not described herein again.
And determining and outputting interaction response information for feeding back to the interaction object based on the interaction history information and the interaction intention indicated by the interaction history information, in particular, based on the interaction history information and at least one intention descriptor associated with the interaction history information. The method is favorable for ensuring that the output interactive response information is matched with the interactive intention of the user, and is favorable for improving the self-service interactive effect and improving the self-service interactive efficiency.
In the embodiment of the disclosure, by acquiring interaction history information, wherein the interaction history information includes input history information and/or response history information, and using a preset intention recognition model, an interaction intention of an interaction object indicated by the interaction history information is determined, and then interaction response information for feeding back to the interaction object is determined and output based on the interaction history information and the interaction intention. Determining interactive response information for feeding back to an interactive object based on the acquired interactive history information and the interactive intention indicated by the interactive history information, which is favorable for improving the matching degree of the output interactive response information on the interactive intention level; through accurate matching interaction input information and providing interaction response information with high accuracy, not only can intelligent interaction efficiency be effectively improved, but also the intelligent interaction effect can be favorably improved.
Fig. 3 schematically shows a flow chart of another interaction processing method according to an embodiment of the present disclosure.
As shown in fig. 3, operation S230 may specifically include operations S310 to S330.
In operation S310, an interaction candidate information base associated with the interaction intention is determined, the interaction candidate information base including at least one input candidate information and at least one response candidate information.
In the embodiment of the present disclosure, specifically, after the interaction intention indicated by the interaction history information is determined, an interaction candidate information base associated with the interaction intention is determined, where the interaction candidate information base includes at least one input candidate information and at least one response candidate information. The interaction candidate information base may be, for example, a standard question-answer base including at least one standard question text and at least one standard answer text, and the standard question text may include, for example, "how much the freight of asking xx express is".
In the above process, for each interaction intention, the interaction candidate information associated with each interaction intention may be stored in a key manner. Specifically, the intention labels of the interaction intentions are used as key names, and the interaction candidate information is used as key values to store the interaction candidate information associated with each interaction intention. And forming an interaction candidate information base associated with the interaction intention by the plurality of interaction candidate information associated with the interaction intention.
Next, in operation S320, target input candidate information that has the highest degree of matching with the interaction history information is determined in the interaction candidate information base based on the interaction history information and the at least one intention descriptor.
In the embodiment of the disclosure, specifically, for any input candidate information, a first matching degree between the interaction history information and the input candidate information is determined; determining a second matching degree between the at least one intention descriptor and the input candidate information; determining a third matching degree for the input candidate information based on the first matching degree and the second matching degree; repeating the steps until a third matching degree aiming at each input candidate information is determined; and determining target input candidate information at least based on the third matching degree for each input candidate information.
The interaction history information includes at least one input history information and/or at least one response history information. For any input candidate information in the interaction candidate information base, when a first matching degree between the interaction history information and the input candidate information is determined, the first matching degree between each piece of interaction history information and the input candidate information is determined. Illustratively, the interaction history information includes 3 history topic texts of a1, a2 and A3 from the user and 2 history question texts of Q1 and Q2 from the virtual customer service, and for the input candidate information "how much the freight of asking xx courier is", first matching degrees between a1, a2, A3, Q1 and Q2 and the input candidate information are respectively determined. The method for determining the first matching degree may include, for example, determining a similarity between a feature vector of the interaction history information and a feature vector of the input candidate information.
In addition, a second matching degree between at least one intention descriptor associated with the interaction history information and each input candidate information in the interaction candidate information base is determined. For any input candidate information, at least one intention descriptor can be used as additional interactive input information, and a feature vector of the interactive input information formed by the at least one intention descriptor is determined. Then, the similarity between the feature vector of the "additional" interactive input information and the feature vector of the input candidate information is determined to obtain a second matching degree between the at least one intention descriptor and the input candidate information.
And for any input candidate information, after determining a first matching degree between each piece of interaction history information and the input candidate information and determining a second matching degree between at least one intention descriptor and the input candidate information, aggregating the first matching degree and the second matching degree by using a matching matrix to obtain a third matching degree for the input candidate information. The above-described operations are repeated, a third matching degree for each input candidate information is determined, and then the input candidate information of which the third matching degree is highest is determined as the target input candidate information based on the third matching degree for each input candidate information.
Fig. 4 is a schematic diagram schematically illustrating a method for determining a third matching degree for a certain candidate input information according to an embodiment of the present disclosure, and as shown in fig. 4, the obtained interaction history information includes a1, Q1, a2, Q2, and A3. For any input candidate information, first matching degrees p1, p2, p3, p4 and p5 between A1, Q1, A2, Q2 and A3 and the input candidate information are respectively determined. Meanwhile, at least one intention descriptor is used as additional interactive input information, and a second matching degree p6 between the at least one descriptor and the input candidate information is determined based on the feature vector associated with the interactive input information. Then, a third degree of matching for the candidate input information is determined using the matching matrix based on the first degree of matching p1, p2, p3, p4, p5, and the second degree of matching p 6. After determining the third matching degree for each candidate input information, determining the candidate input information with the highest third matching degree as the target candidate input information.
In addition to determining the target input candidate information based on the third matching degree for each input candidate information, it is also possible to integrate a plurality of matching degrees by an integration method and then determine the target input candidate information based on the integrated matching degree. Specifically, for any input candidate information, calculating the similarity between the average word vector of at least one intention descriptor and the average word vector of the input candidate information as a fourth matching degree for the input candidate information; calculating the number of the intention descriptors contained in the input candidate information as a fifth matching degree for the input candidate information; calculating the average similarity between all the component words in the input candidate information and at least one intention description word to serve as a sixth matching degree for the input candidate information; and determining target input candidate information based on the third matching degree, the fourth matching degree, the fifth matching degree and the sixth matching degree.
The average similarity between all the constituent words in the input candidate information and at least one of the intention descriptor is the average similarity of all the constituent words, and is obtained by averaging the similarities of all the constituent words. When the similarity of a certain component word is calculated, the similarity between the component word and each intention description word is calculated, and the maximum similarity is determined as the similarity of the component word. After the fourth, fifth, and sixth matching degrees are obtained, the matching degrees including the third matching degree may be integrated by using an integration method such as Xgboost to obtain an integrated matching degree for each input candidate information, and then the input candidate information with the highest integrated matching degree is determined as the target input candidate information.
Next, in operation S330, response candidate information associated with the target input candidate information is output as interaction response information.
In the embodiment of the present disclosure, specifically, after determining the target input candidate information having the highest matching degree with the historical interaction information and the at least one intention keyword, the response candidate information associated with the target input candidate information is used as the interaction response information for feeding back to the interaction object and is output.
And by combining and determining the second matching degree between at least one intention descriptor and each input candidate information, the target input candidate information with the highest matching degree is searched in the interaction candidate information base. The matching degree between the response candidate information associated with the target input candidate information and the interaction history information and the interaction intention indicated by the interaction history information is highest, so that the matching accuracy of the output interaction response information and the interaction intention of the interaction object is effectively improved, and the intelligent interaction efficiency and the intelligent interaction effect between the output interaction response information and the interaction object are improved.
FIG. 5 schematically shows a flow chart of yet another interaction processing method according to an embodiment of the present disclosure.
As shown in fig. 5, operation S230 may specifically include operations S510 to S520.
In operation S510, a preset interaction information generation model is used to perform an encoding operation based on interaction history information, so as to obtain an intermediate feature vector.
In this embodiment of the present disclosure, specifically, the preset interaction information generation model may be, for example, a sequence-to-sequence model, and the encoding operation based on the interaction history information by using the interaction information generation model may specifically include: first, traversing each character in the interaction history information, and traversing each character in the input history information and each character in the response history information when the interaction history information comprises a plurality of input history information and a plurality of response history information. And sequentially taking each character in the interactive history information as the input of the corresponding coding time. In addition, the input of each coding moment also comprises the hidden state output by the adjacent last coding moment. Similarly, the present coding time also outputs a hidden state, and the hidden state output by the present coding time is also used as the input of the next adjacent coding time. For the seq2seq model, the hidden state output at the last encoding time is usually reserved as an intermediate feature vector obtained by the encoding operation based on the interaction history information, and the intermediate feature vector can be considered as semantic information encoding the interaction history information.
Next, in operation S520, a decoding operation based on the intermediate feature vector and the at least one intention descriptor is performed to generate and output interactive response information.
In the embodiment of the present disclosure, specifically, the intermediate feature vector is used as the initial hidden state at the first decoding time, that is, the hidden state output at the last encoding time is used as the initial hidden state at the first decoding time. And outputting the character and the hidden state at each decoding moment of the neural network model, and inputting the character and the hidden state output at the previous decoding moment to the next adjacent decoding moment by taking the character and the hidden state output at the previous decoding moment as input. In addition, in order to improve the matching degree between the generated interactive response information and the interactive intention of the interactive object, an average word vector of at least one intention descriptor is determined, and each decoding time is input by using the average word vector as a representative vector of the at least one intention descriptor. And repeating the operation until the EOS token is decoded and output, wherein the characters output at each decoding moment form interactive response information.
Optionally, an attention mechanism may be introduced to solve the problem that the input sentence is long and has a poor encoding effect, specifically, when performing decoding calculation at time t, in addition to taking the character and hidden state output at the previous decoding time and taking the representation vector of at least one intention descriptor as input, the input at all times in the encoding calculation may be weighted to obtain a context vector, and then the context vector is simultaneously taken as input at each decoding time, so as to generate the interactive response information with a well-matched interactive intention.
The interactive information generation model is a pre-trained model, and can be used for performing coding operation based on interactive sample information in the training process of the interactive information generation model to obtain a sample intermediate characteristic vector, wherein the interactive sample information comprises sample input information and sample response information, and the interactive sample information has a sample label indicating an interactive intention; performing decoding operation based on the sample intermediate characteristic vector to obtain prediction response information; calculating the generation loss of the interactive information generation model based on the predicted response information and the sample response information; calculating a classification loss of the interaction information generation model based on the sample label and the interaction intention indicated by the predicted response information; and determining the total loss of the mutual information generation model based on the generation loss and the classification loss so as to train the model based on the total loss.
Specifically, when interactive response information is generated using the interactive information generation model, p (yt) represents the probability of generating a word at time t, Emb (yt) represents the word vector of the word, Ewe (t; C, z) represents the expected vector value of the generated word at time t:
where p (yt) is the probability of generating a word in the dictionary, which can be trained by the decoder of the mutual information generation model, and V represents a preset common dictionary. According to the sample response information and each timeThe expected vector value Ewe (t; C, z) of the generated word can determine the generation loss L of the mutual information generation modelCE。
Cumulatively averaging the expected vector values of the generated words at all times, and then using the softmax activation function to obtain q (z | Y) which represents the respective probabilities of the intent to generate the statement:
p (z) represents the intention of the expected generated sentence (which can be determined by the sample label of the interaction sample information), q (z | Y) represents the probability of the intention of the actually generated sentence, and the classification loss of the interaction information generation model can be expressed as:
LCLS＝-p(z)log q(z|Y)
based on the generation loss LCEAnd a classification loss LCLSDetermining the total loss L ═ L of the mutual information generation modelCE+μLCLSMu is a definable variable, generating a loss LCEAnd a classification loss LCLSWill act jointly on the generation of the interactive response information.
Fig. 6 schematically shows a block diagram of an interaction processing device according to an embodiment of the present disclosure.
As shown in fig. 6, the apparatus may include an acquisition module 601, a first processing module 602, and a second processing module 603.
Specifically, the obtaining module 601 is configured to obtain interaction history information, where the interaction history information includes input history information and/or response history information; a first processing module 602, configured to determine an interaction intention of an interaction object indicated by the interaction history information, using a preset intention recognition model; a second processing module 603, configured to determine and output interaction response information for feeding back to the interaction object based on the interaction history information and the interaction intention.
In the embodiment of the disclosure, by acquiring interaction history information, wherein the interaction history information includes input history information and/or response history information, and using a preset intention recognition model, an interaction intention of an interaction object indicated by the interaction history information is determined, and then interaction response information for feeding back to the interaction object is determined and output based on the interaction history information and the interaction intention. Determining interactive response information for feeding back to an interactive object based on the acquired interactive history information and the interactive intention indicated by the interactive history information, which is favorable for improving the matching degree of the output interactive response information on the interactive intention level; through accurate matching interaction input information and providing interaction response information with high accuracy, not only can intelligent interaction efficiency be effectively improved, but also the intelligent interaction effect can be favorably improved.
As an alternative embodiment, the second processing module comprises: the first processing submodule is used for determining at least one intention descriptor associated with the interaction history information according to a preset association relationship between the interaction intention and the intention descriptor, wherein the association relationship is obtained by training based on interaction sample data by using an intention recognition model; and the second processing submodule is used for determining interactive response information based on the interactive history information and the at least one intention descriptor.
As an alternative embodiment, the second processing submodule includes: the system comprises a first processing unit, a second processing unit and a third processing unit, wherein the first processing unit is used for determining an interaction candidate information base associated with an interaction intention, and the interaction candidate information base comprises at least one input candidate information and at least one response candidate information; the second processing unit is used for determining target input candidate information with the highest matching degree with the interaction history information in the interaction candidate information base on the basis of the interaction history information and at least one intention descriptor; and the third processing unit is used for taking the response candidate information associated with the target input candidate information as the interaction response information.
As an alternative embodiment, the second processing unit comprises: the first processing subunit is used for determining a first matching degree between the interaction history information and the input candidate information aiming at any input candidate information; a second processing subunit, configured to determine a second degree of matching between the at least one intention descriptor and the input candidate information; a third processing subunit, configured to determine a third matching degree for the input candidate information based on the first matching degree and the second matching degree; repeating the steps until a third matching degree aiming at each input candidate information is determined; a fourth processing subunit, configured to determine the target input candidate information based on at least the third matching degree for each input candidate information.
As an optional embodiment, the fourth processing subunit is specifically configured to: calculating the similarity between the average word vector of at least one intention descriptor and the average word vector of the input candidate information aiming at any input candidate information to serve as a fourth matching degree aiming at the input candidate information; calculating the number of the intention descriptors contained in the input candidate information as a fifth matching degree for the input candidate information; calculating the average similarity between all the component words in the input candidate information and at least one intention description word to serve as a sixth matching degree for the input candidate information; and determining target input candidate information based on the third matching degree, the fourth matching degree, the fifth matching degree and the sixth matching degree.
As an alternative embodiment, the second processing submodule includes: the fourth processing unit is used for generating a model by using preset interaction information, and performing coding operation based on interaction history information to obtain an intermediate feature vector; and the fifth processing unit is used for decoding operation based on the intermediate characteristic vector and at least one intention descriptor so as to generate interactive response information.
As an alternative embodiment, in the training process of the interaction information generation model, the method includes: utilizing an interaction information generation model to perform coding operation based on interaction sample information to obtain a sample intermediate feature vector, wherein the interaction sample information comprises sample input information and sample response information, and the interaction sample information has a sample label indicating an interaction intention; performing decoding operation based on the sample intermediate characteristic vector to obtain prediction response information; calculating the generation loss of the interactive information generation model based on the predicted response information and the sample response information; calculating a classification loss of the interaction information generation model based on the sample label and the interaction intention indicated by the predicted response information; and determining the total loss of the mutual information generation model based on the generation loss and the classification loss so as to train the model based on the total loss.
Alternatively, at least part of the functions of any of the modules, sub-modules, or any of the modules in the obtaining module 601, the first processing module 602, and the second processing module 603 may be implemented in one module. Any one or more of the modules according to the embodiments of the present disclosure may be implemented by being split into a plurality of modules. Any one or more of the modules according to the embodiments of the present disclosure may be implemented at least in part as a hardware circuit, such as a Field Programmable Gate Array (FPGA), a Programmable Logic Array (PLA), a system on a chip, a system on a substrate, a system on a package, an Application Specific Integrated Circuit (ASIC), or may be implemented in any other reasonable manner of hardware or firmware by integrating or packaging the circuit, or in any one of three implementations, or in any suitable combination of any of the software, hardware, and firmware. Or one or more of the modules according to embodiments of the disclosure, may be implemented at least partly as computer program modules which, when executed, may perform corresponding functions.
Fig. 7 schematically illustrates a block diagram of an electronic device suitable for implementing the interaction processing method and apparatus according to an embodiment of the present disclosure. The computer system illustrated in FIG. 7 is only one example and should not impose any limitations on the scope of use or functionality of embodiments of the disclosure.
As shown in fig. 7, a computer system 700 according to an embodiment of the present disclosure includes a processor 701, which can perform various appropriate actions and processes according to a program stored in a Read Only Memory (ROM)702 or a program loaded from a storage section 708 into a Random Access Memory (RAM) 703. The processor 701 may include, for example, a general purpose microprocessor (e.g., a CPU), an instruction set processor and/or associated chipset, and/or a special purpose microprocessor (e.g., an Application Specific Integrated Circuit (ASIC)), among others. The processor 701 may also include on-board memory for caching purposes. The processor 701 may comprise a single processing unit or a plurality of processing units for performing the different actions of the method flows according to embodiments of the present disclosure.
In the RAM 703, various programs and data necessary for the operation of the system 700 are stored. The processor 701, the ROM 702, and the RAM 703 are connected to each other by a bus 704. The processor 701 performs various operations of the method flows according to the embodiments of the present disclosure by executing programs in the ROM 702 and/or the RAM 703. It is noted that the programs may also be stored in one or more memories other than the ROM 702 and RAM 703. The processor 701 may also perform various operations of method flows according to embodiments of the present disclosure by executing programs stored in the one or more memories.
Optionally, the system 700 may also include an input/output (I/O) interface 705, the input/output (I/O) interface 705 also being connected to the bus 704. The system 700 may also include one or more of the following components connected to the I/O interface 705: an input portion 706 including a keyboard, a mouse, and the like; an output section 707 including a display such as a Cathode Ray Tube (CRT), a Liquid Crystal Display (LCD), and the like, and a speaker; a storage section 708 including a hard disk and the like; and a communication section 709 including a network interface card such as a LAN card, a modem, or the like. The communication section 709 performs communication processing via a network such as the internet. A drive 710 is also connected to the I/O interface 706 as needed. A removable medium 711 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is mounted on the drive 710 as necessary, so that a computer program read out therefrom is mounted into the storage section 708 as necessary.
Alternatively, the method flows according to embodiments of the present disclosure may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable storage medium, the computer program containing program code for performing the method illustrated by the flow chart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication section 709, and/or installed from the removable medium 711. The computer program, when executed by the processor 701, performs the above-described functions defined in the system of the embodiment of the present disclosure. Alternatively, the systems, devices, apparatuses, modules, units, etc. described above may be implemented by computer program modules.
The present disclosure also provides a computer-readable storage medium, which may be contained in the apparatus/device/system described in the above embodiments; or may exist separately and not be assembled into the device/apparatus/system. The computer-readable storage medium carries one or more programs which, when executed, implement the method according to an embodiment of the disclosure.
Alternatively, the computer-readable storage medium may be a non-volatile computer-readable storage medium, which may include, for example but is not limited to: a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present disclosure, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium may optionally include one or more memories other than the ROM 702 and/or RAM 703 and/or ROM 702 and RAM 703 described above.
The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams or flowchart illustration, and combinations of blocks in the block diagrams or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
Those skilled in the art will appreciate that various combinations and/or combinations of features recited in the various embodiments and/or claims of the present disclosure can be made, even if such combinations or combinations are not expressly recited in the present disclosure. In particular, various combinations and/or combinations of the features recited in the various embodiments and/or claims of the present disclosure may be made without departing from the spirit or teaching of the present disclosure. All such combinations and/or associations are within the scope of the present disclosure.
The embodiments of the present disclosure have been described above. However, these examples are for illustrative purposes only and are not intended to limit the scope of the present disclosure. Although the embodiments are described separately above, this does not mean that the measures in the embodiments cannot be used in advantageous combination. The scope of the disclosure is defined by the appended claims and equivalents thereof. Various alternatives and modifications can be devised by those skilled in the art without departing from the scope of the present disclosure, and such alternatives and modifications are intended to be within the scope of the present disclosure.
Claims (10)
1. An interaction processing method, comprising:
acquiring interaction history information, wherein the interaction history information comprises input history information and/or response history information;
determining an interaction intention of the interaction object indicated by the interaction history information by using a preset intention recognition model;
and determining and outputting interaction response information for feeding back to the interaction object based on the interaction history information and the interaction intention.
2. The method of claim 1, wherein the determining interaction response information for feedback to the interaction object based on the interaction history information and the interaction intent comprises:
determining at least one intention descriptor associated with the interaction history information according to a preset association relationship between the interaction intention and the intention descriptors, wherein the association relationship is obtained by training based on interaction sample data by using an intention recognition model;
determining the interaction response information based on the interaction history information and the at least one intent descriptor.
3. The method of claim 2, wherein the determining the interaction response information based on the interaction history information and the at least one intent descriptor comprises:
determining an interaction candidate information base associated with the interaction intention, the interaction candidate information base comprising at least one input candidate information and at least one response candidate information;
determining target input candidate information with the highest matching degree with the interaction history information in the interaction candidate information base on the basis of the interaction history information and the at least one intention descriptor;
and using response candidate information associated with the target input candidate information as the interaction response information.
4. The method of claim 3, wherein the determining, in the interaction candidate information base, the target input candidate information that matches the interaction history information most highly based on the interaction history information and the at least one intent descriptor comprises:
aiming at any input candidate information, determining a first matching degree between the interaction history information and the input candidate information;
determining a second degree of match between the at least one intent descriptor and the input candidate information;
determining a third matching degree for the input candidate information based on the first matching degree and the second matching degree;
repeating the steps until a third matching degree aiming at each input candidate information is determined;
determining the target input candidate information based on at least the third matching degree for each of the input candidate information.
5. The method of claim 4, wherein the determining the target input candidate information based at least on a third degree of match for each of the input candidate information comprises:
for any one of the input candidate information, calculating a similarity between the average word vector of the at least one intention descriptor and the average word vector of the input candidate information as a fourth matching degree for the input candidate information;
calculating the number of the intention descriptors contained in the input candidate information as a fifth matching degree for the input candidate information;
calculating an average similarity between all the component words in the input candidate information and the at least one intention descriptor as a sixth matching degree for the input candidate information;
and determining the target input candidate information based on the third matching degree, the fourth matching degree, the fifth matching degree and the sixth matching degree.
6. The method of claim 2, wherein the determining the interaction response information based on the interaction history information and the at least one intent descriptor comprises:
generating a model by using preset interaction information, and performing coding operation based on the interaction history information to obtain an intermediate feature vector; and
performing a decoding operation based on the intermediate feature vector and the at least one intent descriptor to generate the interaction response information.
7. The method of claim 6, comprising, during training of the mutual information generation model:
performing coding operation based on interactive sample information by using the interactive information generation model to obtain a sample intermediate feature vector, wherein the interactive sample information comprises sample input information and sample response information, and the interactive sample information has a sample label indicating an interactive intention;
performing decoding operation based on the sample intermediate characteristic vector to obtain prediction response information;
calculating a generation loss of the interaction information generation model based on the predicted response information and the sample response information;
calculating a classification loss of the interaction information generation model based on the sample labels and the interaction intents indicated by the predicted response information;
and determining the total loss of the interaction information generation model based on the generation loss and the classification loss so as to perform model training based on the total loss.
8. An interaction processing apparatus comprising:
the acquisition module is used for acquiring interaction history information, wherein the interaction history information comprises input history information and/or response history information;
the first processing module is used for determining the interaction intention of the interaction object indicated by the interaction history information by using a preset intention recognition model;
and the second processing module is used for determining and outputting interaction response information for feeding back to the interaction object based on the interaction history information and the interaction intention.
9. An electronic device, comprising:
one or more processors; and
a memory for storing one or more programs,
wherein the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the method of any of claims 1-7.
10. A computer readable storage medium having stored thereon executable instructions which, when executed by a processor, cause the processor to carry out the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011282527.XA CN112347770A (en) | 2020-11-16 | 2020-11-16 | Interactive processing method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011282527.XA CN112347770A (en) | 2020-11-16 | 2020-11-16 | Interactive processing method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112347770A true CN112347770A (en) | 2021-02-09 |
Family
ID=74362922
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011282527.XA Pending CN112347770A (en) | 2020-11-16 | 2020-11-16 | Interactive processing method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112347770A (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109087135A (en) * | 2018-07-25 | 2018-12-25 | 百度在线网络技术（北京）有限公司 | The method for digging and device, computer equipment and readable medium that user is intended to |
| CN111611358A (en) * | 2019-02-25 | 2020-09-01 | 北京嘀嘀无限科技发展有限公司 | Information interaction method and device, electronic equipment and storage medium |
| WO2020215914A1 (en) * | 2019-04-23 | 2020-10-29 | 北京京东尚科信息技术有限公司 | Information processing method, information processing device, and information processing system |
-
2020
- 2020-11-16 CN CN202011282527.XA patent/CN112347770A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109087135A (en) * | 2018-07-25 | 2018-12-25 | 百度在线网络技术（北京）有限公司 | The method for digging and device, computer equipment and readable medium that user is intended to |
| US20200034431A1 (en) * | 2018-07-25 | 2020-01-30 | Baidu Online Network Technology (Bijing ) Co., Ltd. | Method, computer device and readable medium for user's intent mining |
| CN111611358A (en) * | 2019-02-25 | 2020-09-01 | 北京嘀嘀无限科技发展有限公司 | Information interaction method and device, electronic equipment and storage medium |
| WO2020215914A1 (en) * | 2019-04-23 | 2020-10-29 | 北京京东尚科信息技术有限公司 | Information processing method, information processing device, and information processing system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109101537B (en) | Multi-turn dialogue data classification method and device based on deep learning and electronic equipment | |
| CN110377740B (en) | Emotion polarity analysis method and device, electronic equipment and storage medium | |
| WO2020224219A1 (en) | Chinese word segmentation method and apparatus, electronic device and readable storage medium | |
| CN111695352A (en) | Grading method and device based on semantic analysis, terminal equipment and storage medium | |
| CN113313022A (en) | Training method of character recognition model and method for recognizing characters in image | |
| CN111241237A (en) | Intelligent question and answer data processing method and device based on operation and maintenance service | |
| CN112100332A (en) | Word embedding expression learning method and device and text recall method and device | |
| CN111221936B (en) | Information matching method and device, electronic equipment and storage medium | |
| CN112328761A (en) | Intention label setting method and device, computer equipment and storage medium | |
| CN111858843A (en) | Text classification method and device | |
| US11676410B1 (en) | Latent space encoding of text for named entity recognition | |
| CN114387061A (en) | Product pushing method and device, electronic equipment and readable storage medium | |
| CN113268615A (en) | Resource label generation method and device, electronic equipment and storage medium | |
| CN113326702A (en) | Semantic recognition method and device, electronic equipment and storage medium | |
| CN113051380A (en) | Information generation method and device, electronic equipment and storage medium | |
| CN111738791B (en) | Text processing method, device, equipment and storage medium | |
| CN111368066B (en) | Method, apparatus and computer readable storage medium for obtaining dialogue abstract | |
| CN115757731A (en) | Dialogue question rewriting method, device, computer equipment and storage medium | |
| CN113342977B (en) | Invoice image classification method, device, equipment and storage medium | |
| CN110717333A (en) | Method and device for automatically generating article abstract and computer readable storage medium | |
| CN115718889A (en) | Industry classification method and device for company profile | |
| CN116186219A (en) | Man-machine dialogue interaction method, system and storage medium | |
| CN113505293B (en) | Information pushing method and device, electronic equipment and storage medium | |
| CN115358817A (en) | Intelligent product recommendation method, device, equipment and medium based on social data | |
| CN112347770A (en) | Interactive processing method and device |
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 |