CN111429895A - Semantic understanding method and device for multi-round interaction and computer storage medium - Google Patents
Semantic understanding method and device for multi-round interaction and computer storage medium Download PDFInfo
- Publication number
- CN111429895A CN111429895A CN201811572179.2A CN201811572179A CN111429895A CN 111429895 A CN111429895 A CN 111429895A CN 201811572179 A CN201811572179 A CN 201811572179A CN 111429895 A CN111429895 A CN 111429895A
- Authority
- CN
- China
- Prior art keywords
- voice information
- association
- round
- preset
- condition
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 230000003993 interaction Effects 0.000 title claims abstract description 38
- 230000000875 corresponding effect Effects 0.000 claims description 15
- 238000004590 computer program Methods 0.000 claims description 9
- 230000002596 correlated effect Effects 0.000 claims description 6
- 238000010586 diagram Methods 0.000 description 8
- 241000282414 Homo sapiens Species 0.000 description 7
- 230000008569 process Effects 0.000 description 7
- 230000006854 communication Effects 0.000 description 5
- 230000002452 interceptive effect Effects 0.000 description 5
- 238000004891 communication Methods 0.000 description 4
- 235000015277 pork Nutrition 0.000 description 4
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 230000001419 dependent effect Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 241000251468 Actinopterygii Species 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000010411 cooking Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000000796 flavoring agent Substances 0.000 description 1
- 235000019634 flavors Nutrition 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/08—Speech classification or search
- G10L15/18—Speech classification or search using natural language modelling
- G10L15/1822—Parsing for meaning understanding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/08—Speech classification or search
- G10L15/18—Speech classification or search using natural language modelling
- G10L15/183—Speech classification or search using natural language modelling using context dependencies, e.g. language models
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Human Computer Interaction (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Machine Translation (AREA)
Abstract
The application discloses a semantic understanding method, a semantic understanding device and a computer storage medium for multi-round interaction, wherein the semantic understanding method comprises the following steps: acquiring voice information of the current round; analyzing the voice information of the current round according to at least two preset rules to determine the association condition of the voice information of the current round and the voice information of the historical round; judging whether the correlation condition meets a preset condition or not; and responding to a judgment result that the association condition meets a preset condition, and analyzing the voice information of the current round according to the semantic understanding data of the voice information of the historical round so as to obtain the semantic understanding data of the voice information of the current round. The method and the device can accurately judge whether the front round and the back round are related in the multi-round interaction so as to realize accurate understanding of voice.
Description
Technical Field
The present application relates to the field of semantic understanding, and in particular, to a semantic understanding method, a semantic understanding apparatus, and a computer storage medium for multi-round interaction.
Background
Nowadays, in the intelligent development of human living facilities, living facilities similar to intelligent robots can interact with human beings to make human living activities more intelligent. While human-computer interaction involves voice conversation, semantic understanding becomes an important part in the conversation process.
Especially for the case of multiple rounds of interaction, how to realize accurate understanding in more spoken dialog becomes a difficult point of current research.
Disclosure of Invention
The application provides a semantic understanding method, a semantic understanding device and a computer storage medium for multi-round interaction, which aim to solve the problem that accurate understanding of voice in multi-round voice interaction cannot be realized in the prior art.
In order to solve the technical problem, the application provides a semantic understanding method for multi-round interaction, which comprises the steps of obtaining voice information of the current round; analyzing the voice information of the current round according to at least two preset rules to determine the association condition of the voice information of the current round and the voice information of the historical round; judging whether the correlation condition meets a preset condition or not; and responding to a judgment result that the association condition meets a preset condition, and analyzing the voice information of the current round according to the semantic understanding data of the voice information of the historical round so as to obtain the semantic understanding data of the voice information of the current round.
In order to solve the technical problem, the present application provides a semantic understanding apparatus for multi-round interaction, which includes a processor and a memory, where the memory stores a computer program, and the processor is configured to execute the computer program to implement the semantic understanding method.
In order to solve the above technical problem, the present application provides a computer storage medium in which a computer program is stored, and the computer program realizes the above semantic understanding method when executed.
According to the multi-turn interactive semantic understanding method, semantic understanding is conducted on each turn, and when semantic understanding is conducted on the voice information of the current turn, the voice information of the current turn is analyzed through at least two preset rules, so that the association condition of the voice information of the current turn and the voice information of the historical turn can be more accurately determined; and then, when the correlation condition meets a preset condition, analyzing the current turn voice information according to the semantic understanding data of the historical turn voice information, thereby accurately obtaining the semantic understanding data of the current turn voice information.
Drawings
FIG. 1 is a flow chart diagram illustrating an embodiment of a semantic understanding method for multi-round interaction according to the present application;
FIG. 2 is a flow chart of another embodiment of the semantic understanding method for multi-round interaction according to the present application;
FIG. 3 is a flow chart diagram of a semantic understanding method of the present application for multiple rounds of interaction according to yet another embodiment;
FIG. 4 is a schematic structural diagram of an embodiment of the multi-turn interactive system of the present application;
FIG. 5 is a schematic structural diagram of an embodiment of a semantic understanding apparatus for multi-round interaction according to the present application;
FIG. 6 is a schematic structural diagram of an embodiment of a computer storage medium according to the present application.
Detailed Description
In order to make those skilled in the art better understand the technical solution of the present invention, the following describes a semantic understanding method, a semantic understanding apparatus, and a computer storage medium for multiple rounds of interaction provided by the present application in further detail with reference to the accompanying drawings and detailed description.
The application discloses a semantic understanding method of multi-round interaction, belongs to the field of natural language processing, communication between researchers and a computer can be carried out through natural language, task-driven multi-round interaction is researched, and the intention of a user voice is determined through the multi-round interaction, so that the intention of the user is responded or executed. When multiple rounds of interaction are carried out, the user can continue the previous topic and also can change the topic, and for the living intelligent equipment, the user can carry out spoken voice interaction; for this situation, an accurate judgment mechanism is needed to accurately determine the correlation between the current turn speech and the historical turn speech of the user, so that semantic understanding can be more intelligently realized.
Referring to fig. 1 specifically, fig. 1 is a schematic flowchart of an embodiment of a semantic understanding method for multi-round interaction according to the present application, where the semantic understanding method of the present embodiment includes the following steps.
S101: and acquiring the voice information of the current turn.
For a computer performing semantic understanding, current round voice information can be obtained through a voice sensor of the computer, such as a microphone, and the computer can also be in communication connection with other devices, and the current round voice information can be obtained through voice sensors of the other devices. In addition, the computer generally only acquires the current voice information, but does not determine the round information; however, in order to conveniently describe the relationship between the current round and the historical round in the multi-round interaction, the concept of the current round is introduced in the application.
Because the relation between the current round voice information and the historical round voice information is analyzed in the subsequent steps, the current round is at least the second round in the multi-round interaction, and the voice information of the first round can be directly understood semantically.
S102: and analyzing the current turn voice information according to at least two preset rules to determine the association condition of the current turn voice information and the historical turn voice information.
In a multi-round interaction, the embodiment analyzes the voice information of the current round by adopting at least two preset rules, so as to determine the association relationship between the voice information of the current round and the voice information of the historical round. The semantic understanding method is more suitable for human voice interaction with multiple conversation characteristics, can accurately judge whether the current voice is associated with the previous voice or not, and can judge whether the current voice and the previous voice belong to the same topic or not to form multiple rounds of interaction of the same topic.
The preset rule relates to human dialogue features, and thus the preset rule may be a pronoun related rule, an information integrity related rule, a grammar accuracy related rule or an interval time related rule.
Wherein, the indication pronoun related rule is: whether or not "this", "that", etc. appears indicates a pronoun. Analyzing the current round of voice information accordingly may include: when the indication pronouns appear in the current round of voice information, the indication pronouns are related to the historical round of voice information.
The information integrity correlation rule is as follows: whether the semantic slot in semantic understanding can be completely filled. Analyzing the current round of voice information accordingly may include: when the information in the current round of voice information is incomplete, the information is related to the historical round of voice information.
The grammar accuracy related rules are: whether the grammar is accurate or its accuracy. Analyzing the current round of voice information accordingly may include: when the grammar of the current round voice information is inaccurate, the current round voice information is related to the historical round voice information.
The interval time correlation rule is as follows: whether the time interval between the current round of voice information and the last round of voice information exceeds a threshold value. Analyzing the current round of voice information accordingly may include: when the time interval does not exceed the threshold, it is indicated that it is associated with the historical round of voice information.
The preset rule may be set additionally according to the dialog feature, in addition to the above-mentioned ones, and is not limited herein. Further, in this embodiment, at least two preset rules for analyzing the current turn of voice information are both associated with the dialog feature of the preset application domain, that is, the relevant preset rules can be set according to the dialog feature of the preset application domain. For example, if the preset application field is applied to the life field, the dialogue features in the field are spoken, and usually refer to the indication pronouns or omit the dialogue mode of the information, so that the indication pronouns related rule and the information integrity related rule are preferentially adopted for analysis; if the method is applied to the working field, the conversation characteristics in the field are rigorous, the grammar of the conversation process is accurate, and the real-time performance of the conversation is strong; the analysis is preferably performed using syntax-dependent rules and interval-time-dependent rules.
In the embodiment, at least two preset rules are used for analysis, so that multi-dimensional accurate analysis is realized; from the application angle of the intelligent device, different preset rules are set aiming at the application field of the intelligent device, so that the intelligent device is more suitable for conversation in the field, and more accurate semantic analysis is realized.
S103: and judging whether the association condition meets a preset condition or not.
After determining the association condition of the current round voice information and the historical round voice information in the steps, different semantic understanding methods can be applied to different association conditions. Therefore, in step S103, it is determined whether the association condition meets a preset condition, and if the association condition meets the preset condition, the process proceeds to step S104; if the association does not meet the preset condition, the process proceeds to step S105.
The association condition meets the preset condition, namely the current round voice information and the historical round voice information are associated with each other or the association degree is high, so that the current round voice information can be understood by continuously understanding the historical round voice information. If the association condition does not meet the preset condition, the current round voice information and the historical round voice information are not associated with each other or the association degree is low, so that the current round voice information is independently understood again.
The case of the current round voice information and the history round voice information can be understood by the following example of the dialogue.
Asking 1: how is today Shenzhen weather?
Answering 1: the weather is good.
Question 2: harbinge?
Answering 2: heavy snow.
Question 3: recommend a Sichuan dish bar to me?
Answering 3: shredded pork with fish flavor
Question 4: how is this to be done?
Answering 4: … … (Menu providing fish-flavored shredded pork)
For the above dialog, "question 2" is used as the current turn of voice information, and it can be judged that "question 2" is related to "question 1" through the information integrity correlation rule, so that according to the semantic understanding of "question 1", it can be understood that "question 2" is asking for weather, and "answer 2" also answers weather-related information.
The question 3 is used as the current turn voice information, and the question 3 is judged to be irrelevant to the question 2 through at least two preset rules, so that the question 3 is understood again.
The question 4 is used as the current turn voice information, and the relation between the question 4 and the question 3 can be judged through indicating the related rules of pronouns, so that according to the semantic understanding of the question 3, the question 4 can be understood to inquire about how the fish-flavored shredded pork is to be done, and the answer 4 can provide a menu of the fish-flavored shredded pork.
S104: and analyzing the current round voice information according to the semantic understanding data of the historical round voice information to obtain the semantic understanding data of the current round voice information.
After the current round voice information and the historical round voice information are determined to be correlated, the current round voice information can be analyzed according to the semantic understanding data of the historical round voice information, and therefore the semantic understanding data of the current round voice information can be obtained.
Semantic understanding data, namely data generated when speech information is understood, field division is generally performed firstly when semantic understanding is performed, and the definition of the field can be a general field or a cooking field; then, performing intention analysis, and determining the intention by using intention trees of different fields; and after the intention is determined, determining the semantic slot corresponding to the intention. Accordingly, semantic understanding data generally includes domain data, intent data, and semantic slot data. When the voice information of the current round is understood, the missing information can be directly understood by using the semantic understanding data of the historical round without being determined to the user in a polling mode, so that the communication process is smoother, and the user experience is better.
The semantic understanding data of the voice information of the historical round can be the semantic understanding data of the voice information of the previous round or the semantic understanding data of the voice information of the previous rounds.
S105: and clearing the semantic understanding data of the voice information of the historical round, and analyzing the voice information of the current round to obtain the semantic understanding data of the voice information of the current round.
After the fact that the current round voice information is not correlated with the historical round voice information is determined, namely the current round voice information belongs to the other conversation field and is irrelevant to the historical round, semantic understanding needs to be conducted again, therefore, semantic understanding data of the historical round voice information can be eliminated, the current round voice information is analyzed, namely the field information of the current round voice information is determined again, and semantic understanding including intention understanding and filling of semantic slots is conducted on the current round voice information based on the field information.
In this step, the semantic understanding data of the voice information of the historical round is cleared, so that the semantic understanding data of the voice information of the current round is not influenced.
In the embodiment, the current round voice information is analyzed through at least two preset rules, the accurate judgment of the association condition of the current round voice information and the historical round voice information is realized from multiple dimensions, and then the current round voice information is analyzed by determining whether semantic understanding data of the historical round voice information is used or not according to different association conditions so as to accurately understand the current round voice information, so that the whole multi-round interaction process is more intelligent and is more suitable for natural conversation of human beings.
Based on the embodiment shown in fig. 1, there are various ways to apply at least two preset rules, such as the following embodiments.
Referring to fig. 2, fig. 2 is a schematic flowchart of another embodiment of a semantic understanding method for multi-round interaction according to the present application, and the semantic understanding method of the present embodiment includes the following steps.
S201: and acquiring the voice information of the current turn.
Step S201 is similar to step S101, and details are not repeated.
In this embodiment, analyzing the voice information of the current round according to at least two preset rules mainly includes sequentially analyzing and judging, specifically, judging by using a preset rule of a first priority, and if it is judged that the association condition does not meet a preset condition, judging by using a preset rule of a second priority; if the association condition is judged to meet the preset condition, finishing the judgment; and analyzing and judging in sequence from high to low in the priority of at least two preset rules until the process of analyzing and judging in sequence is finished. The method comprises the following specific steps.
S202: and analyzing the voice information of the current turn by using a preset rule to obtain the association condition corresponding to the preset rule.
In this embodiment, the step S202 is repeated for a plurality of times, and each time of execution only uses a single preset rule to analyze the current turn of voice information, so as to obtain the association condition corresponding to the preset rule. The setting of the preset rule, the analysis of the voice information, that is, the determination of the association condition are all similar to step S102, and details are not repeated. It should be noted that, as shown in step S102, the priorities of the at least two preset rules in the present embodiment can also be set according to the dialog features of the application field.
S203: and judging whether the association condition meets a preset condition or not.
In step S202, a single preset rule is used to analyze the current round of voice information, and after determining the association condition, it is determined whether the association condition meets a preset condition, if the association condition meets the preset condition, step S206 is performed, i.e. semantic understanding is directly performed, and the analysis and determination are performed without using the subsequent preset rule; if the association does not meet the preset condition, step S204 is performed.
In step S202, analyzing the current round of voice information by a single preset rule to determine a correlation condition; it is generally determined whether the current round of voice information is correlated with the historical round of voice information. In step S203, it is determined whether the association condition meets a preset condition, that is, whether the association condition is related to each other is determined; and the association condition is that the association corresponds to the association condition meeting the preset condition, and the association condition is that the association does not correspond to the association condition meeting the preset condition.
S204: and judging whether the preset rule is the preset rule with the lowest priority.
When the preset rule is used to judge that the association condition is not in accordance with the preset condition, judging whether the preset rule is the preset rule with the lowest priority, and if the preset rule is not the preset rule with the lowest priority, re-performing the step S202 according to the preset rule with the next priority; if the preset rule is the preset rule with the lowest priority, it is determined that the association condition does not meet the preset condition, and step S205 is performed.
S205: and clearing the semantic understanding data of the voice information of the historical round, and analyzing the voice information of the current round to obtain the semantic understanding data of the voice information of the current round.
S206: and analyzing the current round voice information according to the semantic understanding data of the historical round voice information to obtain the semantic understanding data of the current round voice information.
The steps S205 to S206 are similar to the steps S104 to S105, and detailed description thereof is omitted.
The embodiment analyzes the voice information of the current turn by using the preset rules in the order of the priority from high to low, thereby determining the association condition. Furthermore, in this embodiment, the priority of the preset rule may also be determined according to the dialog feature of the application field, that is, the preset rule most related to the dialog feature is preferentially adopted to analyze the current round of voice information, so that semantic understanding of the dialog in the application field can be performed more accurately.
Referring to fig. 3, fig. 3 is a schematic flowchart of a semantic understanding method according to another embodiment of the present application, where the semantic understanding method includes the following steps.
S301: and acquiring the voice information of the current turn.
Step S301 is similar to step S101, and detailed description thereof is omitted.
In this embodiment, at least two preset rules are used to comprehensively analyze the current round of voice information, so as to obtain the association degree between the current round of voice information and the historical round of voice information. The method comprises the following specific steps.
S302: and respectively analyzing the current round voice information by utilizing at least two preset rules to obtain at least two association scores corresponding to each preset rule.
In step S302, a preset rule is used to analyze the current round of voice information to obtain an association score, and the preset rule is no longer a simple judgment, but relates to the analysis of metrics. For example, in the interval time correlation rule, it can be analyzed to which interval time period the interval time belongs, and different interval time periods correspond to different association scores.
S303: and calculating the association degree of the current round voice information and the historical round voice information by combining at least two association scores and the weight of each association score.
After at least two relevance scores corresponding to each preset rule are obtained, the relevance scores are comprehensively solved to obtain a relevance degree, and the solving adopts a mode of combining the relevance scores with weights. The weight of the association score is positively correlated with the priority of the preset rule corresponding to the association score, and the priority of the preset rule can be set according to the conversation characteristics of the applied field.
The higher the priority associated with the conversation feature, the higher the priority and the higher the weight of the corresponding association score, and thus the calculated association degree can accurately reflect the human natural conversation situation of the application field.
S304: and judging whether the association degree exceeds an association degree threshold value.
After obtaining the association degree, determining the association degree by using an association degree threshold, and if the association degree exceeds the association degree threshold and corresponds to the association condition meeting the preset condition, performing step S305; if the correlation degree does not exceed the correlation degree threshold value corresponding to the correlation condition not meeting the preset condition, go to step S306.
S305: and analyzing the current round voice information according to the semantic understanding data of the historical round voice information to obtain the semantic understanding data of the current round voice information.
S306: removing semantic understanding data of historical turn voice information, and analyzing current turn voice information to obtain semantic understanding data of current turn voice information
The steps S305 to S306 are similar to the steps S104 to S105, and detailed description thereof is omitted.
In the embodiment, each preset rule is utilized to analyze to obtain a plurality of corresponding association scores, and the association degree is obtained by combining the weight calculation of the association scores for judgment, wherein the weight of the association score is positively correlated with the priority of the corresponding preset rule, and the priority is determined by the conversation characteristics of the application field, so that the accurate understanding of the conversation in the application field can be realized.
Based on the semantic understanding method of the multi-round interaction, a multi-round interaction system can be constructed. Referring to fig. 4, fig. 4 is a schematic structural diagram of an embodiment of a multi-turn interactive system according to the present application, where the multi-turn interactive system 100 of the present embodiment includes a voice recognition module 11, a semantic understanding module 12, a dialog management module 13, a language generation module 14, a voice broadcast module 15, and a command execution module 16.
The voice recognition module (ASR)11 converts voice information into text information, the text information is transmitted to a semantic understanding module (N L U)12 to be understood, when multiple rounds of conversations occur, a conversation management module (DM)13 is used for determining the incidence relation between the voice information of the current round and the voice information of the historical round, the conversation management module 13 determines the incidence relation by the method, then the semantic understanding module (N L U)12 is used for understanding to determine semantic understanding data of the voice information of the current round, the conversation management module (DM)13 determines reply content or an execution instruction by the semantic understanding data, for the reply content, voice reply is achieved through a language generation module (N L G)14 and a voice broadcasting module (TTS)15, and for the execution instruction, the execution instruction is executed through an instruction execution module 16.
The multi-round interactive system can accurately understand the user language and realize high fluency of human-computer voice interaction.
The method is applied to hardware equipment, and can realize multi-round voice interaction. Referring to fig. 5 in detail, fig. 5 is a schematic structural diagram of an embodiment of a semantic understanding apparatus for multi-round interaction according to the present application, where the semantic understanding apparatus 200 of the present embodiment includes a processor 21 and a memory 22. Wherein, the memory 22 stores a computer program, and the processor 21 is configured to execute the computer program to implement the semantic understanding method of the above-mentioned multiple rounds of interaction.
The processor 21 is configured to obtain voice information of a current turn; analyzing the voice information of the current round according to at least two preset rules preset in the memory 22 to determine the association condition between the voice information of the current round and the voice information of the historical round; judging whether the correlation condition meets a preset condition or not; in response to the determination result that the association condition meets the preset condition, the current round voice information is analyzed according to the semantic understanding data of the historical round voice information in the memory 22 to obtain the semantic understanding data of the current round voice information.
The processor 21 may be an integrated circuit chip having signal processing capability. The processor 21 may also be a general purpose processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The semantic understanding apparatus 200 may be an intelligent home appliance, and implements an intelligent conversation in home life, and the rule preset in the corresponding home appliance is determined according to the conversation feature of the home domain to which the rule is applied. The semantic understanding apparatus 200 may also be a server, and the intelligent home appliance is connected to the server, and combines the functions of the server to implement multiple rounds of voice interaction.
Referring to fig. 6, fig. 6 is a schematic structural diagram of an embodiment of a computer storage medium according to the present application, where the computer storage medium 300 of the present embodiment includes a computer program 31, which can be executed to implement the method in the foregoing embodiment.
The computer storage medium 300 of this embodiment may be a medium that can store program instructions, such as a usb disk, a removable hard disk, a Read-only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk, or may also be a server that stores the program instructions, and the server may send the stored program instructions to other devices for operation, or may self-operate the stored program instructions.
In the several embodiments provided in the present application, it should be understood that the disclosed method and apparatus may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, a division of a module or a unit is merely a logical division, and an actual implementation may have another division, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
Units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.
In addition, functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.
The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application may be substantially implemented or contributed to by the prior art, or all or part of the technical solution may be embodied in a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, a network device, or the like) or a processor (processor) to execute all or part of the steps of the method according to the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
The above description is only for the purpose of illustrating embodiments of the present application and is not intended to limit the scope of the present application, and all modifications of equivalent structures and equivalent processes, which are made by the contents of the specification and the drawings of the present application or are directly or indirectly applied to other related technical fields, are also included in the scope of the present application.
Claims (10)
1. A semantic understanding method of multi-round interaction, the method comprising:
acquiring voice information of the current round;
analyzing the current round voice information according to at least two preset rules to determine the association condition of the current round voice information and the historical round voice information;
judging whether the association condition meets a preset condition or not;
and responding to a judgment result that the association condition meets the preset condition, and analyzing the current turn voice information according to the semantic understanding data of the historical turn voice information to obtain the semantic understanding data of the current turn voice information.
2. The method according to claim 1, wherein the at least two preset rules each have a priority;
analyzing the current turn voice information according to at least two preset rules to determine the association condition of the current turn voice information and the historical turn voice information; judging whether the association condition meets a preset condition or not; the method comprises the following steps:
analyzing the voice information of the current turn by sequentially utilizing each preset rule according to the sequence of the priority from high to low so as to obtain the association condition; wherein the association condition corresponds to the utilized preset rule; judging whether the association condition meets a preset condition or not;
until the association condition is judged to be in accordance with a preset condition, or until the current round voice information is analyzed by using a preset rule with the lowest priority.
3. The method according to claim 2, wherein the determining whether the association condition meets a preset condition comprises:
and judging whether the association conditions are associated with each other, wherein the association conditions are associated with each other and correspond to the association conditions meeting the preset conditions, and the association conditions are not associated with each other and correspond to the association conditions not meeting the preset conditions.
4. The method according to claim 1, wherein the analyzing the voice information of the current turn according to at least two preset rules to determine the association between the voice information of the current turn and the voice information of the historical turn comprises:
analyzing the current round voice information according to at least two preset rules to calculate the association degree of the current round voice information and the historical round voice information;
the judging whether the association condition meets a preset condition includes:
judging whether the association degree exceeds an association degree threshold, wherein the association degree exceeds the association degree threshold corresponding to that the association condition meets a preset condition, and the association degree does not exceed the association degree threshold corresponding to that the association condition does not meet the preset condition.
5. The method according to claim 4, wherein the analyzing the voice information of the current turn according to at least two preset rules to calculate the association degree between the voice information of the current turn and the voice information of the historical turn comprises:
analyzing the current round voice information by using the at least two preset rules respectively to obtain at least two association scores of the current round and the historical round corresponding to each preset rule;
calculating the association degree by combining the at least two association scores and the weight of each association score; and the weight of the association score is positively correlated with the priority of the preset rule corresponding to the association score.
6. The method according to claim 1, wherein the at least two preset rules are each associated with a dialog feature of a preset application domain.
7. The method according to claim 1, wherein the at least two preset rules comprise: at least two of a pronoun related rule, an information integrity related rule, a grammar accuracy related rule, and an interval time related rule are indicated.
8. The method according to claim 7, wherein the indication pronoun related rule has a higher priority than the information integrity related rule, the information integrity related rule has a higher priority than the grammar accuracy related rule, and the grammar accuracy related rule has a higher priority than the interval time related rule.
9. A semantic understanding apparatus for multi-round interaction, the semantic understanding apparatus comprising a processor and a memory; the memory has stored therein a computer program for execution by the processor to implement the steps of the method according to any one of claims 1-8.
10. A computer storage medium, characterized in that the computer storage medium stores a computer program which, when executed, implements the steps of the method according to any one of claims 1-8.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811572179.2A CN111429895B (en) | 2018-12-21 | 2018-12-21 | Semantic understanding method and device for multi-round interaction and computer storage medium |
| PCT/CN2019/123807 WO2020125457A1 (en) | 2018-12-21 | 2019-12-06 | Multi-round interaction semantic understanding method and apparatus, and computer storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811572179.2A CN111429895B (en) | 2018-12-21 | 2018-12-21 | Semantic understanding method and device for multi-round interaction and computer storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111429895A true CN111429895A (en) | 2020-07-17 |
| CN111429895B CN111429895B (en) | 2023-05-05 |
Family
ID=71101987
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811572179.2A Active CN111429895B (en) | 2018-12-21 | 2018-12-21 | Semantic understanding method and device for multi-round interaction and computer storage medium |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN111429895B (en) |
| WO (1) | WO2020125457A1 (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112084768A (en) * | 2020-08-06 | 2020-12-15 | 珠海格力电器股份有限公司 | Multi-round interaction method and device and storage medium |
| CN112164401A (en) * | 2020-09-18 | 2021-01-01 | 广州小鹏汽车科技有限公司 | Voice interaction method, server and computer-readable storage medium |
| CN112256229A (en) * | 2020-09-11 | 2021-01-22 | 北京三快在线科技有限公司 | Man-machine voice interaction method and device, electronic equipment and storage medium |
| CN112417107A (en) * | 2020-10-22 | 2021-02-26 | 联想(北京)有限公司 | Information processing method and device |
| CN112992137A (en) * | 2021-01-29 | 2021-06-18 | 青岛海尔科技有限公司 | Voice interaction method and device, storage medium and electronic device |
| CN113035180A (en) * | 2021-03-22 | 2021-06-25 | 建信金融科技有限责任公司 | Voice input integrity judgment method and device, electronic equipment and storage medium |
| CN113380241A (en) * | 2021-05-21 | 2021-09-10 | 珠海格力电器股份有限公司 | Semantic interaction adjusting method and device, voice equipment and storage medium |
| CN113435196A (en) * | 2021-06-22 | 2021-09-24 | 平安科技（深圳）有限公司 | Intention recognition method, device, equipment and storage medium |
| CN113555018A (en) * | 2021-07-20 | 2021-10-26 | 海信视像科技股份有限公司 | Voice interaction method and device |
| CN113673253A (en) * | 2021-08-13 | 2021-11-19 | 珠海格力电器股份有限公司 | Semantic interaction method and device and electronic equipment |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113656562A (en) * | 2020-11-27 | 2021-11-16 | 话媒(广州)科技有限公司 | Multi-round man-machine psychological interaction method and device |
| CN112632251B (en) * | 2020-12-24 | 2023-12-29 | 北京百度网讯科技有限公司 | Reply content generation method, device, equipment and storage medium |
| CN115022733B (en) * | 2022-06-17 | 2023-09-15 | 中国平安人寿保险股份有限公司 | Digest video generation method, digest video generation device, computer device and storage medium |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130218553A1 (en) * | 2012-02-16 | 2013-08-22 | Kabushiki Kaisha Toshiba | Information notification supporting device, information notification supporting method, and computer program product |
| US20150341499A1 (en) * | 2014-05-20 | 2015-11-26 | Hootsuite Media Inc. | Method and system for managing voice calls in association with social media content |
| CN107665706A (en) * | 2016-07-29 | 2018-02-06 | 科大讯飞股份有限公司 | Rapid Speech exchange method and system |
| CN107785018A (en) * | 2016-08-31 | 2018-03-09 | 科大讯飞股份有限公司 | More wheel interaction semantics understanding methods and device |
| CN108037905A (en) * | 2017-11-21 | 2018-05-15 | 北京光年无限科技有限公司 | A kind of interaction output method and intelligent robot for intelligent robot |
| CN108388638A (en) * | 2018-02-26 | 2018-08-10 | 出门问问信息科技有限公司 | Semantic analytic method, device, equipment and storage medium |
| CN108962261A (en) * | 2018-08-08 | 2018-12-07 | 联想(北京)有限公司 | Information processing method, information processing unit and bluetooth headset |
| CN108959412A (en) * | 2018-06-07 | 2018-12-07 | 出门问问信息科技有限公司 | Generation method, device, equipment and the storage medium of labeled data |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105589844B (en) * | 2015-12-18 | 2017-08-08 | 北京中科汇联科技股份有限公司 | It is a kind of to be used to take turns the method for lacking semantic supplement in question answering system more |
| CN105704013B (en) * | 2016-03-18 | 2019-04-19 | 北京光年无限科技有限公司 | Topic based on context updates data processing method and device |
| CN107799116A (en) * | 2016-08-31 | 2018-03-13 | 科大讯飞股份有限公司 | More wheel interacting parallel semantic understanding method and apparatus |
| CN106503156B (en) * | 2016-10-24 | 2019-09-03 | 北京百度网讯科技有限公司 | Man-machine interaction method and device based on artificial intelligence |
| CN106777013B (en) * | 2016-12-07 | 2020-09-11 | 科大讯飞股份有限公司 | Conversation management method and device |
-
2018
- 2018-12-21 CN CN201811572179.2A patent/CN111429895B/en active Active
-
2019
- 2019-12-06 WO PCT/CN2019/123807 patent/WO2020125457A1/en active Application Filing
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130218553A1 (en) * | 2012-02-16 | 2013-08-22 | Kabushiki Kaisha Toshiba | Information notification supporting device, information notification supporting method, and computer program product |
| US20150341499A1 (en) * | 2014-05-20 | 2015-11-26 | Hootsuite Media Inc. | Method and system for managing voice calls in association with social media content |
| CN107665706A (en) * | 2016-07-29 | 2018-02-06 | 科大讯飞股份有限公司 | Rapid Speech exchange method and system |
| CN107785018A (en) * | 2016-08-31 | 2018-03-09 | 科大讯飞股份有限公司 | More wheel interaction semantics understanding methods and device |
| CN108037905A (en) * | 2017-11-21 | 2018-05-15 | 北京光年无限科技有限公司 | A kind of interaction output method and intelligent robot for intelligent robot |
| CN108388638A (en) * | 2018-02-26 | 2018-08-10 | 出门问问信息科技有限公司 | Semantic analytic method, device, equipment and storage medium |
| CN108959412A (en) * | 2018-06-07 | 2018-12-07 | 出门问问信息科技有限公司 | Generation method, device, equipment and the storage medium of labeled data |
| CN108962261A (en) * | 2018-08-08 | 2018-12-07 | 联想(北京)有限公司 | Information processing method, information processing unit and bluetooth headset |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112084768A (en) * | 2020-08-06 | 2020-12-15 | 珠海格力电器股份有限公司 | Multi-round interaction method and device and storage medium |
| CN112256229A (en) * | 2020-09-11 | 2021-01-22 | 北京三快在线科技有限公司 | Man-machine voice interaction method and device, electronic equipment and storage medium |
| CN112256229B (en) * | 2020-09-11 | 2024-05-14 | 北京三快在线科技有限公司 | Man-machine voice interaction method and device, electronic equipment and storage medium |
| CN112164401A (en) * | 2020-09-18 | 2021-01-01 | 广州小鹏汽车科技有限公司 | Voice interaction method, server and computer-readable storage medium |
| CN112164401B (en) * | 2020-09-18 | 2022-03-18 | 广州小鹏汽车科技有限公司 | Voice interaction method, server and computer-readable storage medium |
| CN112417107A (en) * | 2020-10-22 | 2021-02-26 | 联想(北京)有限公司 | Information processing method and device |
| CN112992137A (en) * | 2021-01-29 | 2021-06-18 | 青岛海尔科技有限公司 | Voice interaction method and device, storage medium and electronic device |
| CN113035180A (en) * | 2021-03-22 | 2021-06-25 | 建信金融科技有限责任公司 | Voice input integrity judgment method and device, electronic equipment and storage medium |
| CN113380241B (en) * | 2021-05-21 | 2024-03-08 | 珠海格力电器股份有限公司 | Semantic interaction adjusting method and device, voice equipment and storage medium |
| CN113380241A (en) * | 2021-05-21 | 2021-09-10 | 珠海格力电器股份有限公司 | Semantic interaction adjusting method and device, voice equipment and storage medium |
| CN113435196A (en) * | 2021-06-22 | 2021-09-24 | 平安科技（深圳）有限公司 | Intention recognition method, device, equipment and storage medium |
| CN113435196B (en) * | 2021-06-22 | 2022-07-29 | 平安科技（深圳）有限公司 | Intention recognition method, device, equipment and storage medium |
| CN113555018A (en) * | 2021-07-20 | 2021-10-26 | 海信视像科技股份有限公司 | Voice interaction method and device |
| CN113555018B (en) * | 2021-07-20 | 2024-05-28 | 海信视像科技股份有限公司 | Voice interaction method and device |
| CN113673253B (en) * | 2021-08-13 | 2023-07-07 | 珠海格力电器股份有限公司 | Semantic interaction method and device and electronic equipment |
| CN113673253A (en) * | 2021-08-13 | 2021-11-19 | 珠海格力电器股份有限公司 | Semantic interaction method and device and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2020125457A1 (en) | 2020-06-25 |
| CN111429895B (en) | 2023-05-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111429895B (en) | Semantic understanding method and device for multi-round interaction and computer storage medium | |
| CN112365894B (en) | AI-based composite voice interaction method and device and computer equipment | |
| CN108446322B (en) | Method and device for realizing intelligent question-answering system | |
| US9583102B2 (en) | Method of controlling interactive system, method of controlling server, server, and interactive device | |
| CN108664472B (en) | Natural language processing method, device and equipment | |
| CN108153780B (en) | Man-machine conversation device and method for realizing man-machine conversation | |
| CN109460463B (en) | Model training method, device, terminal and storage medium based on data processing | |
| CN109514586B (en) | Method and system for realizing intelligent customer service robot | |
| CN110782962A (en) | Hearing language rehabilitation device, method, electronic equipment and storage medium | |
| CN105006235A (en) | Speech endpointing based on word comparisons | |
| CN106847305B (en) | Method and device for processing recording data of customer service telephone | |
| JP6677419B2 (en) | Voice interaction method and apparatus | |
| CN108962233A (en) | Voice dialogue processing method and system for voice dialogue platform | |
| CN111429899A (en) | Speech response processing method, device, equipment and medium based on artificial intelligence | |
| CN104038473A (en) | Method of audio ad insertion, device, equipment and system | |
| CN109524013B (en) | Voice processing method, device, medium and intelligent equipment | |
| CN110600008A (en) | Voice wake-up optimization method and system | |
| CN106205622A (en) | Information processing method and electronic equipment | |
| CN108897517B (en) | Information processing method and electronic equipment | |
| CN112183098A (en) | Session processing method and device, storage medium and electronic device | |
| US20120053937A1 (en) | Generalizing text content summary from speech content | |
| CN111968643A (en) | Intelligent recognition method, robot and computer readable storage medium | |
| CN112242135A (en) | Voice data processing method and intelligent customer service device | |
| CN115731915A (en) | Active dialogue method and device for dialogue robot, electronic device and storage medium | |
| CN112509567B (en) | Method, apparatus, device, storage medium and program product for processing voice data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |