CN113076408A - Session information processing method and device - Google Patents
Session information processing method and device Download PDFInfo
- Publication number
- CN113076408A CN113076408A CN202110305854.0A CN202110305854A CN113076408A CN 113076408 A CN113076408 A CN 113076408A CN 202110305854 A CN202110305854 A CN 202110305854A CN 113076408 A CN113076408 A CN 113076408A
- Authority
- CN
- China
- Prior art keywords
- vector
- determining
- conversation
- state vector
- current wheel
- 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
- 238000003672 processing method Methods 0.000 title claims description 7
- 230000010365 information processing Effects 0.000 title description 9
- 239000013598 vector Substances 0.000 claims abstract description 225
- 238000000034 method Methods 0.000 claims abstract description 32
- 230000000875 corresponding effect Effects 0.000 claims description 75
- 230000002596 correlated effect Effects 0.000 claims description 5
- 238000004458 analytical method Methods 0.000 abstract description 14
- 238000010586 diagram Methods 0.000 description 5
- 238000004590 computer program Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000007792 addition Methods 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000005215 recombination Methods 0.000 description 1
- 230000006798 recombination Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation or dialogue systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/205—Parsing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
- G06F40/35—Discourse or dialogue representation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Human Computer Interaction (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- User Interface Of Digital Computer (AREA)
- Machine Translation (AREA)
Abstract
The application discloses a method and a device for processing session information, wherein the method comprises the following steps: determining an expected vector and an actual vector corresponding to the previous pair of dialogues; determining a first state vector corresponding to the current wheel conversation according to the expected vector and the actual vector; determining a target state vector corresponding to the current wheel conversation according to the first state vector corresponding to the current wheel conversation and the conversation information of the current wheel conversation; the judgment of the dialog state of the current round and the identification of the user intention are realized in combination with the context semantics in the multi-round dialog; and thereby avoid a situation in which erroneous dialog state judgments in the above disturb the semantic analysis of the context in connection with the above, resulting in erroneous judgments of the user's intention.
Description
Technical Field
The present application relates to the field of computer technologies, and in particular, to a method and an apparatus for processing session information.
Background
With the maturity of man-machine conversation technology, the intelligent customer service system has replaced artificial customer service to a certain extent. In some cases, the intelligent customer service system may determine the user's intent in conjunction with multiple rounds of conversations with the user to address the relevant issue for the user. In other words, in multiple rounds of conversations, the intelligent customer service system needs to perform semantic analysis in relation to the context of the conversation to realize the judgment of the user's intention.
In the prior art, a state vector can be determined according to the previous dialog content, and the state vector can include the judgment of the intelligent customer service system on the user intention temporarily in the previous dialog. And then the conversation content of the previous round and the current round is combined, so that the effect of 'contacting the above' can be achieved in semantic analysis.
However, the judgment of the user's intention included in the state vector is also only a temporary analysis result, that is, the judgment is not necessarily correct. Therefore, in the prior art, the state vector is used as a basis for associating context in semantic analysis, so that misjudgment on the user intention is easy to occur, and the accuracy rate needs to be improved.
Disclosure of Invention
The application provides a method and a device for processing session information.
In a first aspect, the present application provides a method for processing session information, including:
determining an expected vector and an actual vector corresponding to the previous pair of dialogues;
determining a first state vector corresponding to the current wheel conversation according to the expected vector and the actual vector;
and determining a target state vector corresponding to the current wheel conversation according to the first state vector corresponding to the current wheel conversation and the conversation information of the current wheel conversation.
Preferably, the determining the expected vector corresponding to the previous dialog comprises:
determining expected keywords according to the historical state vector corresponding to the previous dialog;
and determining the expected vector according to the expected keyword.
Preferably, the determining the actual vector corresponding to the previous pair of dialogues includes:
and determining the actual vector according to the dialogue information of the previous dialogue.
Preferably, the determining a first state vector corresponding to the current wheel dialog according to the expected vector and the actual vector includes:
determining similarity indices of the desired vector and the actual vector;
respectively determining the weight coefficients of the expected vector and the actual vector according to the similarity indexes;
determining the first state vector according to the expected vector, the actual vector and the weight coefficient.
Preferably, the determining the weight coefficients of the desired vector and the actual vector respectively according to the similarity index includes:
determining a first weight coefficient of the expected vector according to the similarity index; the first weight coefficient is positively correlated with the similarity index;
determining a second weight coefficient of the actual vector according to the similarity index; the second weight coefficient is inversely related to the similarity index.
Preferably, before determining the target state vector corresponding to the current wheel dialog, the method further includes:
and determining a second state vector corresponding to the current round of conversation according to the conversation information of the previous round of conversation.
Preferably, the determining, according to the first state vector corresponding to the current wheel session and the session information of the current wheel session, the target state vector corresponding to the current wheel session includes:
and determining a target state vector corresponding to the current wheel conversation according to the first state vector, the second state vector and the conversation information of the current wheel conversation.
Preferably, the method further comprises the following steps:
and determining the output information corresponding to the next wheel telephone according to the target state vector corresponding to the current wheel telephone.
Preferably, the method further comprises the following steps:
and determining an expected vector corresponding to the next wheel microphone according to the output information corresponding to the next wheel microphone.
In a second aspect, the present application provides a device for processing session information, including:
a history vector determination module, configured to determine an expected vector and an actual vector corresponding to the previous pair of dialogues;
the first state vector determining module is used for determining a first state vector corresponding to the current wheel conversation according to the expected vector and the actual vector;
and the target state vector determining module is used for determining the target state vector corresponding to the current wheel conversation according to the first state vector corresponding to the current wheel conversation and the conversation information of the current wheel conversation.
Compared with the related art, the conversation information processing method and the conversation information processing device provided by the application provide a brand-new conversation information processing mode.
Drawings
Fig. 1 is a schematic flowchart illustrating a session information processing method according to an embodiment of the present application;
fig. 2 is a flowchart illustrating a processing method of session information according to another embodiment of the present application;
fig. 3 is a schematic structural diagram illustrating a session information processing apparatus according to an embodiment of the present application.
Detailed Description
In order to make the objects, features and advantages of the present application more obvious and understandable, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is apparent that the described embodiments are only a part of the embodiments of the present application, and not all the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
The intelligent customer service system determines the intention of the user by combining multiple rounds of conversations with the user, so that relevant problems are solved for the user; that is, semantic analysis can be performed in connection with the context of the dialog, so that the user's intention can be judged. In some cases, the system may determine a state vector based on the previous session content, and the state vector may include a determination that the system was temporarily interested in the user's intent during the previous session. And then the conversation content of the previous round and the current round is combined, so that the effect of 'contacting the above' can be achieved in semantic analysis.
However, the above-described determination of the intention of the user in the state vector is merely a temporary analysis result, that is, the determination is not necessarily correct. Therefore, in the prior art, the state vector is used as a basis for associating context in semantic analysis, so that misjudgment on the user intention is easy to occur, and the accuracy rate needs to be improved.
Therefore, the embodiment of the present application will provide a method for processing session information. As shown in fig. 1, the method in this embodiment includes the following steps:
and 101, determining an expected vector and an actual vector corresponding to the previous pair of dialogs.
In the scenario of multiple rounds of conversations, the specific content of the conversation may be referred to as conversation information. Typically, a question and answer from the intelligent customer service system to the customer is defined as a round of conversation. In a round of dialog, dialog information output by the system may be referred to as output information, and dialog information received by the system from a user may be referred to as input information. From the dialog information of the previous dialog, the corresponding desired vector and the actual vector can be determined.
The expected vector corresponding to the previous session represents some information that the system would otherwise expect to learn through the previous session before the previous session began. Specifically, determining the expected vector corresponding to the previous dialog comprises: determining expected keywords according to the historical state vector corresponding to the previous dialog; based on the desired keywords, a desired vector is determined.
Assume that in the middle of an earlier conversation, the user asks the system for "weather", indicating the user's "intent". The dialog states determined in conjunction with this "intent" in the semantic analysis may constitute a historical state vector. Further, the system needs to make clear in the previous dialog that the user asks for "where weather" in order to accurately meet the user's intention. Thus in the last dialog, the output information of the system may be "where you need to ask about weather" in order to reply to the user with input information about "location". That is, the expected keyword corresponding to the previous dialog is the "place"; this desired keyword may also be commonly referred to in the art as a "slot". The expected vector can be embodied by the expected keyword of 'location'.
However, the user's intent in the dialog may change, and the input information given by the user may or may not be canonical. The content of the input information actually given by the user in the previous dialog may or may not match the expectations of the system. In this step, the corresponding actual vector can be determined according to the input information actually given by the user in the previous round of conversation. The actual vector embodies the information that the user actually expressed in the previous dialog.
And 102, determining a first state vector corresponding to the current wheel dialog according to the expected vector and the actual vector.
It will be appreciated that the desired vector may or may not be strongly correlated with the actual vector. In combination with the above example, when the expected keyword of the expected vector is "location", if the actual vector contains information about "location", the correlation between the two is strong; if the information about the "location" is not included, the correlation between the two is weak.
Given the strong correlation of the desired vector with the actual vector, it is possible to account for the questions posed by the system that the user answers accurately. This can indicate to some extent that the previous determination of the dialog state by the system was correct, or that the system was accurate for the recognition of the user's intention.
On the contrary, if the correlation between the expected vector and the actual vector is weak, it may indicate that the intention of the user is changed; or the system fails to accurately recognize the user intention, so that the system and the user cannot understand each other, and the conversation is obstructed. Namely, the system is considered to have certain error in the judgment of the conversation state before.
Therefore, the dialog state after the last dialog is finished and before the current dialog is started can be confirmed again according to the expected vector and the actual vector, namely the first state vector is determined. Specifically, the similarity index of the desired vector and the actual vector may be determined; respectively determining the weight coefficients of the expected vector and the actual vector according to the similarity index; a first state vector is determined based on the desired vector, the actual vector, and the weight coefficients.
The similarity index is an index for quantitatively measuring the strength of the correlation between the expected vector and the actual vector. The calculation process of the similarity index in this embodiment is not limited, and any algorithm that can achieve the same or similar effect may be combined in this embodiment.
It is understood that the higher the similarity index, i.e. the stronger the correlation between the expected vector and the actual vector, the more accurate the system has been in the previous determination of the dialog state. The first state vector may now be determined more according to "above", i.e. the desired vector determined from the historical state vectors, to continue and strengthen the context relevance. On the contrary, if the similarity index is lower, that is, the correlation between the expected vector and the actual vector is weaker, it indicates that the system has a larger judgment error for the dialog state. The first state vector may be determined at this time more according to the "latest situation of the user", that is, the actual vector determined according to the input information of the previous round of dialog, to weaken the relevance of the context, to attempt to redetermine a possible new intention of the user, or an intention that is not recognized.
That is, the first weight coefficient of the desired vector should be determined according to the similarity index; the first weight coefficient is positively correlated with the similarity index. Determining a second weight coefficient of the actual vector according to the similarity index; the second weight coefficient is inversely related to the similarity index. And further weighting and calculating the expected vector, the actual vector, the first weight coefficient and the second weight coefficient to determine a first state vector.
And 103, determining a target state vector corresponding to the current wheel conversation according to the first state vector corresponding to the current wheel conversation and the conversation information of the current wheel conversation.
In this embodiment, the output information of the system at the current wheel may be determined based on the first state vector. Or determining the output information of the system at the current wheel according to other state vectors of the current wheel. The output information of the system in the current round is determined, for example, from a second state vector determined based on the dialog information of the previous round of dialog. After the system provides the output information of the current wheel to the user, the input information fed back by the user based on the output information of the current wheel can be obtained. The output information and the input information of the current round will constitute the complete dialog information of the current round.
Further, a target state vector corresponding to the current wheel dialogue can be determined according to the first state vector corresponding to the current wheel and the dialogue information of the current wheel. And judging the conversation semantic content actually related to the current round according to the conversation information of the current round. From the first state vector, semantic analysis in connection with the above may be achieved in connection with several previous dialog rounds.
In addition, because the weight configuration of the expected vector and the actual vector exists in the first state vector, the context relevance can be continued and strengthened under the condition that the original conversation state is judged correctly; in the case that the original dialog state is judged incorrectly, the user may try to judge the possible new intention or the unrecognized intention again. Therefore, the situation that in the prior art, the wrong conversation state judgment disturbs the semantic analysis of the context of the contact, so that the wrong judgment of the user intention occurs is avoided.
According to the technical scheme, the beneficial effects of the embodiment are as follows: determining a first state vector by using the expected vector and the actual vector, and determining a template state vector of the current wheel by using the first state vector; the judgment of the dialog state of the current round and the identification of the user intention are realized in combination with the context semantics in the multi-round dialog; and thereby avoid a situation in which erroneous dialog state judgments in the above disturb the semantic analysis of the context in connection with the above, resulting in erroneous judgments of the user's intention.
Fig. 1 shows only a basic embodiment of the method described in the present application, and based on this, certain optimization and expansion can be performed, and other preferred embodiments of the method can also be obtained.
Fig. 2 is a schematic diagram illustrating another embodiment of a session information processing method according to the present application. The present embodiment is further developed on the basis of the foregoing embodiments. The method specifically comprises the following steps:
The contents of the above steps 201 to 202 are the same as those of the previous embodiment, and are not described herein again.
And step 203, determining a second state vector corresponding to the current wheel dialog according to the dialog information of the previous round of dialog.
In the embodiment, the conversation state is further judged according to the conversation information of the previous round of conversation, and the intention expressed by the user in the previous round of conversation is determined; i.e. a second state vector is determined. The combination of the second state vector when analyzing the current round of dialog also enables to some extent the linked context in multiple rounds of dialog.
And 204, determining a target state vector corresponding to the current wheel conversation according to the first state vector, the second state vector and the conversation information of the current wheel conversation.
In this embodiment, since the second state vector is already determined, in the process of determining the target state vector, the first state vector, the second state vector and the dialog information of the current dialog are determined together. Therefore, more information related to the above language context can be involved in the target state vector, and the accuracy of semantic analysis on the current wheel conversation information can be further improved.
And step 205, determining output information corresponding to the next wheel telephone according to the target state vector corresponding to the current wheel telephone.
And step 206, determining an expected vector corresponding to the next wheel dialog according to the output information corresponding to the next wheel dialog.
In fact, any one of the dialog rounds may be the current round in this embodiment. Any current round session may also be considered as the "previous round" relative to its next round session. That is, the above process of determining the target state vector of the current session can be performed in a loop in multiple sessions.
Specifically, the output information corresponding to the next wheel telephone can be determined according to the target state vector corresponding to the current wheel telephone; and meanwhile, determining an expected vector corresponding to the next wheel microphone according to the output information corresponding to the next wheel microphone. And then starting the next dialog, and determining the target state vector corresponding to the next dialog in the same way.
Therefore, the method in the embodiment can continuously determine the conversation state according to the conversation information of each round and the previous conversation information along with the progress of multiple rounds of conversations, so that the intention of the user can be more accurately judged, and the experience of man-machine conversation of the intelligent customer service system is improved.
Fig. 3 is a schematic diagram illustrating a specific embodiment of a session information processing apparatus according to the present application. The apparatus of this embodiment is a physical apparatus for performing the method described in FIGS. 1-2. The technical solution is essentially the same as that in the above embodiment, and the corresponding description in the above embodiment is also applicable to this embodiment. The device in the embodiment comprises:
a history vector determining module 301, configured to determine an expected vector and an actual vector corresponding to a previous pair of dialogues;
a first state vector determining module 302, configured to determine, according to the expected vector and the actual vector, a first state vector corresponding to the current wheel dialog;
and the target state vector determining module 303 is configured to determine a target state vector corresponding to the current wheel session according to the first state vector corresponding to the current wheel session and the session information of the current wheel session.
In addition, on the basis of the embodiment shown in fig. 3, it is preferable that:
the history vector determination module 301 includes:
an expected vector determining unit 311, configured to determine an expected keyword according to a historical state vector corresponding to a previous dialog; based on the desired keywords, a desired vector is determined.
And an actual vector determining unit 312, configured to determine an actual vector according to the dialog information of the previous dialog.
The first state vector determination module 302 includes:
a similarity index determination unit 321, configured to determine similarity indexes of the desired vector and the actual vector.
A weight determining unit 322, configured to determine weight coefficients of the desired vector and the actual vector according to the similarity index.
A first state vector determination unit 323 for determining a first state vector based on the desired vector, the actual vector and the weight coefficients.
The weight determination unit 322 includes:
a first weighting subunit 3221, configured to determine a first weighting coefficient of the desired vector according to the similarity index; the first weight coefficient is positively correlated with the similarity index.
A second weighting subunit 3222, configured to determine a second weighting coefficient of the actual vector according to the similarity index; the second weight coefficient is inversely related to the similarity index.
Further comprising:
and the second state vector determining module 304 is configured to determine, according to the dialog information of the previous dialog, a second state vector corresponding to the current dialog.
The target state vector determination module 303 determines a target state vector corresponding to the current round of dialog according to the first state vector, the second state vector, and the dialog information of the current round of dialog.
In addition to the above-described methods and apparatus, embodiments of the present application may also be a computer program product comprising computer program instructions that, when executed by a processor, cause the processor to perform the steps in the methods according to the various embodiments of the present application described in the "exemplary methods" section of this specification, above.
The computer program product may be written with program code for performing the operations of embodiments of the present application in any combination of one or more programming languages, including an object oriented programming language such as Java, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device, or entirely on the remote computing device or server.
Furthermore, embodiments of the present application may also be a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, cause the processor to perform steps in a method according to various embodiments of the present application described in the "exemplary methods" section above of this specification.
The computer-readable storage medium may take any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may include, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
The foregoing describes the general principles of the present application in conjunction with specific embodiments, however, it is noted that the advantages, effects, etc. mentioned in the present application are merely examples and are not limiting, and they should not be considered essential to the various embodiments of the present application. Furthermore, the foregoing disclosure of specific details is for the purpose of illustration and description and is not intended to be limiting, since the foregoing disclosure is not intended to be exhaustive or to limit the disclosure to the precise details disclosed.
The block diagrams of devices, apparatuses, systems referred to in this application are only given as illustrative examples and are not intended to require or imply that the connections, arrangements, configurations, etc. must be made in the manner shown in the block diagrams. These devices, apparatuses, devices, systems may be connected, arranged, configured in any manner, as will be appreciated by those skilled in the art. Words such as "including," "comprising," "having," and the like are open-ended words that mean "including, but not limited to," and are used interchangeably therewith. The words "or" and "as used herein mean, and are used interchangeably with, the word" and/or, "unless the context clearly dictates otherwise. The word "such as" is used herein to mean, and is used interchangeably with, the phrase "such as but not limited to".
It should also be noted that in the devices, apparatuses, and methods of the present application, the components or steps may be decomposed and/or recombined. These decompositions and/or recombinations are to be considered as equivalents of the present application.
The previous description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other aspects without departing from the scope of the application. Thus, the present application is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
The foregoing description has been presented for purposes of illustration and description. Furthermore, the description is not intended to limit embodiments of the application to the form disclosed herein. While a number of example aspects and embodiments have been discussed above, those of skill in the art will recognize certain variations, modifications, alterations, additions and sub-combinations thereof.
Claims (10)
1. A processing method of session information comprises the following steps:
determining an expected vector and an actual vector corresponding to the previous pair of dialogues;
determining a first state vector corresponding to the current wheel conversation according to the expected vector and the actual vector;
and determining a target state vector corresponding to the current wheel conversation according to the first state vector corresponding to the current wheel conversation and the conversation information of the current wheel conversation.
2. The method of claim 1, the determining the expected vector corresponding to the previous pair of dialogs comprising:
determining expected keywords according to the historical state vector corresponding to the previous dialog;
and determining the expected vector according to the expected keyword.
3. The method of claim 1, the determining the actual vector corresponding to the previous pair of dialogs comprising:
and determining the actual vector according to the dialogue information of the previous dialogue.
4. The method of claim 1, wherein determining a first state vector corresponding to a current wheel talk according to the desired vector and the actual vector comprises:
determining similarity indices of the desired vector and the actual vector;
respectively determining the weight coefficients of the expected vector and the actual vector according to the similarity indexes;
determining the first state vector according to the expected vector, the actual vector and the weight coefficient.
5. The method of claim 4, wherein determining the weighting coefficients of the desired vector and the actual vector according to the similarity index comprises:
determining a first weight coefficient of the expected vector according to the similarity index; the first weight coefficient is positively correlated with the similarity index;
determining a second weight coefficient of the actual vector according to the similarity index; the second weight coefficient is inversely related to the similarity index.
6. The method of claim 1, further comprising, prior to determining the target state vector corresponding to the current wheel dialog:
and determining a second state vector corresponding to the current round of conversation according to the conversation information of the previous round of conversation.
7. The method of claim 6, wherein determining the target state vector corresponding to the current wheel session according to the first state vector corresponding to the current wheel session and the session information of the current wheel session comprises:
and determining a target state vector corresponding to the current wheel conversation according to the first state vector, the second state vector and the conversation information of the current wheel conversation.
8. The method of any of claims 1 to 7, further comprising:
and determining the output information corresponding to the next wheel telephone according to the target state vector corresponding to the current wheel telephone.
9. The method of claim 8, further comprising:
and determining an expected vector corresponding to the next wheel microphone according to the output information corresponding to the next wheel microphone.
10. A processing apparatus of session information, comprising:
a history vector determination module, configured to determine an expected vector and an actual vector corresponding to the previous pair of dialogues;
the first state vector determining module is used for determining a first state vector corresponding to the current wheel conversation according to the expected vector and the actual vector;
and the target state vector determining module is used for determining the target state vector corresponding to the current wheel conversation according to the first state vector corresponding to the current wheel conversation and the conversation information of the current wheel conversation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110305854.0A CN113076408B (en) | 2021-03-19 | 2021-03-19 | Session information processing method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110305854.0A CN113076408B (en) | 2021-03-19 | 2021-03-19 | Session information processing method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113076408A true CN113076408A (en) | 2021-07-06 |
| CN113076408B CN113076408B (en) | 2024-09-17 |
Family
ID=76613254
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110305854.0A Active CN113076408B (en) | 2021-03-19 | 2021-03-19 | Session information processing method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113076408B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180285348A1 (en) * | 2016-07-19 | 2018-10-04 | Tencent Technology (Shenzhen) Company Limited | Dialog generation method, apparatus, and device, and storage medium |
| CN109582767A (en) * | 2018-11-21 | 2019-04-05 | 北京京东尚科信息技术有限公司 | Conversational system processing method, device, equipment and readable storage medium storing program for executing |
| KR20190103504A (en) * | 2018-02-12 | 2019-09-05 | 보리 주식회사 | Continuous Conversation Method and Its System by Automating Conversation Scenario Collection |
| CN110309275A (en) * | 2018-03-15 | 2019-10-08 | 北京京东尚科信息技术有限公司 | A kind of method and apparatus that dialogue generates |
| CN110334201A (en) * | 2019-07-18 | 2019-10-15 | 中国工商银行股份有限公司 | A kind of intension recognizing method, apparatus and system |
| CN111339781A (en) * | 2020-02-10 | 2020-06-26 | 科大讯飞华南人工智能研究院（广州）有限公司 | Intention recognition method and device, electronic equipment and storage medium |
| CN112445902A (en) * | 2019-09-04 | 2021-03-05 | 深圳Tcl数字技术有限公司 | Method for identifying user intention in multi-turn conversation and related equipment |
-
2021
- 2021-03-19 CN CN202110305854.0A patent/CN113076408B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180285348A1 (en) * | 2016-07-19 | 2018-10-04 | Tencent Technology (Shenzhen) Company Limited | Dialog generation method, apparatus, and device, and storage medium |
| KR20190103504A (en) * | 2018-02-12 | 2019-09-05 | 보리 주식회사 | Continuous Conversation Method and Its System by Automating Conversation Scenario Collection |
| CN110309275A (en) * | 2018-03-15 | 2019-10-08 | 北京京东尚科信息技术有限公司 | A kind of method and apparatus that dialogue generates |
| CN109582767A (en) * | 2018-11-21 | 2019-04-05 | 北京京东尚科信息技术有限公司 | Conversational system processing method, device, equipment and readable storage medium storing program for executing |
| CN110334201A (en) * | 2019-07-18 | 2019-10-15 | 中国工商银行股份有限公司 | A kind of intension recognizing method, apparatus and system |
| CN112445902A (en) * | 2019-09-04 | 2021-03-05 | 深圳Tcl数字技术有限公司 | Method for identifying user intention in multi-turn conversation and related equipment |
| CN111339781A (en) * | 2020-02-10 | 2020-06-26 | 科大讯飞华南人工智能研究院（广州）有限公司 | Intention recognition method and device, electronic equipment and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 杨成彪;吕荣荣;吴刚;: "一种基于记忆网络的多轮对话下的意图识别方法", 电子技术与软件工程, no. 10 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113076408B (en) | 2024-09-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108428447B (en) | Voice intention recognition method and device | |
| US20090076798A1 (en) | Apparatus and method for post-processing dialogue error in speech dialogue system using multilevel verification | |
| CN112468659B (en) | Quality evaluation method, device, equipment and storage medium applied to telephone customer service | |
| CN110875059B (en) | Method and device for judging reception end and storage device | |
| CN103337241B (en) | Voice recognition method and device | |
| WO2021073179A1 (en) | Named entity identification method and device, and computer-readable storage medium | |
| CN111627423A (en) | VAD tail point detection method, device, server and computer readable medium | |
| CN114067786A (en) | Voice recognition method and device, electronic equipment and storage medium | |
| CN113779208A (en) | Method and device for man-machine conversation | |
| CN112767916A (en) | Voice interaction method, device, equipment, medium and product of intelligent voice equipment | |
| CN111611358A (en) | Information interaction method and device, electronic equipment and storage medium | |
| CN114155839A (en) | Voice endpoint detection method, device, equipment and storage medium | |
| CN118113831A (en) | Question-answer data processing method and device, electronic equipment and storage medium | |
| CN115394318A (en) | Audio detection method and device | |
| CN115509485A (en) | Filling-in method and device of business form, electronic equipment and storage medium | |
| CN113889091A (en) | Voice recognition method and device, computer readable storage medium and electronic equipment | |
| CN112863496B (en) | Voice endpoint detection method and device | |
| CN109273004B (en) | Predictive speech recognition method and device based on big data | |
| CN110222161B (en) | Intelligent response method and device for conversation robot | |
| CN111931507A (en) | Method, apparatus, medium, and device for acquiring a tuple set for implementing a session | |
| CN113076408A (en) | Session information processing method and device | |
| CN111970311B (en) | Session segmentation method, electronic device and computer readable medium | |
| CN112687296B (en) | Audio disfluency identification method, device, equipment and readable storage medium | |
| CN114417881A (en) | Sensitive word detection method and device, electronic equipment and storage medium | |
| CN114203169A (en) | Method, device and equipment for determining voice recognition result and storage medium |
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 |