CN112988956B - Method and device for automatically generating dialogue, and method and device for detecting information recommendation effect - Google Patents
Method and device for automatically generating dialogue, and method and device for detecting information recommendation effect Download PDFInfo
- Publication number
- CN112988956B CN112988956B CN201911298133.0A CN201911298133A CN112988956B CN 112988956 B CN112988956 B CN 112988956B CN 201911298133 A CN201911298133 A CN 201911298133A CN 112988956 B CN112988956 B CN 112988956B
- Authority
- CN
- China
- Prior art keywords
- dialogue
- information
- script
- target
- user
- 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
- 238000000034 method Methods 0.000 title claims abstract description 73
- 230000000694 effects Effects 0.000 title claims abstract description 29
- 238000013515 script Methods 0.000 claims abstract description 146
- 238000001514 detection method Methods 0.000 claims abstract description 43
- 230000009471 action Effects 0.000 claims abstract description 40
- 230000006399 behavior Effects 0.000 claims abstract description 37
- 238000004088 simulation Methods 0.000 claims abstract description 35
- 238000006243 chemical reaction Methods 0.000 claims description 18
- 230000015654 memory Effects 0.000 claims description 17
- 238000003860 storage Methods 0.000 claims description 12
- 230000001960 triggered effect Effects 0.000 claims description 4
- 230000000875 corresponding effect Effects 0.000 description 34
- 238000010586 diagram Methods 0.000 description 19
- 238000013459 approach Methods 0.000 description 12
- 238000004891 communication Methods 0.000 description 12
- 238000012545 processing Methods 0.000 description 11
- 230000008569 process Effects 0.000 description 9
- 238000005457 optimization Methods 0.000 description 6
- 238000011161 development Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 5
- 230000005291 magnetic effect Effects 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 4
- 230000005236 sound signal Effects 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 238000012937 correction Methods 0.000 description 3
- 230000007547 defect Effects 0.000 description 3
- 241000989913 Gunnera petaloidea Species 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 2
- 125000004122 cyclic group Chemical group 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000007726 management method Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000003825 pressing Methods 0.000 description 2
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical compound [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 230000002085 persistent effect Effects 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
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
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Human Computer Interaction (AREA)
- Machine Translation (AREA)
Abstract
The invention discloses a method and a device for automatically generating a dialogue, wherein the method comprises the following steps: determining a simulation user, a dialogue field and a dialogue intention of the dialogue, and generating a dialogue target in a semantic level; determining personalized information of the simulated user; acquiring a pre-established dialogue action script corresponding to the dialogue field and the dialogue intention; extracting a script instruction from the dialogue action script, and filling slot position information in the script instruction according to the dialogue target to obtain the script instruction of the dialogue; generating a dialogue text according to the script instruction of the dialogue and the personalized information; and outputting the dialogue text as a simulation result of user dialogue input. The invention also discloses a method and a device for detecting the information recommendation effect. By utilizing the invention, various dialogue behaviors of the user chat can be completely and effectively simulated, and further the detection of the information recommendation system can be conveniently realized.
Description
Technical Field
The invention relates to the field of information recommendation, in particular to a method and a device for automatically generating a dialogue, and also relates to a method and a device for detecting information recommendation effect.
Background
Instant chat tools, which can deliver instant messages between two or more users, are one of the most popular internet communication tools. In the prior art, some instant chat tools judge entities and texts contained in user input, and provide relevant services such as searching, information recommending and the like according to the current chat intention of the user, so that the chat tools become more intelligent, and the user experience is enhanced. However, most of the current systems for recommending information in chat scenes only depend on text information in current user input strings, and the provided information recommending service generally lacks pertinence and has poor recommending effect. Therefore, how to detect the information recommendation quality, find out the problem, and further perform targeted optimization and promotion on the existing information recommendation system is an important problem to be solved at present. In the development, detection and optimization process of the information recommendation system, if the real chat data of the user is directly used, the potential risk on the privacy of the user is brought.
Disclosure of Invention
On one hand, the embodiment of the invention provides a method and a device for automatically generating a dialogue, which can completely and effectively simulate various dialogue behaviors of user chatting.
On the other hand, the embodiment of the invention provides a method and a device for detecting information recommendation effect, which can conveniently realize the detection of an information recommendation system.
Therefore, the invention provides the following technical scheme:
a method of automatically generating a dialog, the method comprising:
determining a simulation user, a dialogue field and a dialogue intention of the dialogue, and generating a dialogue target in a semantic level;
Determining personalized information of the simulated user;
Acquiring a pre-established dialogue action script corresponding to the dialogue field and the dialogue intention;
Extracting a script instruction from the dialogue action script, and filling slot position information in the script instruction according to the dialogue target to obtain a target script instruction of the dialogue;
Generating a dialogue text according to the target script instruction of the dialogue and the personalized information;
And outputting the dialogue text as a simulation result for simulating the dialogue input of the user.
Optionally, the semantic-level dialogue goal includes a series of slots and slot value information corresponding to the slots.
Optionally, the generating the semantic-level dialog target includes:
acquiring text content of dialogue target description of natural language;
Extracting entity information from the text content;
And filling the entity information into preset slots to generate a series of slots and slot value information corresponding to the slots.
Optionally, the personalized information of the simulated user includes any one or more of the following: age, sex, personality, speech style.
Optionally, the dialog intention includes: informing of intent and asking for intent.
Optionally, the method further comprises: pre-establishing a template library;
The generating the dialogue text according to the target script instruction of the dialogue and the personalized information comprises the following steps:
searching the template library according to the target script instruction of the current dialogue and the personalized information to obtain a dialogue template;
and filling the dialogue template according to the slot position information in the target script instruction of the dialogue, so as to obtain dialogue texts.
Optionally, the method further comprises: pre-establishing a dialogue generating model;
The generating the dialogue text according to the target script instruction of the dialogue and the personalized information comprises the following steps:
and inputting the information in the target script instruction of the current dialogue and the personalized information into the dialogue generating model to obtain a dialogue text.
Optionally, the method further comprises: a knowledge base containing historical input information and recommended dialogue content is established in advance;
The generating the dialogue text according to the target script instruction of the dialogue and the personalized information comprises the following steps:
Searching the knowledge base according to the target script instruction of the current dialogue and the personalized information to obtain recommended dialogue content;
And taking the recommended dialogue content as dialogue text.
Optionally, the method further comprises: a knowledge base containing historical input information and recommended dialogue contents and a dialogue conversion model are established in advance;
the generating a dialogue text according to the target script instruction, the personalized information and the dialogue target comprises:
retrieving the knowledge base according to the target script instruction to obtain recommended dialogue content;
And inputting the personalized information and the recommended dialogue content into the dialogue conversion model to obtain dialogue texts.
An information recommendation effect detection method, the method comprising:
simulating user input behaviors by using a pre-established dialogue behavior script to automatically generate a series of dialogue information, and inputting the dialogue information into an information recommendation system;
If the recommended information of the information recommending system aiming at the dialogue information is received, inputting the dialogue information and the recommended information into a pre-established correlation judging model, and obtaining a correlation detection result according to the output of the correlation judging model;
and recording the correlation detection result.
Alternatively, a series of dialogue information is generated using the method of automatically generating a dialogue described above.
Optionally, the method further comprises:
and scoring the information recommendation system according to the correlation detection result to obtain a scoring result.
An apparatus for automatically generating a dialog, the apparatus comprising:
the setting module is used for determining the simulation user, the dialogue field and the dialogue intention of the dialogue;
The dialogue target generation module is used for generating a dialogue target of a semantic level;
the personalized information determining module is used for determining personalized information of the simulated user;
The script acquisition module is used for acquiring a pre-established dialogue action script corresponding to the dialogue field and the dialogue intention;
The script instruction generation module is used for extracting script instructions from the dialogue action scripts and filling slot position information in the script instructions according to the dialogue targets to obtain script instructions of the dialogue;
The dialogue generating module is used for generating dialogue texts according to the script instructions of the dialogue and the personalized information;
And the output module is used for outputting the dialogue text as a simulation result of user dialogue input.
Optionally, the semantic-level dialogue goal includes a series of slots and slot value information corresponding to the slots.
Optionally, the dialog target generation module includes:
The text receiving unit is used for acquiring text contents of natural language description;
an entity extraction unit for extracting entity information from the text content;
And the target generation unit is used for filling the entity information into preset slots and generating a series of slots and slot value information corresponding to the slots.
Optionally, the personalized information of the simulated user includes any one or more of the following: age, sex, personality, speech style.
Optionally, the dialog intention includes: informing of intent and asking for intent.
Optionally, the apparatus further comprises: the template library establishing module is used for establishing a template library in advance;
The dialogue generation module comprises:
the first retrieval unit is used for retrieving the template library according to the script instruction of the current dialogue and the personalized information to obtain a dialogue template;
And the slot filling unit is used for filling the dialogue template according to the slot information in the script instruction of the current dialogue to obtain a dialogue text.
Optionally, the apparatus further comprises: the dialogue generation model building module is used for pre-building a dialogue generation model;
The dialogue generating module is specifically configured to input the information in the script instruction of the current dialogue and the personalized information into the dialogue generating model to obtain a dialogue text.
Optionally, the apparatus further comprises: the knowledge base establishing module is used for establishing a knowledge base containing historical input information and recommended dialogue contents in advance;
the dialogue generating module is specifically configured to retrieve the knowledge base according to the script instruction of the current dialogue and the personalized information, obtain recommended dialogue content, and use the recommended dialogue content as a dialogue text.
Optionally, the apparatus further comprises:
The knowledge base establishing module is used for establishing a knowledge base containing historical input information and recommended dialogue contents in advance;
the dialogue conversion model building module is used for pre-building a dialogue conversion model;
The dialogue generation module comprises:
the second retrieval unit is used for retrieving the knowledge base according to the script instruction to obtain recommended dialogue content;
and the text conversion unit is used for inputting the personalized information and the recommended dialogue content into the dialogue conversion model to obtain dialogue texts.
An information recommendation effect detection apparatus, the apparatus comprising:
A dialogue simulator for simulating user input behaviors by using a pre-established dialogue behavior script to automatically generate a series of dialogue information and inputting the dialogue information into an information recommendation system;
the detection module is used for inputting the dialogue information and the recommended information into a pre-established correlation judgment model when receiving the recommended information of the information recommendation system aiming at the dialogue information, and obtaining a correlation detection result according to the output of the correlation judgment model;
And the recording module is used for recording the correlation detection result. The dialogue module includes the following claims
Optionally, the dialog simulator comprises the means for automatically generating dialog as described above.
Optionally, the apparatus further comprises:
And the scoring module is used for scoring the information recommendation system according to the correlation detection result to obtain a scoring result.
A computer device, comprising: one or more processors, memory;
The memory is configured to store computer-executable instructions and the processor is configured to execute the computer-executable instructions to implement the methods described above.
A readable storage medium having stored thereon instructions that are executed to implement the method described previously.
When user input behaviors need to be simulated, the method and the device for automatically generating the dialogue determine the dialogue field and the dialogue intention, generate the dialogue target at the semantic level, determine the personalized information of the simulated user, obtain the script instruction of the dialogue by utilizing the pre-established dialogue behavior script corresponding to the dialogue field and the dialogue intention, generate the dialogue text according to the script instruction of the dialogue and the personalized information, and output the dialogue text as the simulation result of the user input behaviors, thereby being capable of completely and effectively simulating various dialogue behaviors of the user chat.
The information recommendation effect detection method and device provided by the embodiment of the invention are mainly used for detecting the effect of information recommendation of a system which carries out information recommendation in a dialogue scene such as a chat scene using a chat tool. By simulating the dialogue requirement and intention of the real user to generate dialogue sentences, the defect of detection by using the real chat data of the user is avoided. The method can completely simulate various dialogue behaviors of user chatting, so that the information recommendation effect of the information recommendation system in various different dialogue environments and in the single-round dialogue or multi-round dialogue process can be comprehensively detected, and effective help is provided for development, detection and optimization of the information recommendation system.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments described in the present application, and other drawings may be obtained according to these drawings for a person having ordinary skill in the art.
FIG. 1 is a flow chart of a method of automatically generating a dialog in accordance with an embodiment of the present invention;
FIG. 2 is a block diagram of an apparatus for automatically generating a dialog in accordance with an embodiment of the present invention;
FIG. 3 is a flowchart of a method for detecting an information recommendation effect according to an embodiment of the present invention;
FIG. 4 is another flowchart of a method for detecting an information recommendation effect according to an embodiment of the present invention;
FIG. 5 is a block diagram showing a structure of an information recommendation effect detecting apparatus according to an embodiment of the present invention;
FIG. 6 is a block diagram of an apparatus for automatically generating a dialog, according to an exemplary embodiment;
fig. 7 is a schematic structural diagram of a server according to an embodiment of the present invention.
Detailed Description
In order to make the solution of the embodiment of the present invention better understood by those skilled in the art, the embodiment of the present invention is further described in detail below with reference to the accompanying drawings and embodiments.
The embodiment of the invention provides a method and a device for automatically generating a dialogue, which are used for establishing dialogue action scripts corresponding to various dialogue fields and dialogue intentions in advance, determining the dialogue field and the dialogue intentions when user input actions need to be simulated, generating a dialogue target at a semantic level, determining personalized information of a simulated user, generating a dialogue text by using the dialogue action scripts corresponding to the dialogue field and the dialogue intentions, and outputting the dialogue text as a simulation result of the user input actions.
As shown in fig. 1, a flowchart of a method for automatically generating a dialogue according to an embodiment of the invention includes the following steps:
Step 101, determining the simulation user, the dialogue field and the dialogue intention of the dialogue, and generating the dialogue target of the semantic layer.
In practical applications, a plurality of simulation users with different characteristics can be set, and each simulation user sets a user name, that is, different user names are used for distinguishing different simulation users.
The dialogue field can be various, such as star eight diagrams, news, movies, travel, etc. Each time a user session is simulated, the corresponding session field can be set as desired.
The dialog intention refers to the purpose of simulating a user utterance (typically at the semantic level). In the embodiment of the present invention, the dialog intention may be set as two of the following: informing of intent and asking for intent. Wherein the informing intention means that the purpose of the utterance is to inform a certain information, such as "I are the iron powder of Yang Mi-! "the dialogue intention of this sentence is a notification, which may be formally expressed as" intention=notification "; the query intent refers to the purpose of an utterance to query for certain information, such as "do you know the birth date of Yang Mi? "the dialog intention of this sentence is a query, which may be formally expressed as" intention=query ". Of course, according to actual needs, other kinds of intention are also possible, and the embodiment of the present invention is not limited thereto.
The semantic-level dialogue goal includes a series of slots and slot value information corresponding to the slots, i.e. "slot-slot value" information, such as: star = wang someplace, tv series = give a full account orders, news =? . There are slots with known slot values, and slots with unknown slot values, i.e. information that the user wishes to obtain.
In practical application, when simulating user dialogue each time, the user can directly input dialogue target in semantic level; or inputting a dialogue target in a natural language layer, and then generating a dialogue target in a semantic layer according to the dialogue target in the natural language layer, wherein the specific process is as follows: and acquiring text content of dialogue target description of natural language, extracting entity information from the text content, filling the entity information into preset slots, and generating a series of slot-slot value information serving as a semantic dialogue target. For example, dialogue goals for the input natural language are: know about the news of a certain television show give a full account. First, entity information is extracted therefrom: "wang somebody", "drama", "give a full account reams", "news", then these entity information are filled into some preset slots, so as to obtain the dialogue goal of semantic level.
Step 102, personalized information of the simulated user is determined.
In practical application, the personalized information of the simulation user can be set when the simulation user dialogues each time; corresponding user portraits can also be preset for user names representing different simulation users, wherein the user portraits contain all personalized information of the simulation users, and the personalized information comprises any one or more of the following components: age, sex, personality, speech style. For example, the personalized information in the user portraits with the user names of "king" and "king" is: age=21 years, sex=male, character=liveness, speaking style=humour, and the like. Therefore, after the user name of the current dialogue is set, the user portrait information is directly read, so that the personalized information corresponding to the simulated user can be obtained, the dialogue generation efficiency is improved, and the personalized information corresponding to each user name can be flexibly and conveniently adjusted according to the dialogue environment requirement.
Step 103, obtaining a pre-established dialogue action script corresponding to the dialogue domain and the dialogue intention.
In practical application, dialogue action scripts corresponding to various dialogue fields and dialogue intents can be established in advance, and put into a script library. And searching a script library according to the dialogue field and the dialogue intention to obtain a corresponding dialogue action script when required.
The dialog behavior script mainly includes a series of semantic information to be issued to complete a specific dialog goal. In each dialog behavior script, a series of script instructions are included that simulate the text generation instructions generated by a series of dialog actions of a user.
The contents of each dialog behavior script file are as follows:
1. A continuous segment of script in the script file represents the process by which the user continuously inputs information to achieve the goal. Wherein a line of scripts records a specific chat behavior triggered by a specific intention of the user.
2. The intention of the chat action of the user, the slot information to be expressed and the like are recorded in each row of script.
For example, when the dialogue field is "the star eight diagrams" and the dialogue intention is "the inquiry", the script instruction is extracted as follows:
1. Domain = star eight diagrams, intention = inquiry drama, slot information = { drama }
The dialogue acts corresponding to the script instructions are: generating sentences for inquiring names of television drama;
2. domain = star eight diagrams, intent = query news, slot information = { television show, star name }
The dialogue acts corresponding to the script instructions are as follows: a statement is generated asking for star news.
And 104, extracting a script instruction from the dialogue action script, and filling slot position information in the script instruction according to the dialogue target to generate a target script instruction of the current dialogue.
Taking the script instruction as an example, the generated target script instruction of the current dialogue is as follows:
Domain = star eight diagrams, intention = inquiry drama, slot information = { drama = give a full account reams }
Domain = star eight diagrams, intention = query news, slot information = { television series = give a full account order, star name = wang someplace }
And 105, generating a dialogue text according to the target script instruction of the dialogue and the personalized information.
In a particular application, dialog text may be generated in a variety of ways, such as: (1) a template-based approach; (2) a manner based on the generation formula; (3) a manner based on a retrievable formula; (4) hybrid-based approach.
The process of generating the dialog text in the above four different ways is described in detail below.
(1) Template-based approach
In a template-based approach, a template library may be pre-established, with a series of related templates corresponding to each domain, each template also differing for different personalized information. When a dialogue text is generated, firstly, searching the template library according to the target script instruction of the current dialogue and the personalized information to obtain a dialogue template, and then filling the dialogue template according to the slot position information in the target script instruction of the current dialogue to obtain the dialogue text.
For example, for the target script instruction of "field=star eight diagrams, intention=inquiry drama, slot information= { drama= give a full account }" described above, the template library is searched under the condition of "field=star eight diagrams" and "intention=inquiry drama" to obtain the following two templates:
template a: "do you pay attention to { drama } recently? "(character = stable in sinking)
Template B: "haha, { TV play } recently very fire-! "(character=lively)
Since the personalized information of the simulation user has "character=liveness", the template B is selected.
Then filling the template B according to 'slot information = { television play = give a full account order }' in the target script instruction of the current dialogue, and generating a dialogue text as follows: "haha, give a full account makes the most recent fire o-! ".
Continuing to search the template library for the target script instruction of the above-mentioned "field=star eight diagrams, intention=query news, slot information= { television series= give a full account order, star name=wang somehow }", and taking the "field=star eight diagrams" and "intention=query news" as conditions to obtain the following two templates:
template a: "do you know the role of { star } in this section { TV play? "(user personality = stable)
Template B: "listening to { stars } is excellent in this section { drama }, and does not know how does it be? "(user character = liveness)
Since the personalized information of the simulation user has "character=liveness", the template B is selected.
Then filling the template B according to the 'slot information = { television play = give a full account order, star name = wang somewhere }' in the script instruction of the current dialogue, and generating a dialogue text as follows: "what is best seen in this section give a full account, how does it know the specific situation? ".
And generating the series of personalized natural languages to obtain simulation results of the input behaviors of the two sentences of users.
(2) Based on the way of generating
In a manner based on the generation formula, a dialogue generation model may be established in advance. The dialogue generation model may adopt a neural network model, for example, a seq2seq model for fitting the input information-reply relationship may be constructed based on RNN (RecurrentNeuralNetwork, cyclic neural network), LSTM (Long Short-Term Memory network), GRU (Gated Recurrent Unit, gated cyclic unit) and other structures.
And when the dialogue text is generated, inputting the information in the target script instruction of the dialogue and the personalized information into the dialogue generation model to obtain the dialogue text.
(3) Search-based mode
In the search-based approach, a knowledge base including historical input information and recommended dialog content needs to be established in advance.
And when a dialogue text is generated, searching the knowledge base according to the field, the slot position information and the personalized information contained in the target script instruction of the dialogue, finding similar historical input information, obtaining recommended dialogue content corresponding to the historical input information, and taking the recommended dialogue content as the dialogue text.
For example, the knowledge base includes the history input information "give a full account orders of a main show of wang" and the recommended dialogue content corresponding to the history input information is "excellent in hearing and speaking a main show of a series, and how is the specific situation? ".
The target script instruction of this dialogue is as follows:
Domain = star eight diagrams, intent = query news, slot information = { television series = give a full account orders, star name = wang someplace };
the personalized information has character=liveness;
Then, according to the target script instruction and the personalized information, similar historical input information ' give a full account orders of a king certain main show ' is found in the knowledge base, and further dialogue content ' excellent in hearing and speaking of the king certain show in the drama ' corresponding to the historical input information ' is obtained, and how is the specific situation? ".
It should be noted that, the determination of the similar historical input information may be determined according to the semantic correlation between the two, and the calculation of the correlation may use the prior art, which will not be described in detail herein.
(4) Based on a hybrid approach
In the hybrid-based approach, a knowledge base and a session conversion model including historical input information and recommended session content need to be established in advance.
The dialogue conversion model is similar to the model structure based on the generation formula, and the seq2seq model fitting the input information-reply relation can be constructed by using the structures based on RNN, LSTM, GRU and the like.
When a dialogue text is generated, firstly searching the knowledge base according to the target script instruction of the dialogue, finding out history input information similar to the target script instruction, and further obtaining recommended dialogue content corresponding to the history input information; and then inputting the personalized information and the recommended dialogue content into the dialogue conversion model to obtain dialogue texts.
This method not only inherits the advantages of the search formula-based method (3) described above, namely that the dialogue content is smoothly generated and is rich in information, but also enjoys the advantages of the generation formula-based method in terms of diversity and originality.
For example, the recommended dialogue content obtained by using the search knowledge base is "excellent in hearing and speaking about what is shown in the series, how is something no matter what is? By inputting the recommended dialogue content to the encoding end of the dialogue conversion model, and inputting the slot "character=lively", "star=wang somewhere, and" tv series= give a full account command "which is desired to affect decoding to the decoding end, the recommended content rich in the change can be obtained, for example: ' good points o! Listening to the king is excellent in give a full account years, and immediately looking at-! ".
And 106, outputting the dialogue text as a simulation result for simulating the dialogue input of the user.
When user input behaviors need to be simulated, the method for automatically generating the dialogue determines the dialogue field and the dialogue intention, generates a dialogue target at a semantic level, determines personalized information of a simulated user, obtains a target script instruction of the dialogue by utilizing a pre-established dialogue behavior script corresponding to the dialogue field and the dialogue intention, then generates a dialogue text according to the target script instruction of the dialogue and the personalized information, and outputs the dialogue text as a simulation result for simulating the user input behaviors, thereby being capable of completely and effectively simulating various dialogue behaviors of user chatting.
Correspondingly, the embodiment of the invention also provides a device for automatically generating the dialogue, and the device is a structural block diagram as shown in fig. 2.
In this embodiment, the apparatus comprises the following modules:
A setting module 201, configured to determine a simulated user, a dialogue domain, and a dialogue intention of the present dialogue;
A dialogue target generation module 202, configured to generate a semantic-level dialogue target;
A personalized information determining module 203, configured to determine personalized information of the simulated user;
A script obtaining module 204, configured to obtain a pre-established dialogue action script corresponding to the dialogue domain and the dialogue intention;
The script instruction generating module 205 is configured to extract a script instruction from the dialogue action script, and fill slot information in the script instruction according to the dialogue target, so as to obtain a script instruction of the present dialogue;
A dialogue generation module 206, configured to generate a dialogue text according to the script instruction of the current dialogue and the personalized information;
And the output module 207 is used for outputting the dialogue text as a simulation result of user dialogue input.
In the embodiment of the invention, the dialogue target at the semantic level comprises a series of slots and slot values corresponding to the slots, namely 'slot-slot value' information. In practical applications, each time a user dialogue is simulated, the user can directly input a dialogue target of a semantic level to the dialogue target generation module 202; a natural language-level dialog target may also be input, and then the dialog target generating module 202 generates a semantic-level dialog target according to the natural language-level dialog target, and accordingly, a specific structure of the dialog target generating module 202 may include the following units:
The text receiving unit is used for acquiring text contents of natural language description;
an entity extraction unit for extracting entity information from the text content;
and the target generation unit is used for filling the entity information into preset slots and generating a series of slot-slot value information.
In the embodiment of the invention, the personalized information of the simulation user comprises, but is not limited to, any one or more of the following: age, sex, personality, speech style.
In practical application, the user may input the personalized information of the simulated user to the personalized information determining module 203 each time the simulated user dialogues; corresponding user portraits may be preset for user names representing different simulation users, where the user portraits include all personalized information of the simulation users, and after the user names of the current dialogue are set, the personalized information determining module 203 reads the personalized information corresponding to the simulation users from the corresponding user portraits.
In an embodiment of the present invention, the dialog intention may include: informing of intent and asking for intent. Of course, according to actual needs, other kinds of intention are also possible, and the embodiment of the present invention is not limited thereto.
In practical application, dialogue action scripts corresponding to various dialogue fields and dialogue intents can be established in advance, and the foot dialogue action scripts are put into a script library. The dialog behavior script mainly includes a series of semantic information to be issued to complete a specific dialog goal. In each dialog behavior script, a series of script instructions are included that simulate the text generation instructions generated by a series of dialog actions of a user.
Accordingly, the script obtaining module 204 may search a script library according to the dialog field and the dialog intention to obtain a corresponding dialog behavior script.
In practical applications, the dialog generation module 206 may generate dialog text in a variety of ways, such as: (1) a template-based approach; (2) a manner based on the generation formula; (3) a manner based on a retrievable formula; (4) hybrid-based approach.
If a template-based approach is employed, the apparatus may further comprise: a template library creation module (not shown) for creating a template library in advance. Accordingly, the dialog generation module 206 may include the following elements:
the first retrieval unit is used for retrieving the template library according to the script instruction of the current dialogue and the personalized information to obtain a dialogue template;
And the slot filling unit is used for filling the dialogue template according to the slot information in the script instruction of the current dialogue to obtain a dialogue text.
If a manner based on a generation formula is adopted, the device can further comprise: a dialogue generation model creation module (not shown) for creating a dialogue generation model in advance. Accordingly, the dialogue generation module 206 may input the information in the script command of the current dialogue and the personalized information into the dialogue generation model to obtain dialogue text.
If a search-based approach is employed, the apparatus may further comprise: a knowledge base establishing module (not shown) for establishing a knowledge base containing history input information and recommended dialogue content in advance. Accordingly, the dialogue generation module 206 may retrieve the knowledge base according to the script command of the current dialogue and the personalized information to obtain recommended dialogue content, and use the recommended dialogue content as dialogue text.
If a hybrid-based approach is employed, the apparatus may further comprise: a knowledge base building module and a dialogue conversion model building module (not shown), wherein: the knowledge base establishing module is used for establishing a knowledge base containing historical input information and recommended dialogue content in advance; the dialogue conversion model building module is used for pre-building a dialogue conversion model. Accordingly, the dialog generation module 206 may include the following elements:
the second retrieval unit is used for retrieving the knowledge base according to the script instruction to obtain recommended dialogue content;
and the text conversion unit is used for inputting the personalized information and the recommended dialogue content into the dialogue conversion model to obtain dialogue texts.
The template library building module, the knowledge base building module, the dialogue generation model building module, and the dialogue conversion model building module, which are not shown, may be part of the apparatus of the present invention, or may be independent of the apparatus of the present invention, and the embodiment of the present invention is not limited thereto.
When user input behaviors need to be simulated, the device for automatically generating the dialogue determines the dialogue field and the dialogue intention, generates a dialogue target of a semantic layer, determines personalized information of a simulated user, obtains a script instruction of the dialogue by utilizing a pre-established dialogue behavior script corresponding to the dialogue field and the dialogue intention, then generates a dialogue text according to the script instruction of the dialogue and the personalized information, and outputs the dialogue text as a simulation result of the user input behaviors, thereby being capable of completely and effectively simulating various dialogue behaviors of user chatting.
Correspondingly, the embodiment of the invention also provides an information recommendation effect detection method and device, which utilize a pre-established dialogue behavior script to simulate the user input behavior to automatically generate a series of dialogue information, and input the dialogue information into an information recommendation system; if the recommended information of the information recommending system aiming at the dialogue information is received, inputting the dialogue information and the recommended information into a pre-established correlation judging model, and obtaining a correlation detection result according to the output of the correlation judging model; and recording the correlation detection result.
As shown in fig. 3, a flowchart of a method for detecting an information recommendation effect according to an embodiment of the present invention includes the following steps:
step 301, automatically generating a series of dialogue information by simulating user input behaviors by using a pre-established dialogue behavior script, and inputting the dialogue information into an information recommendation system.
The process of generating the dialogue information may refer to the descriptions in the embodiments of the method for automatically generating the dialogue according to the present invention, which are not described herein.
Step 302, if the recommended information of the information recommendation system for the dialogue information is received, inputting the dialogue information and the recommended information into a pre-established correlation judgment model, and obtaining a correlation detection result according to the output of the correlation judgment model.
And step 303, recording the correlation detection result. The correlation detection result includes: correlated, uncorrelated.
In another embodiment of the information recommendation effect detection method of the present invention, as shown in fig. 4, the method may further include step 304 of scoring the information recommendation system according to the correlation detection result to obtain a scoring result. For example, the number of correlations and uncorrelated in the correlation detection result is counted, and the correlation rate of the recommendation information is calculated.
The information recommendation effect detection method provided by the embodiment of the invention is mainly used for detecting the effect of information recommendation of a system for carrying out information recommendation in a dialogue scene such as a chat scene using a chat tool. By simulating the dialogue requirement and intention of the real user to generate dialogue sentences, the defect of detection by using the real chat data of the user is avoided. The method can completely simulate various dialogue behaviors of user chatting, so that the information recommendation effect of the information recommendation system in various different dialogue environments and in the single-round dialogue or multi-round dialogue process can be comprehensively detected, and effective help is provided for development, detection and optimization of the information recommendation system.
Correspondingly, the embodiment of the invention also provides an information recommending effect detecting device, as shown in fig. 5, which is a structural block diagram of the device.
In this embodiment, the apparatus comprises the following modules:
A dialogue simulator 501 for automatically generating a series of dialogue information by simulating user input actions using a pre-established dialogue action script, and inputting the dialogue information to the information recommendation system 500;
The detection module 502 is configured to, when receiving recommendation information of the information recommendation system 500 for the dialogue information, input the dialogue information and the recommendation information into a pre-established correlation judgment model, and obtain a correlation detection result according to output of the correlation judgment model;
A recording module 503, configured to record the correlation detection result. The dialogue module includes the following claims
The specific structure of the dialogue module may refer to the aforementioned device for automatically generating dialogue, and will not be described in detail herein.
Further, in another embodiment of the information recommendation effect detection apparatus of the present invention, the apparatus may further include: and the scoring module (not shown) is used for scoring the information recommendation system according to the correlation detection result to obtain a scoring result.
The information recommendation effect detection device provided by the embodiment of the invention mainly aims at detecting the effect of information recommendation of a system for carrying out information recommendation in a dialogue scene such as a chat scene using a chat tool. By simulating the dialogue requirement and intention of the real user to generate dialogue sentences, the defect of detection by using the real chat data of the user is avoided. The method can completely simulate various dialogue behaviors of user chatting, so that the information recommendation effect of the information recommendation system in various different dialogue environments and in the single-round dialogue or multi-round dialogue process can be comprehensively detected, and effective help is provided for development, detection and optimization of the information recommendation system.
According to the information recommendation effect detection method and device provided by the embodiment of the invention, dialogue sentences are generated by simulating dialogue requirements and intentions of real users, so that the output simulated user dialogue accords with a dialogue scene of actual application; moreover, user dialogs with different personalities, different fields and different intentions can be generated, and the richness and the universality of dialog information are ensured. Compared with manual dialogue data and real user dialogue data which can be used only by long-time data collection and cleaning, the method can be used for generating available dialogue data in a large scale in a short time, effectively shortens the evaluation time of the recommendation effect of the information recommendation system, and improves the development and optimization speed of the information recommendation system.
The specific manner in which the various modules perform the operations in the apparatus of the above embodiments have been described in detail in connection with the embodiments of the method, and will not be described in detail herein.
It should be noted that the method and apparatus of the embodiments of the present invention may be applied to various terminal devices, such as a mobile phone, a computer, a notebook, and so on.
Fig. 6 is a block diagram illustrating an apparatus 800 for a method of automatically generating a dialog, according to an example embodiment. For example, apparatus 800 may be a mobile phone, computer, digital broadcast terminal, messaging device, game console, tablet device, medical device, exercise device, personal digital assistant, or the like.
Referring to fig. 6, apparatus 800 may include one or more of the following components: a processing component 802, a memory 804, a power component 806, a multimedia component 808, an audio component 810, an input/output (I/O) interface 812, a sensor component 814, and a communication component 816.
The processing component 802 generally controls overall operation of the apparatus 800, such as operations associated with display, telephone calls, data communications, camera operations, and recording operations. Processing element 802 may include one or more processors 820 to execute instructions to perform all or part of the steps of the methods described above. Further, the processing component 802 can include one or more modules that facilitate interactions between the processing component 802 and other components. For example, the processing component 802 may include a multimedia module to facilitate interaction between the multimedia component 808 and the processing component 802.
The memory 804 is configured to store various types of data to support operations at the device 800. Examples of such data include instructions for any application or method operating on the device 800, contact data, phonebook data, messages, pictures, videos, and the like. The memory 804 may be implemented by any type or combination of volatile or nonvolatile memory devices such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disk.
The power component 806 provides power to the various components of the device 800. The power components 806 may include a power management system, one or more power sources, and other components associated with generating, managing, and distributing power for the device 800.
The multimedia component 808 includes a screen between the device 800 and the user that provides an output interface. In some embodiments, the screen may include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from a user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensor may sense not only the boundary of a touch or slide action, but also the duration and pressure associated with the touch or slide operation. In some embodiments, the multimedia component 808 includes a front camera and/or a rear camera. The front camera and/or the rear camera may receive external multimedia data when the device 800 is in an operational mode, such as a shooting mode or a video mode. Each front camera and rear camera may be a fixed optical lens system or have focal length and optical zoom capabilities.
The audio component 810 is configured to output and/or input audio signals. For example, the audio component 810 includes a Microphone (MIC) configured to receive external audio signals when the device 800 is in an operational mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signals may be further stored in the memory 804 or transmitted via the communication component 816. In some embodiments, audio component 810 further includes a speaker for outputting audio signals.
The I/O interface 812 provides an interface between the processing component 802 and peripheral interface modules, which may be a keyboard, click wheel, buttons, etc. These buttons may include, but are not limited to: homepage button, volume button, start button, and lock button.
The sensor assembly 814 includes one or more sensors for providing status assessment of various aspects of the apparatus 800. For example, the sensor assembly 814 may detect an on/off state of the device 800, a relative positioning of the components, such as a display and keypad of the apparatus 800, the sensor assembly 814 may also detect a change in position of the apparatus 800 or one component of the apparatus 800, the presence or absence of user contact with the apparatus 800, an orientation or acceleration/deceleration of the apparatus 800, and a change in temperature of the apparatus 800. The sensor assembly 814 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. The sensor assembly 814 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 814 may also include an acceleration sensor, a gyroscopic sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.
The communication component 816 is configured to facilitate communication between the apparatus 800 and other devices, either in a wired or wireless manner. The device 800 may access a wireless network based on a communication standard, such as WiFi,2G or 3G, or a combination thereof. In one exemplary embodiment, the communication part 816 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, the communication component 816 further includes a Near Field Communication (NFC) module to facilitate short range communications. For example, the NFC module may be implemented based on Radio Frequency Identification (RFID) technology, infrared data association (IrDA) technology, ultra Wideband (UWB) technology, bluetooth (BT) technology, and other technologies.
In an exemplary embodiment, the apparatus 800 may be implemented by one or more Application Specific Integrated Circuits (ASICs), digital Signal Processors (DSPs), digital Signal Processing Devices (DSPDs), programmable Logic Devices (PLDs), field Programmable Gate Arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic elements for executing the methods described above.
In an exemplary embodiment, a non-transitory computer readable storage medium is also provided, such as memory 804 including instructions executable by processor 820 of apparatus 800 to perform the above-described key-miss-touch error correction method. For example, the non-transitory computer readable storage medium may be ROM, random Access Memory (RAM), CD-ROM, magnetic tape, floppy disk, optical data storage device, etc.
A non-transitory computer readable storage medium, which when executed by a processor of a mobile terminal, causes the mobile terminal to perform a key false touch error correction method, the method comprising: in the process of inputting by a user, obtaining pressing information when each key is triggered; determining false triggering keys according to the obtained pressing information; correcting errors of the false triggering keys; and determining each candidate word corresponding to the corrected complete input string.
Fig. 7 is a schematic structural diagram of a server according to an embodiment of the present invention. The server 1900 may vary considerably in configuration or performance and may include one or more central processing units (Central Processing Units, CPUs) 1922 (e.g., one or more processors) and memory 1932, one or more storage mediums 1930 (e.g., one or more mass storage devices) that store applications 1942 or data 1944. Wherein the memory 1932 and storage medium 1930 may be transitory or persistent. The program stored in the storage medium 1930 may include one or more modules (not shown), each of which may include a series of instruction operations on a server. Still further, a central processor 1922 may be provided in communication with a storage medium 1930 to execute a series of instruction operations in the storage medium 1930 on the server 1900.
The server 1900 may also include one or more power supplies 1926, one or more wired or wireless network interfaces 1950, one or more input/output interfaces 1958, one or more keyboards 1956, and/or one or more operating systems 1941, such as Windows Server, mac OS XTM, unixTM, linuxTM, freeBSDTM, and the like.
A non-transitory computer readable storage medium, which when executed by a processor of an apparatus, causes the apparatus to perform the above-described key-press error correction method.
Other embodiments of the invention will be apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention following, in general, the principles of the invention and including such departures from the present disclosure as come within known or customary practice within the art to which the invention pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the invention being indicated by the following claims.
It is to be understood that the invention is not limited to the precise arrangements and instrumentalities shown in the drawings, which have been described above, and that various modifications and changes may be effected without departing from the scope thereof. The scope of the invention is limited only by the appended claims.
The foregoing description of the preferred embodiments of the invention is not intended to limit the invention to the precise form disclosed, and any such modifications, equivalents, and alternatives falling within the spirit and scope of the invention are intended to be included within the scope of the invention.
Claims (10)
1. A method for automatically generating a dialog, for application to a dialog simulator, the method comprising:
determining at least one simulation user, dialogue field and dialogue intention of the dialogue, and generating a dialogue target of a semantic layer;
Determining personalized information of the at least one simulated user; different simulated users correspond to different personalized information;
Acquiring a pre-established dialogue action script corresponding to the dialogue field and the dialogue intention; the dialogue action script comprises a series of script instructions corresponding to the dialogue target, wherein the script instructions are used for simulating text generation instructions generated by a series of dialogue actions of a user; each row of script instructions contained in the dialogue action script is used for representing a primary dialogue action triggered based on one dialogue intention of a simulated user, and script instructions of different rows in the dialogue action script correspond to different dialogue actions;
Extracting a script instruction from the dialogue action script, and filling slot position information in the script instruction according to the dialogue target to obtain a target script instruction of the dialogue;
Generating a dialogue text according to the target script instruction of the dialogue and the personalized information;
and outputting the dialogue text as a simulation result for simulating the dialogue input of the real user, wherein the simulation result is at least used for being sent to an information recommendation system to conduct information recommendation.
2. The method according to claim 1, wherein the method further comprises: pre-establishing a template library;
The generating the dialogue text according to the target script instruction of the dialogue and the personalized information comprises the following steps:
searching the template library according to the target script instruction of the current dialogue and the personalized information to obtain a dialogue template;
and filling the dialogue template according to the slot position information in the target script instruction of the dialogue, so as to obtain dialogue texts.
3. The method according to claim 1, wherein the method further comprises: pre-establishing a dialogue generating model;
The generating the dialogue text according to the target script instruction of the dialogue and the personalized information comprises the following steps:
and inputting the information in the target script instruction of the current dialogue and the personalized information into the dialogue generating model to obtain a dialogue text.
4. The method according to claim 1, wherein the method further comprises: a knowledge base containing historical input information and recommended dialogue content is established in advance;
The generating the dialogue text according to the target script instruction of the dialogue and the personalized information comprises the following steps:
Searching the knowledge base according to the target script instruction of the current dialogue and the personalized information to obtain recommended dialogue content;
And taking the recommended dialogue content as dialogue text.
5. The method according to claim 1, wherein the method further comprises: a knowledge base containing historical input information and recommended dialogue contents and a dialogue conversion model are established in advance;
the generating a dialogue text according to the target script instruction, the personalized information and the dialogue target comprises:
retrieving the knowledge base according to the target script instruction to obtain recommended dialogue content;
And inputting the personalized information and the recommended dialogue content into the dialogue conversion model to obtain dialogue texts.
6. An information recommendation effect detection method, characterized in that the method comprises:
Simulating user input behaviors by using a pre-established dialogue behavior script to automatically generate a series of dialogue information, and inputting the dialogue information into an information recommendation system; the dialog information being generated on the basis of the method of any of claims 1-5;
If the recommended information of the information recommending system aiming at the dialogue information is received, inputting the dialogue information and the recommended information into a pre-established correlation judging model, and obtaining a correlation detection result according to the output of the correlation judging model;
and recording the correlation detection result.
7. An apparatus for automatically generating a dialog, for application to a dialog simulator, the apparatus comprising:
the setting module is used for determining at least one simulation user, dialogue field and dialogue intention of the dialogue;
The dialogue target generation module is used for generating a dialogue target of a semantic level;
The personalized information determining module is used for determining personalized information of the at least one simulation user; different simulated users correspond to different personalized information;
The script acquisition module is used for acquiring a pre-established dialogue action script corresponding to the dialogue field and the dialogue intention; the dialogue action script comprises a series of script instructions corresponding to the dialogue target, wherein the script instructions are used for simulating text generation instructions generated by a series of dialogue actions of a user; each row of script instructions contained in the dialogue action script is used for representing a primary dialogue action triggered based on one dialogue intention of a simulated user, and script instructions of different rows in the dialogue action script correspond to different dialogue actions;
The script instruction generation module is used for extracting script instructions from the dialogue action scripts and filling slot position information in the script instructions according to the dialogue targets to obtain script instructions of the dialogue;
The dialogue generating module is used for generating dialogue texts according to the script instructions of the dialogue and the personalized information;
And the output module is used for outputting the dialogue text as a simulation result for simulating the dialogue input of the real user, and the simulation result is at least used for sending the simulation result to the information recommendation system to conduct information recommendation.
8. An information recommendation effect detection apparatus, characterized in that the apparatus comprises:
A dialogue simulator for simulating user input behaviors by using a pre-established dialogue behavior script to automatically generate a series of dialogue information and inputting the dialogue information into an information recommendation system; the dialog information being generated on the basis of the method of any of claims 1-5;
the detection module is used for inputting the dialogue information and the recommended information into a pre-established correlation judgment model when receiving the recommended information of the information recommendation system aiming at the dialogue information, and obtaining a correlation detection result according to the output of the correlation judgment model;
and the recording module is used for recording the correlation detection result.
9. A computer device, comprising: one or more processors, memory;
the memory is for storing computer executable instructions and the processor is for executing the computer executable instructions to implement the method of any one of claims 1 to 5.
10. A readable storage medium having stored thereon instructions executable to implement the method of any of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911298133.0A CN112988956B (en) | 2019-12-17 | 2019-12-17 | Method and device for automatically generating dialogue, and method and device for detecting information recommendation effect |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911298133.0A CN112988956B (en) | 2019-12-17 | 2019-12-17 | Method and device for automatically generating dialogue, and method and device for detecting information recommendation effect |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112988956A CN112988956A (en) | 2021-06-18 |
| CN112988956B true CN112988956B (en) | 2024-07-12 |
Family
ID=76341825
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911298133.0A Active CN112988956B (en) | 2019-12-17 | 2019-12-17 | Method and device for automatically generating dialogue, and method and device for detecting information recommendation effect |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112988956B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114118101B (en) * | 2021-11-26 | 2022-12-09 | 北京百度网讯科技有限公司 | Dialogue data generation method and device, equipment and medium |
| CN115860013B (en) * | 2023-03-03 | 2023-06-02 | 深圳市人马互动科技有限公司 | Dialogue message processing method, device, system, equipment and medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106844627A (en) * | 2017-01-20 | 2017-06-13 | 竹间智能科技（上海）有限公司 | On-line study method and device based on conversational system |
| CN109271498A (en) * | 2018-09-14 | 2019-01-25 | 南京七奇智能科技有限公司 | The natural language interaction method and system of Virtual robot |
Family Cites Families (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104836720B (en) * | 2014-02-12 | 2022-02-25 | 北京三星通信技术研究有限公司 | Method and device for information recommendation in interactive communication |
| CN106991094A (en) * | 2016-01-21 | 2017-07-28 | 何钰威 | Foreigh-language oral-speech is spoken learning system, method and computer program |
| WO2018036555A1 (en) * | 2016-08-25 | 2018-03-01 | 腾讯科技（深圳）有限公司 | Session processing method and apparatus |
| US10572801B2 (en) * | 2017-11-22 | 2020-02-25 | Clinc, Inc. | System and method for implementing an artificially intelligent virtual assistant using machine learning |
| CN108829797A (en) * | 2018-04-25 | 2018-11-16 | 苏州思必驰信息科技有限公司 | Multiple agent dialog strategy system constituting method and adaptive approach |
| CN109471798A (en) * | 2018-10-25 | 2019-03-15 | 杭州任你说智能科技有限公司 | A kind of automatization test system and its test method of musical technique module |
| CN109857848A (en) * | 2019-01-18 | 2019-06-07 | 深圳壹账通智能科技有限公司 | Interaction content generation method, device, computer equipment and storage medium |
| CN110134775B (en) * | 2019-05-10 | 2021-08-24 | 中国联合网络通信集团有限公司 | Question and answer data generation method and device and storage medium |
| CN110188177A (en) * | 2019-05-28 | 2019-08-30 | 北京搜狗科技发展有限公司 | Talk with generation method and device |
| CN110362651A (en) * | 2019-06-11 | 2019-10-22 | 华南师范大学 | Dialogue method, system, device and the storage medium that retrieval and generation combine |
-
2019
- 2019-12-17 CN CN201911298133.0A patent/CN112988956B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106844627A (en) * | 2017-01-20 | 2017-06-13 | 竹间智能科技（上海）有限公司 | On-line study method and device based on conversational system |
| CN109271498A (en) * | 2018-09-14 | 2019-01-25 | 南京七奇智能科技有限公司 | The natural language interaction method and system of Virtual robot |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112988956A (en) | 2021-06-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110634483B (en) | Man-machine interaction method and device, electronic equipment and storage medium | |
| TWI720062B (en) | Voice input method, device and terminal equipment | |
| CN106098060B (en) | Method and device for error correction processing of voice | |
| CN107666536B (en) | Method and device for searching terminal | |
| CN112068711A (en) | Information recommendation method and device of input method and electronic equipment | |
| CN112988956B (en) | Method and device for automatically generating dialogue, and method and device for detecting information recommendation effect | |
| CN112291614A (en) | Video generation method and device | |
| CN110728981A (en) | Interactive function execution method and device, electronic equipment and storage medium | |
| CN112631435A (en) | Input method, device, equipment and storage medium | |
| CN113936697B (en) | Voice processing method and device for voice processing | |
| CN111629270A (en) | Candidate item determination method and device and machine-readable medium | |
| CN110110046B (en) | Method and device for recommending entities with same name | |
| CN110213062B (en) | Method and device for processing message | |
| US20230067387A1 (en) | Method for music generation, electronic device, storage medium cross reference to related applications | |
| CN111984767A (en) | Information recommendation method and device and electronic equipment | |
| CN113127613B (en) | Chat information processing method and device | |
| CN111831132A (en) | Information recommendation method and device and electronic equipment | |
| CN113923517B (en) | Background music generation method and device and electronic equipment | |
| CN112837668B (en) | Voice processing method and device for processing voice | |
| CN108874172A (en) | input method and device | |
| CN109558017B (en) | Input method and device and electronic equipment | |
| CN109388250B (en) | Information recommendation method and device and electronic equipment | |
| CN112445451A (en) | Music playing method and device and electronic equipment | |
| WO2017128884A1 (en) | Key processing method and apparatus, and apparatus for key processing | |
| CN110838291B (en) | Input method and device and electronic equipment |
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 |