CN113192505A - Session information generation method and device and service robot - Google Patents
Session information generation method and device and service robot Download PDFInfo
- Publication number
- CN113192505A CN113192505A CN202110479849.1A CN202110479849A CN113192505A CN 113192505 A CN113192505 A CN 113192505A CN 202110479849 A CN202110479849 A CN 202110479849A CN 113192505 A CN113192505 A CN 113192505A
- Authority
- CN
- China
- Prior art keywords
- word slot
- information
- service
- session
- reply
- 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
Images
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; 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
- G10L15/1815—Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; 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
Abstract
The embodiment of the specification relates to the technical field of artificial intelligence, and particularly discloses a session information generation method, a session information generation device and a service robot, wherein the method comprises the following steps: receiving service requirement information of a user; calling a reply dialect corresponding to the service demand information and a word slot identifier in the reply dialect; the word slot mark at least comprises an original word slot mark corresponding to the original word slot and a mapping word slot mark corresponding to the mapping word slot; extracting an original word slot value corresponding to an original word slot identifier in a reply language based on the service demand information and the corresponding associated service information; extracting a mapping word slot value corresponding to the corresponding mapping word slot identifier based on an original word slot value associated with the mapping word slot identifier in the answering operation and a preset word slot filling mode; and correspondingly filling the original word slot value and the mapping word slot value into the positions of the original word slot identifier and the mapping word slot identifier of the reply language to generate service reply information corresponding to the service demand information, thereby improving the flexibility of reply information generation.
Description
Technical Field
The present specification relates to the technical field of artificial intelligence, and in particular, to a session information generation method, an apparatus, and a service robot.
Background
Nowadays, with the high-speed development of internet technology, customer service staff expand from traditional telephone customer service to various channels such as APP, WeChat, webpage and the like, enterprises can more conveniently provide services for users, but customer service also faces the problems that service channels are more diversified, the number of served customers is increased rapidly and the like. With the rise of artificial intelligence, the development of the conversation robot greatly saves human resources and greatly improves the response speed of service users. In the session process, the service robot can generate service reply information based on user demand information or service information stored in the service system. However, the representation form of the user demand information or the service information stored in the service system is usually fixed, so that the generated service reply information is rigid, and bad experience is brought to the user.
Disclosure of Invention
An object of the embodiments of the present specification is to provide a session information generation method, a session information generation device, and a service robot, which can improve flexibility of session information generation.
The present specification provides a session information generation method, a device and a service robot, which are implemented in the following ways:
a session information generation method is applied to a service robot, and comprises the following steps: receiving service requirement information of a user; calling a reply dialect corresponding to the service demand information and a word slot identifier in the reply dialect; the word slot mark at least comprises an original word slot mark corresponding to an original word slot and a mapping word slot mark corresponding to a mapping word slot; the original word slot is a word slot filled based on the service demand information and the corresponding associated service information; the mapping word slot is a word slot filled based on the associated original word slot and a preset word slot filling mode; extracting an original word slot value corresponding to the original word slot identifier in the reply grammar based on the service demand information and the corresponding associated service information; extracting a mapping word slot value corresponding to the corresponding mapping word slot identifier based on the original word slot value associated with the mapping word slot identifier in the answering operation and a preset word slot filling mode; correspondingly filling the original word slot value and the mapping word slot value into the positions of the original word slot identifier and the mapping word slot identifier of the reply grammar to generate service reply information corresponding to the service demand information; and feeding back the service reply information to the user.
In other embodiments of the method provided herein, the method further comprises: acquiring session context information corresponding to a session identifier of the service demand information, wherein the session context information at least comprises an extracted word slot value and a corresponding word slot identifier under the session identifier; correspondingly, an original word slot value corresponding to the original word slot identifier or a mapping word slot value corresponding to the mapping word slot identifier in the reply sentence is extracted by combining the context information of the conversation.
In other embodiments of the method provided in this specification, the invoking a reply language corresponding to the service requirement information includes: acquiring session transfer information of a service scene to which the service demand information belongs; the session transfer information comprises at least one session node and a transfer relationship between the session nodes; the session node is used for representing different processing links in the service processing process of the corresponding service scene; at least part of the session nodes are associated with reply dialogs related to corresponding processing links; determining a session node which is corresponding to the service demand information in the session flow information and is associated with a reply dialog as a target session node based on the intention recognition result of the service demand information; and taking the reply dialect associated with the target session node as the reply dialect corresponding to the service requirement information.
In other embodiments of the method provided in this specification, at least some session nodes in the session transfer information are associated with word slot identifiers related to corresponding processing links; accordingly, whether the circulation condition satisfies the word slot value determination filled with the word slot identifier associated with the session node.
In other embodiments of the method provided herein, the method further comprises: acquiring a streamed session node under the session identifier of the service demand information; correspondingly, the session node which is associated with the reply dialog and corresponds to the service requirement information in the session circulation information is determined based on the circulated session node and the intention identification result of the service requirement information.
In other embodiments of the method provided in this specification, the original word slot value and the mapped word slot value are normalized, so that the normalized original word slot value and the normalized mapped word slot value are correspondingly filled in the positions where the original word slot identifier and the mapped word slot identifier of the answer grammar are located; wherein the normalization process is constructed based on a pre-packaged normalization process function.
On the other hand, the embodiments of the present specification further provide a session information generating apparatus, applied to a service robot, the apparatus including: the receiving module is used for receiving the service requirement information of the user; the calling module is used for extracting a reply dialect corresponding to the service demand information and a word slot identifier in the reply dialect; the word slot mark at least comprises an original word slot mark corresponding to an original word slot and a mapping word slot mark corresponding to a mapping word slot; the original word slot is a word slot filled based on the service demand information and the corresponding associated service information; the mapping word slot is a word slot filled based on the associated original word slot and a preset word slot filling mode; a word slot value extraction module, configured to determine, based on the service requirement information and the corresponding associated service information, an original word slot value corresponding to the original word slot identifier in the reply grammar; determining a mapping word slot value corresponding to the corresponding mapping word slot identifier based on the original word slot value associated with the mapping word slot identifier in the reply technology and a preset word slot filling mode; the filling module is used for correspondingly filling the original word slot value and the mapping word slot value into the positions of the original word slot identifier and the mapping word slot identifier of the reply speech to generate service reply information corresponding to the service demand information; and the feedback module is used for feeding back the service reply information to the user.
In other embodiments of the apparatus provided in this specification, the word groove value extraction module includes: a first obtaining unit, configured to obtain session context information corresponding to a session identifier of the service demand information, where the session context information at least includes a word slot value extracted under the session identifier and a corresponding word slot identifier; and the extracting unit is used for extracting an original word slot value corresponding to the original word slot identifier or a mapping word slot value corresponding to the mapping word slot identifier in the reply language by combining the context information of the conversation.
In other embodiments of the apparatus provided in this specification, the invoking module includes: the second obtaining unit is used for obtaining the session transfer information of the service scene to which the service demand information belongs; the session transfer information comprises at least one session node and a transfer relationship between the session nodes; the session node is used for representing different processing links in the service processing process of the corresponding service scene; at least part of the session nodes are associated with reply dialogs related to corresponding processing links; a session node determination unit, configured to determine, based on an intention recognition result of the service demand information, a session node associated with a reply dialog corresponding to the service demand information in the session flow information, as a target session node; a reply-to-talk determining unit, configured to use the reply talk associated with the target session node as the reply talk corresponding to the service requirement information.
In another aspect, this specification further provides a service robot, where the service robot includes at least one processor and a memory for storing processor-executable instructions, and the instructions, when executed by the processor, implement the steps of the method according to any one or more of the foregoing embodiments.
In the session information generation method, the session information generation device, and the service robot provided in one or more embodiments of the present specification, the mapping word slot is further configured under the condition that the original information provided by the user or the original information form of the service system is not changed, so that the original information provided by the user or the original information of the service system is appropriately transformed, and the flexibility of the finally generated reply information can be improved. Meanwhile, the original information extracted in the session transfer process can be stored, and the efficiency of generating subsequent reply information is improved on the basis of ensuring the reply flexibility.
Drawings
In order to more clearly illustrate the embodiments of the present specification or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, it is obvious that the drawings in the following description are only some embodiments described in the present specification, and for those skilled in the art, other drawings can be obtained according to the drawings without any creative effort. In the drawings:
fig. 1 is a schematic flow chart of an implementation of a session information generation method provided in this specification;
fig. 2 is a schematic diagram of session flow information in an example provided by the present specification;
fig. 3 is a schematic diagram of session flow information in an example provided by the present specification;
fig. 4 is a schematic block diagram of a session information generation apparatus provided in this specification;
fig. 5 is a schematic block diagram of a session information generation apparatus in an example provided in this specification;
FIG. 6 is a block diagram of a configuration management device in one example provided herein;
fig. 7 is a schematic block diagram of an information preprocessing apparatus in an example provided in the present specification;
fig. 8 is a schematic block diagram of a word slot filling apparatus in one example provided in this specification;
fig. 9 is a block configuration diagram of an information post-processing apparatus in one example provided in the present specification;
FIG. 10 is a block diagram of an intent translation device in one example provided in this specification;
FIG. 11 is a block diagram of a data storage device in one example provided herein;
fig. 12 is a schematic diagram of a session information generation flow in an example provided in this specification.
Detailed Description
In order to make those skilled in the art better understand the technical solutions in the present specification, the technical solutions in one or more embodiments of the present specification will be clearly and completely described below with reference to the drawings in one or more embodiments of the present specification, and it is obvious that the described embodiments are only a part of the embodiments of the specification, and not all embodiments. All other embodiments obtained by a person skilled in the art based on one or more embodiments of the present specification without making any creative effort shall fall within the protection scope of the embodiments of the present specification.
In a service scenario of service handling or service consultation, a user can initiate a service handling or service consultation request through a user terminal. The user terminal can also be a computer, an IPAD, a smart watch and the like. The service robot may be, for example, a device that a service party can realize intelligent interaction, such as a server and an intelligent service terminal in a service system of the service party.
For example, a service application may be installed in the user terminal, and the user may initiate a service transaction or service consultation request through the service application. The service robot may construct a conversation window based on the request and configure a conversation identification to identify the conversation information involved in the conversation window. The service robot can send the conversation window to the user terminal so that the user terminal can display based on the business application. Accordingly, the user may enter information in the conversation window. The user can input information in the modes of words, pictures, voice and the like. The information input by the user can be used as business requirement information and sent to the service robot. The service robot can generate corresponding service reply information according to the service demand information, and feed back the service reply information to the user terminal so as to display the service reply information in the session window.
In the process of communication between the user terminal and the service robot, the amount of information covered by the service demand information is usually small, even if the pre-stored associated service information can be extracted from the service system to generate service reply information. However, since the form of the service information pre-stored in the service scene is generally fixed, when the service robot replies based on the extracted information, the reply content is rigid, and the user experience is not good.
Fig. 1 is a flowchart illustrating an embodiment of a session information generation method provided in this specification. As shown in fig. 1, embodiments of the present specification further provide a session information generation method, which may be applied to a service robot. Accordingly, the method may comprise the following steps.
S20: and receiving the service requirement information of the user.
The user can send the service requirement information through the user terminal. As shown in the above scenario example, an interactive window may be presented in the user terminal through which the user may interact with the service robot. The information input by the user in the interactive window can be used as the service requirement information of the user.
Or, the user may initiate voice interaction at the user terminal, and may use the voice interaction request as a request for service handling or service consultation. The service robot may establish a voice interactive communication connection with the user terminal and assign a session identification to the established communication connection. The session identification may be, for example, a randomly generated character. Alternatively, the session identifier may be constructed based on the user identifier and the timestamp. The service robot can carry out conversation communication with the user in a voice interaction mode, and conversation information related in the conversation communication process can be identified through a conversation identifier. Correspondingly, the voice information recorded by the user in the voice interaction can be used as the service requirement information of the user.
Alternatively, the user may also interact directly with the service robot. For example, the service robot interacts with the service robot by means of voice, information input in a display interface of the service robot, and the like. Accordingly, the voice information of the user or other types of input information received by the service robot can be used as the business requirement information of the user. Of course, the user may also perform session communication with the service robot in other manners, which is not limited herein.
The service requirement information of the user can exist in various forms, such as text information, image information, voice information and the like. In order to facilitate the generation of the consultation response information, the service requirement information can be preprocessed. For example, non-text information can be converted into text information, and filtering processing can be performed on sensitive words, stop words and the like.
S22: calling a reply dialect corresponding to the service demand information and a word slot identifier in the reply dialect; the word slot mark at least comprises an original word slot mark corresponding to an original word slot and a mapping word slot mark corresponding to a mapping word slot; the original word slot is a word slot filled based on the service demand information and the corresponding associated service information; the mapping word slot is a word slot filled based on the associated original word slot and a preset word slot filling mode.
After receiving the service requirement information, the service robot can call a reply language corresponding to the service requirement information. The answering speech is a mode of answering the service robot to the service demand information. The answer dialect may be configured in advance, for example, different answer dialects may be configured for different user intentions, and after the service robot receives the service requirement information, the service robot may identify a corresponding user intention first, and then extract a corresponding answer dialect. Of course, a model may also be generated by constructing a reply language, so as to analyze the business requirement information based on the model to determine a reply language corresponding to the business requirement information. Of course, other methods may also be used to extract the reply language corresponding to the service requirement information, which is not limited herein.
In some embodiments, the reply utterance may include at least one of a word slot identifier to be filled with a word slot and partially fixed textual information. For example, the answer dialog may be "ask you for # name # i", where "name" is the word slot identifier, "ask you for … … i" as fixed text information. The word slot value corresponding to the word slot identifier can be at least extracted from the service requirement information and the corresponding associated service information. The associated service information may be service information stored in a service system and associated with the service demand information. For example, the service requirement information of the user is "you are, i want to do a credit card", the service robot may extract information such as name, gender, and identification number of the user from the service system based on the user identifier or the incoming call number, and may describe the extracted information as associated service information corresponding to the service requirement information.
The service requirement information or the information contained in the associated service information extracted from the service system is limited, and the information representation form is usually fixed, so that the reply form is rigid and inflexible, and gives a bad experience to the user. If the reply information is generated by using the intelligent context algorithm, the intelligent context algorithm is generally complex, has high input cost and is not suitable for business handling or consulting application scenes with simple processing logic. In this embodiment, the word slot filled based on the service requirement information and the corresponding associated service information may be described as an original word slot. And further configures the mapped word slots of the original word slots as needed. The word slot value of the mapping word slot can be the deformation and processing of the word slot value of the original word slot, so that the function of making the conversation smooth and natural is achieved, and the flexibility of the business reply sentence is improved. Correspondingly, the mapping word slot may be a word slot filled based on the associated original word slot and a preset word slot filling manner.
For example, the service robot extracts the name of the user from the business requirement information or the associated business information as masculine XX and gender female. If a more flexible conversation is desired in the following dialog, such as to call the user as a woman in europe and sun, what the user's last name is not known only by the value of the slot of the original slot, and the user can only continue to ask for more information on the premise of knowing the user's name and gender, such as "ask for your last name? "or" should i call your woman or mr? ", greatly reducing the conversational experience.
As shown in table 1 below, the mapping word slot corresponding to the original word slot, the word slot filling manner of the mapping word slot, and the like may be configured in advance as needed, so as to generate the service reply information based on the mapping word slot in practical application, and improve the flexibility of the reply information. Table 1 shows some pre-configured original word slots and associated mapped word slots.
TABLE 1
| Original word slot mark | Mapping word slot identification | Original word slot value | Mapping word slot values | Normalization process |
| Full name of user | User surname | Air conditioner | Air conditioner | Extracting user surnames |
| Gender of user | Title to be called | For male | Mr. first | Air conditioner |
| Gender of user | Title to be called | Woman | Female | Air conditioner |
Accordingly, when corresponding service reply information is generated based on a reply dialect, the word slot identifier R in the reply dialect may be extracted. If the word slot identifier R is the original word slot, the word slot value of the word slot identifier R may be extracted from the service requirement information or the associated service information. If the word slot identifier R is a mapping word slot, the service robot can acquire a word slot identifier Q of an original word slot associated with the word slot identifier, and further extract a word slot value of the word slot identifier Q from the service demand information or the associated service information; and generating a word slot value of the word slot identifier R based on the word slot filling mode of the word slot identifier R and the word slot value of the word slot identifier Q.
Of course, the service robot may also store the word slot identifier and the corresponding extracted word slot value in the session context information of the service reply information after associating them. The session context information may refer to information that the service requirement information relates to in a session advancing process. The session context information may be identified by a session identification. Correspondingly, in the process of filling the word slot value, the word slot value corresponding to the word slot identifier may be extracted from the session context information, and if the word slot value is not extracted, the word slot value of the original word slot may be extracted from the service requirement information or the associated service information, or the word slot value of the mapped word slot may be generated based on the word slot value of the associated original word slot and a preset word slot filling manner.
Correspondingly, in some embodiments, the service robot may obtain session context information corresponding to the session identifier of the service requirement information, where the session context information at least includes the extracted word slot value and the corresponding word slot identifier under the session identifier. Accordingly, the original word slot value corresponding to the original word slot identifier or the mapped word slot value corresponding to the mapped word slot identifier in the reply dialog may be determined in combination with the context information.
For example, for the configuration information shown in table 1 above. If the word slot filling manner of the mapping word slot is default filling, such as the first row in table 1, and the mapping word slot value (the fourth column) is configured as a null value, the mapping word slot value may be directly filled with the associated original word slot value. If not default population, the word slot values of the mapping word slot may be pre-configured, such as the second and third rows in Table 1, with the mapping word slot values (fourth column) configured as "lady" and "mr". If the word slot identifier R is "title", the original word slot identifier Q "user gender" associated with "title" may be obtained according to the word slot configuration information, and then the original word slot value of the original word slot identifier "user gender" is extracted from the session context information, and if the original word slot value "woman" is extracted, the mapped word slot value "woman" of the mapped word slot identifier "title" may be determined based on the word slot configuration information.
Or, a normalization process in the word slot configuration information may also be obtained, and if the normalization process is configured to be non-empty, a corresponding normalization process function may be called based on the configuration information to perform a normalization process on the mapping word slot value, so as to generate the service reply information using the mapping word slot value after the normalization process. And if the standardized processing configuration is null, generating the service reply information by using the mapping word slot value. Table 2 shows an example of the normalized original word bin value and the mapped word bin value.
TABLE 2
Of course, the normalization process described above may also be performed on the original map values. Accordingly, in some embodiments, the service robot may perform normalization on the original word slot value and the mapped word slot value, so as to correspondingly fill the normalized original word slot value and the normalized mapped word slot value into the positions of the original word slot identifier and the mapped word slot identifier of the answer grammar. Wherein the normalization process may be constructed based on a pre-packaged normalization process function. The initial extracted word slot value is standardized through some pre-packaged functions, so that the finally obtained word slot value is more in line with the actual requirement.
Based on the above embodiment, the service robot may extract the word slot value of each word slot identifier in the reply utterance using steps S24, S26. If the word slot identifier corresponds to the original word slot, step S24 may be executed to increase the word slot value of the corresponding word slot identifier. If the word slot id corresponds to the mapped word slot, step S26 may be executed to increase the word slot value of the corresponding word slot id.
S24: and extracting an original word slot value corresponding to the original word slot identifier in the reply dialog based on the service requirement information and the corresponding associated service information.
S26: and extracting the mapping word slot value corresponding to the corresponding mapping word slot identifier based on the original word slot value associated with the mapping word slot identifier in the reply technology and a preset word slot filling mode.
S28: and correspondingly filling the original word slot value and the mapping word slot value into the positions of the original word slot identifier and the mapping word slot identifier of the reply grammar to generate the service reply information corresponding to the service demand information.
S210: and feeding back the service reply information to the user.
The service robot may feed back the service reply information to the user after generating the service reply information based on step S28. Such as may be sent to a user terminal of a user for presentation at the user terminal. Alternatively, the service robot may also display the service reply information to the user directly through a display or in a voice form.
In the solution provided by the above embodiment, the mapping word slot is further configured under the condition that the original information provided by the user or the original information form of the service system is not changed, so that the original information provided by the user or the original information of the service system is appropriately deformed, and the flexibility of the finally generated reply information can be improved. Meanwhile, the original information extracted in the session transfer process can be stored, and the efficiency of generating subsequent reply information is improved on the basis of ensuring the reply flexibility.
In other embodiments, the session transfer information for different service scenarios may also be preconfigured. The session transfer information may include at least one session node and a transfer relationship between the session nodes. The session node may be used to characterize different processing links in the service processing process of the corresponding service scenario. The flow relationship may include which session nodes have an association therebetween, a flow direction between the nodes having the association, and the like. The flow relation can also comprise flow conditions among the session nodes, and when the corresponding flow conditions are met, the session nodes jump to another session node. Reply dialogs or word slot identifications needed to associate respective processing links with the session node may also be made. The word slot identifier associated with the session node may be used to characterize the extracted service information required by the service scenario under the session node. The reply dialog may be used to characterize the dialog that the traffic scenario needs to use when replying under this session node.
By configuring the session nodes, the stream-to-stream relation between the session nodes and the required reply dialogs or word slot identifiers, the reply dialogs, the service information and the like required by the current service requirement of the user in the multi-turn session process can be better determined, and the accuracy of the reply of the current service requirement is improved.
The session nodes, the circulation relationship of the session nodes, and the word slots or dialogs associated with the session nodes included in each service scene can be configured by service personnel according to experience; or an intelligent scene configuration generation model can be constructed by combining the existing business processing flow and the reply information, and the intelligent generation model is utilized to determine the scene configuration information. Of course, the intelligent generation model of scene configuration may be constructed in other ways.
If the service requirement information is sent through the session window, the service robot may first determine whether the service requirement information is the first piece of service requirement information under the session identifier based on the session identifier. If yes, intention identification can be carried out on the service requirement information to determine a service scene corresponding to the service requirement information, and then session circulation information under the service scene is obtained.
The session flow information may be stored in the form of a knowledge graph, as shown in fig. 2. A, B, C in fig. 2 indicate session nodes, and session node a is associated with word slot 1 and word slot 2, session node B is associated with word slot 3, and session node C is associated with word slot 4. A. A line between B and A, C indicates that there is an association between a and B, an association between a and C, and arrows indicate "X (word slot 1 ═ a & word slot 2 ═ B), and Y (word slot 1 ═ C & word slot 2 ═ d)" on a flow direction line between session nodes, which indicate flow conditions. Of course, the session flow information may also be stored in other manners, such as in the form of a table, which is not limited herein.
For example, the service requirement information of the user is "you are, i want to do a credit card". The service robot can identify the intention of the service requirement information so as to determine that the service scene corresponding to the service requirement information is handled by a credit card. Then, the session transfer information corresponding to the service scenario "credit card transaction" can be obtained. After the session transfer information is acquired, the service robot may further determine that the first service processing link related to the service demand information is "basic information of the user is confirmed". The session node involved in the "confirm user's basic information" processing link can be obtained, as shown in fig. 3.
Referring to fig. 3, the session node involved in the "confirm user's basic information" processing link may include "acquire user's basic information", "confirm whether or not the user applies for", "ask for user's identity card number", and the like, and is configured with a word slot "name", "gender", "identity card number" associated with "acquire user's basic information", and "confirm whether or not the user applies for" associated reply "you good, ask you for # name # called # i", "ask for user's identity card number" associated reply "you good, please provide a complete identity card number, and we need to inquire your personal information first. And the association relationship (represented by a connecting line), the circulation direction (represented by an arrow) and the circulation conditions of 'success in acquiring the user information' and 'failure in acquiring the user information' among the session nodes are also configured.
The first session node "acquiring the user basic information" may be determined based on the flow-to-flow relationship between the session nodes. For example, the service robot may access the database of the service system according to the incoming call number or the user identifier, and extract the word slot value of the word slot identifier associated with the session node: the name of the user is Zhang III, the name of the user is male, and the identification number of the user is XXXX. Wherein, the word slot values of the corresponding word slot marks are Zhang III, Man and XXXX.
After extracting the word slot value, the service robot may associate the word slot identifier with the word slot value and store the associated word slot identifier in the session context information. And jumping to a session node to confirm whether the user applies for the session node according to corresponding streaming conditions. The session node determines whether the user applies for the 'associated answer dialog' good, asks for that the user is the name # of the # name #, and the service robot can take the answer dialog as the answer dialog corresponding to the service requirement information. The service robot can obtain the word slot identifier "name" and "title" in the reply sentence, and first extract the word slot value of the corresponding word slot identifier from the context information of the conversation, such as the word slot value "zhang san" of the "name". For the word slot identifier 'appellation', the service robot may extract the original word slot 'gender' associated with the 'appellation', extract the word slot value 'male' of the 'gender' from the session context information, and extract the word slot value 'mr' of the 'appellation' based on the preset word slot filling mode of the word slot value 'male' and the 'appellation'. The extracted word slot value can be filled into the position of the corresponding word slot identifier of the reply dialect, and the service reply information' you good ask you for mr. zhang.
If the service robot does not inquire the basic information of the user, the service robot can jump to a session node to inquire the identity card number of the user according to the circulation condition, and obtains a reply technique associated with the session node as the reply technique of the service demand information. The reply language does not relate to the word slot identifier, and the reply language can be directly used as a business reply message "you are good, please provide a complete identification number, and we need to inquire your personal information first". After receiving the identity card number input by the user, the service robot can execute the next session node, extract the basic information of the user based on the identity card number, and extract the word slot value from the basic information based on the word slot corresponding to the session node.
After extracting the basic information again or confirming the basic information of the user, the service robot can process the next session node again. For example, after the service robot receives the reply information sent by the user, the service robot may identify the intention of the new service requirement information after taking the reply information of the user as the new service requirement information, so as to determine the session identifier to which the service requirement information belongs. In some embodiments, the service robot may store the session node, which is circulated in the service reply information that generates each service demand information, as a circulated session node, in the session context information. When the intention recognition is performed on the new service requirement information, the session context information corresponding to the session identifier may be extracted, and the intention recognition is performed on the service requirement information based on the session nodes that have circulated in the session context information, so as to determine the session nodes associated with the reply dialect corresponding to the service requirement information in the session circulation information, so as to further improve the accuracy of positioning the session nodes corresponding to the new service requirement information, and further improve the accuracy of generating the service reply information.
If the user asks you to be Zhang Sanzai, then the user replies yes. The service robot may assist in intent recognition based on the user's business need information, session context information, to determine the next business process link as "ask the user what type of credit card the user wants to handle". The session nodes involved in the service processing element can then be extracted from the session flow information. And extracting a first session node, a second session node and the like based on the circulation relationship so as to generate corresponding service reply information based on the word slot identification or the reply operation associated with each session node.
In the above embodiment, each service processing link is further subdivided into a plurality of session nodes, different association information (such as different word slot identifiers and answer conventions) is configured for each session node, and then each session node is connected based on a flow relationship, so that the session processing flow of each service processing link is clearer and clearer, and the flexibility of session answer is further improved.
Correspondingly, in some embodiments, the session transfer information of the service scene to which the service demand information belongs may be acquired. The session transfer information includes at least one session node and a transfer relationship between the session nodes. The session node is used for representing different processing links in the service processing process of the corresponding service scene. At least part of the session nodes are associated with reply dialogs related to corresponding processing links; and determining a session node which corresponds to the service demand information in the session flow information and is associated with a reply dialog as a target session node based on the intention recognition result of the service demand information. And taking the reply dialect associated with the target session node as the reply dialect corresponding to the service requirement information. By configuring the session transfer information and determining the reply dialect corresponding to the service demand information based on the session transfer information, the accuracy and the simplicity of the generation of the service reply information corresponding to the service demand information can be further improved.
At least part of the session nodes in the session transfer information can also be associated with word slot identifiers related to corresponding processing links. When forwarding to a session node associated with a word slot identifier, a word slot value corresponding to the corresponding word slot identifier may be extracted, and the extracted word slot value may be stored in session context information, so as to be used when generating service reply information based on a reply technology.
Whether the flow condition is met or not can be determined by the word slot value filled by the word slot identifier associated with the session node, and if the word slot value corresponding to the word slot identifier associated with the first session node in fig. 3 is filled, the flow is switched to the session node to determine whether the user applies for the word slot value; if the word slot value corresponding to the word slot identifier associated with the first session node is null, the flow is transferred to the session node "inquire the user identification number". Alternatively, as shown in fig. 2, the word slot filling type is different, and the direction of the flow is also different. The circulation condition is determined based on the filled word slot value, so that the representation of the circulation condition is more consistent with the actually extracted information, namely more consistent with the actual conversation scene, and the accuracy of the service reply information is further improved.
In other implementations, the service robot may further obtain a session node that has been streamed under the session identifier of the service requirement information. And determining the corresponding session node associated with the reply dialog in the session flow information of the service demand information based on the circulated session node and the intention recognition result of the service demand information, so as to improve the accuracy and efficiency of determining the service reply information.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. For details, reference may be made to the description of the related embodiments of the related processing, and details are not repeated herein.
The foregoing description has been directed to specific embodiments of this disclosure. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.
As shown in fig. 4, based on the method provided by the foregoing embodiment, the present specification further provides a session information generating apparatus, where the apparatus may include:
the receiving module 40 may be configured to receive service requirement information of a user.
The retrieval module 42 may be configured to extract a reply dialect corresponding to the service requirement information and a word slot identifier in the reply dialect; the word slot mark at least comprises an original word slot mark corresponding to an original word slot and a mapping word slot mark corresponding to a mapping word slot; the original word slot is a word slot filled based on the service demand information and the corresponding associated service information; the mapping word slot is a word slot filled based on the associated original word slot and a preset word slot filling mode.
A word slot value extraction module 44, configured to determine an original word slot value corresponding to an original word slot identifier in the reply grammar based on the service requirement information and the corresponding associated service information; and determining a mapping word slot value corresponding to the corresponding mapping word slot identifier based on the original word slot value associated with the mapping word slot identifier in the reply technology and a preset word slot filling mode.
The filling module 46 may be configured to correspondingly fill the original word slot value and the mapping word slot value into positions of the original word slot identifier and the mapping word slot identifier of the reply grammar, so as to generate service reply information corresponding to the service demand information.
A feedback module 48, which may be configured to feed back the service reply information to the user.
In other embodiments, the word-groove value extraction module may include:
the first obtaining unit may be configured to obtain session context information corresponding to a session identifier of the service requirement information, where the session context information at least includes a word slot value extracted under the session identifier and a corresponding word slot identifier.
And the extracting unit may be configured to extract, in combination with the context information of the conversation, an original word slot value corresponding to the original word slot identifier in the reply sentence or a mapped word slot value corresponding to the mapped word slot identifier.
In other embodiments, the retrieving module may include:
the second obtaining unit may be configured to obtain session transfer information of a service scene to which the service demand information belongs; the session transfer information comprises at least one session node and a transfer relationship between the session nodes; the session node is used for representing different processing links in the service processing process of the corresponding service scene; at least some of the session nodes are associated with reply dialogs to which the respective processing elements relate.
The session node determination unit may be configured to determine, as a target session node, a session node associated with a reply dialog corresponding to the service requirement information in the session flow information based on the intention recognition result of the service requirement information.
The reply-to-speech determination unit may be configured to use a reply speech associated with the target session node as a reply speech corresponding to the service requirement information.
It should be noted that the above-mentioned apparatus may also include other embodiments according to the description of the above-mentioned embodiments. The specific implementation manner may refer to the description of the related method embodiment, and is not described in detail herein.
Based on the session information generation method or device provided by the above embodiments, the present specification also provides an example of a session information generation device.
As shown in fig. 5, the apparatus may include a configuration management apparatus 1, an information preprocessing apparatus 2, a word slot filling apparatus 3, an information post-processing apparatus 4, an intention recognition apparatus 5, and a data storage apparatus 6. Wherein, the configuration management device 1 is connected with the data storage device 6; the information preprocessing device 2 and the word slot filling device 3 are connected with the information post-processing device 4; the information preprocessing device 2 is also connected with an intention recognition device 5 and a data storage device 6 respectively; the word groove filling means 3 are also connected to the intention recognition means 5.
The configuration management device 1 is responsible for completing the configuration of scene-related data.
The information preprocessing device 2 is used for receiving and analyzing the user request information, then completing scene recognition through the intention recognition device 5, and acquiring scene configuration information and current session context information through the data storage device 6.
The word slot filling means 3 is for acquiring the preprocessed information from the information preprocessing means 2 and acquiring the supplied element information from the intention recognition means 5 to complete the word slot filling.
The information post-processing device 4 is used for obtaining the word slot filling result from the word slot filling device 3, then finishing the session node skip judgment of the session flow according to the scene configuration information, and finally constructing a response message to return.
The intention recognition means 5 is used to provide an intention recognition service.
The data storage 6 is responsible for storing scene configuration information and session context information.
Fig. 6 is a schematic configuration diagram of the configuration management apparatus 1. As shown in fig. 6, the configuration management apparatus 1 may include a word slot configuration unit 101 and a flow configuration unit 102.
The word slot configuration unit 101 is responsible for completing configuration management of the word slot related data.
The process configuration unit 102 is responsible for completing configuration management of process data, forming a flowchart composed of nodes and lines connected with each other, and implementing multi-turn session control flow of the service scenario represented by the flowchart. The nodes represent different processing links in the scene, and of course, one processing link may correspond to a plurality of session nodes. The conversation node can bind or speak one or more word slots and dialogs in the relevant word slot configuration unit 101 according to needs, and is used for representing word slots of a scene needing to be filled under the node; the line represents the jump condition that needs to be met for a state change to occur. As shown in fig. 2, if the word slot filling result of the multi-turn session at the node a is that the word slot 1 is a and the word slot 2 is B, after the branch jump judgment, the branch X is satisfied, and the flow jumps to the node B to continue processing, thereby determining the flow trend.
Fig. 7 is a schematic configuration diagram of the information preprocessing apparatus 2. As shown in fig. 7, the information preprocessing device 2 may include a request parsing unit 201, an information filtering unit 202, and a session management unit 203.
The request parsing unit 201 is responsible for parsing a request message sent by a service server and identifying a user intention.
The information filtering unit 202 is responsible for performing filtering processing on sensitive words, stop words and the like on the request original data.
The session management unit 203 is responsible for obtaining current session context information from the data storage 6.
Fig. 8 is a schematic structural view of the word tank filling device 3. As shown in fig. 8, the word slot filling apparatus 3 may include a word slot obtaining unit 301, an original word slot filling unit 302, a mapped word slot filling unit 303, and a normalization processing unit 304.
The word slot acquiring unit 301 is responsible for acquiring a word slot list that needs to be filled in the current interaction stage from the process configuration unit 102.
The original word slot filling unit 302 is responsible for acquiring information provided by the user or information extracted from the business system from the intention recognition device 5, and completing original word slot filling. The original word slot filling completely reserves the information directly provided by the user or the information stored in the service system. The whole filling process comprises the following two steps:
in a first step, the call intention recognition means 5 performs element extraction from the original input of the user.
Examples are as follows:
the user requests: recharge for mobile phones of nine, two, three, four, five, six and eight
Original word slot: mobile phone number
And (3) element extraction result: yao nine two three four five seven eight nine
And secondly, finishing customized information extraction or conversion through standardization processing.
Example 1:
original information: yao nine two three four five seven eight nine
And (3) standardization treatment: conversion to Arabic numeral representation
And (3) element extraction result: 19912345678
Example 2:
original information: 1.2345
And (3) standardization treatment: reserved 2 decimal place
And (3) element extraction result: 1.23
The mapping word slot filling unit 303 is responsible for acquiring mapping word slot configuration information to be filled from the scene configuration information. The mapping word slot configuration information contains the mapping relation with the original word slot, and the filling of the mapping word slot is completed according to the value of the original word slot. The mapping word slot is the deformation and processing of the original word slot, is mostly used in the answering art, and plays a role in making the conversation smooth and natural. For example, the robot knows that the real name of the client is euonyna and gender, the information is stored in the conversation context through the original word slot, and if the client is called as miss-ohmits in the next conversation, the mapping word slot can be configured for further extracting surname information and defining a nickname appellation.
The overall filling process of the mapping word slot comprises the following two steps:
firstly, acquiring configuration information which comprises a mapping word slot list and a mapping relation;
secondly, calculating a mapping word slot value of the mapping word slot according to the configuration information in the first step;
and step three, calling a standardization processing unit to calculate the final value of the mapping word slot.
And the standardization processing unit 304 is responsible for realizing various unified logics. According to business or technical needs, the data can be processed by encapsulating some functions, so that the effects of facilitating understanding of users, unifying formats, facilitating subsequent processing (such as input of a model) and the like are achieved.
Fig. 9 is a schematic configuration diagram of the information post-processing apparatus 4. As shown in fig. 9, the information post-processing apparatus 4 includes a flow control unit 401 and a response construction unit 402.
The process control unit 401 is responsible for determining the process trend according to the word slot filling result and completing the transition of the session state.
The response constructing unit 402 is responsible for acquiring corresponding configuration information according to the result after the flow jump, and generating a response message (i.e. service reply information).
Fig. 10 is a schematic view of the structure of the intention identifying device 5. As shown in fig. 10, the intention recognition apparatus 5 may include a model management unit 501, a model service unit 502.
The model management unit 501 is responsible for management operations such as model training and publishing.
And the model service unit 502 is responsible for responding to the model calling request and giving a model processing result.
Fig. 11 is a schematic configuration diagram of the data storage device 6. As shown in fig. 11, the data storage device 6 includes a database storage unit 601 and a cache storage unit 602.
The database storage unit 601 is responsible for providing database read-write services.
And the cache storage unit 602 is responsible for providing cache read-write service.
Based on the above device example, the present specification also provides an example of a session information generation method. As shown in fig. 12, the business processing flow is described by taking the credit card application of the user as an example, and includes the following steps:
in step 701, data configuration related to multiple rounds of sessions, such as word slots and flows (including nodes and jump conditions), is completed in advance through the configuration management device 1.
Assume that the application credit card scenario configures the following word slot:
| word slot mark | Word groove type | Information extraction | Normalization process |
| Name (I) | Original word slot | Obtaining a user name | Air conditioner |
| Sex | Original word slot | Obtaining user gender | Air conditioner |
| Title to be called | Mapping word slot | Air conditioner | Air conditioner |
| Surname family name | Mapping word slot | Air conditioner | Extracting surname |
And configuring the following mapping relations:
| original word slot mark | Mapping word slot identification | Original word slot value | Mapping word slot values |
| Name (I) | Surname family name | Air conditioner | Air conditioner |
| Sex | Title to be called | For male | Mr. first |
| Sex | Title to be called | Woman | Female |
The user: you good, i want to work with a credit card.
The service robot inquires the basic information of the user according to the incoming call number: the name of the user is Zhang III, the sex is male, and the identification number is XXXX.
And 703, for the brand-new session, completing user intention analysis through the intention identification device 5, and returning a result to enter a corresponding service scene according to the model.
The intelligent customer service robot calls the intention recognition device 5 to judge that the current service scene is a credit card application.
The service robot calls the data storage device 6 to acquire the first node configuration of the business scene of applying for the credit card, namely, the identity of the user is confirmed, the answer configuration is good, and the question is asked that the name is $ title.
The service robot calls the word slot filling device 3 to obtain the word slot for confirming the binding of the user identity node: name, gender.
The service robot fills the word slot according to the word slot configuration in step 701, and the result of filling the word slot is as follows: name is Zhang III, gender is male.
And 707, acquiring a corresponding mapping word slot list according to the configuration information of each word slot in the list, and completing the filling of the mapping word slot according to the configuration of the mapping relation.
The service robot is configured according to the mapping relation table to complete the first-stage filling of the mapping word slot, namely, the surname is Zhang III and the name is Mr. the service robot is configured according to the mapping relation table; and then, according to the configuration of the word slot table, finishing the filling of the second stage of the mapping word slot, namely, the surname is named as Mr.
In step 708, the information post-processing device 4 completes branch jump judgment according to the word slot filling result and the current node configuration, jumps to the target node associated with the reply word technique, replaces the word slot identification name included in the reply word technique with a word slot value, and generates a response message.
The service robot: you do you ask you about how your is mr. Zhang III.
The embodiment can improve the flexibility of the service robot in answering, and can quickly and accurately position the required answering dialogues based on the conversation circulation information and improve the accuracy of the answering information generation. And based on the word slot associated with the session node involved in the session transfer process, the word slot value required subsequently is quickly extracted, and the efficiency of generating subsequent reply information is improved.
The present specification also provides a service robot that may include at least one processor and a memory for storing processor-executable instructions that, when executed by the processor, perform steps comprising the method of any one or more of the embodiments described above. The memory may include physical means for storing information, typically by digitizing the information for storage on a medium using electrical, magnetic or optical means. The storage medium may include: devices that store information using electrical energy, such as various types of memory, e.g., RAM, ROM, etc.; devices that store information using magnetic energy, such as hard disks, floppy disks, tapes, core memories, bubble memories, and usb disks; devices that store information optically, such as CDs or DVDs. Of course, there are other ways of storing media that can be read, such as quantum memory, graphene memory, and so forth.
It should be noted that the embodiments of the present disclosure are not limited to the cases where the data model/template is necessarily compliant with the standard data model/template or the description of the embodiments of the present disclosure. Certain industry standards, or implementations modified slightly from those described using custom modes or examples, may also achieve the same, equivalent, or similar, or other, contemplated implementations of the above-described examples. The embodiments using these modified or transformed data acquisition, storage, judgment, processing, etc. may still fall within the scope of the alternative embodiments of the present description.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment. In the description of the specification, reference to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the specification. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.
The above description is only an example of the present specification, and is not intended to limit the present specification. Various modifications and alterations to this description will become apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present specification should be included in the scope of the claims of the present specification.
Claims (10)
1. A session information generation method applied to a service robot, the method comprising:
receiving service requirement information of a user;
calling a reply dialect corresponding to the service demand information and a word slot identifier in the reply dialect; the word slot mark at least comprises an original word slot mark corresponding to an original word slot and a mapping word slot mark corresponding to a mapping word slot; the original word slot is a word slot filled based on the service demand information and the corresponding associated service information; the mapping word slot is a word slot filled based on the associated original word slot and a preset word slot filling mode;
extracting an original word slot value corresponding to the original word slot identifier in the reply grammar based on the service demand information and the corresponding associated service information;
extracting a mapping word slot value corresponding to the corresponding mapping word slot identifier based on the original word slot value associated with the mapping word slot identifier in the answering operation and a preset word slot filling mode;
correspondingly filling the original word slot value and the mapping word slot value into the positions of the original word slot identifier and the mapping word slot identifier of the reply grammar to generate service reply information corresponding to the service demand information;
and feeding back the service reply information to the user.
2. The method of claim 1, further comprising:
acquiring session context information corresponding to a session identifier of the service demand information, wherein the session context information at least comprises an extracted word slot value and a corresponding word slot identifier under the session identifier;
correspondingly, an original word slot value corresponding to the original word slot identifier or a mapping word slot value corresponding to the mapping word slot identifier in the reply sentence is extracted by combining the context information of the conversation.
3. The method according to claim 1, wherein the retrieving the answer dialog corresponding to the service requirement information comprises:
acquiring session transfer information of a service scene to which the service demand information belongs; the session transfer information comprises at least one session node and a transfer relationship between the session nodes; the session node is used for representing different processing links in the service processing process of the corresponding service scene; at least part of the session nodes are associated with reply dialogs related to corresponding processing links;
determining a session node which is corresponding to the service demand information in the session flow information and is associated with a reply dialog as a target session node based on the intention recognition result of the service demand information;
and taking the reply dialect associated with the target session node as the reply dialect corresponding to the service requirement information.
4. The method according to claim 3, wherein at least some session nodes in the session transfer information are associated with word slot identifiers related to corresponding processing links;
accordingly, whether the circulation condition satisfies the word slot value determination filled with the word slot identifier associated with the session node.
5. The method of claim 3, further comprising:
acquiring a streamed session node under the session identifier of the service demand information;
correspondingly, the session node which is associated with the reply dialog and corresponds to the service requirement information in the session circulation information is determined based on the circulated session node and the intention identification result of the service requirement information.
6. The method according to claim 1, wherein the original word slot value and the mapped word slot value are normalized, so that the normalized original word slot value and the normalized mapped word slot value are correspondingly filled into positions where the original word slot identifier and the mapped word slot identifier of the answer grammar are located; wherein the normalization process is constructed based on a pre-packaged normalization process function.
7. A session information generation apparatus applied to a service robot, the apparatus comprising:
the receiving module is used for receiving the service requirement information of the user;
the calling module is used for extracting a reply dialect corresponding to the service demand information and a word slot identifier in the reply dialect; the word slot mark at least comprises an original word slot mark corresponding to an original word slot and a mapping word slot mark corresponding to a mapping word slot; the original word slot is a word slot filled based on the service demand information and the corresponding associated service information; the mapping word slot is a word slot filled based on the associated original word slot and a preset word slot filling mode;
a word slot value extraction module, configured to determine, based on the service requirement information and the corresponding associated service information, an original word slot value corresponding to the original word slot identifier in the reply grammar; determining a mapping word slot value corresponding to the corresponding mapping word slot identifier based on the original word slot value associated with the mapping word slot identifier in the reply technology and a preset word slot filling mode;
the filling module is used for correspondingly filling the original word slot value and the mapping word slot value into the positions of the original word slot identifier and the mapping word slot identifier of the reply speech to generate service reply information corresponding to the service demand information;
and the feedback module is used for feeding back the service reply information to the user.
8. The apparatus of claim 7, wherein the word-bin-value extracting module comprises:
a first obtaining unit, configured to obtain session context information corresponding to a session identifier of the service demand information, where the session context information at least includes a word slot value extracted under the session identifier and a corresponding word slot identifier;
and the extracting unit is used for extracting an original word slot value corresponding to the original word slot identifier or a mapping word slot value corresponding to the mapping word slot identifier in the reply language by combining the context information of the conversation.
9. The apparatus of claim 7, wherein the retrieving module comprises:
the second obtaining unit is used for obtaining the session transfer information of the service scene to which the service demand information belongs; the session transfer information comprises at least one session node and a transfer relationship between the session nodes; the session node is used for representing different processing links in the service processing process of the corresponding service scene; at least part of the session nodes are associated with reply dialogs related to corresponding processing links;
a session node determination unit, configured to determine, based on an intention recognition result of the service demand information, a session node associated with a reply dialog corresponding to the service demand information in the session flow information, as a target session node;
a reply-to-talk determining unit, configured to use the reply talk associated with the target session node as the reply talk corresponding to the service requirement information.
10. A service robot, characterized in that it comprises at least one processor and a memory for storing processor-executable instructions, which when executed by the processor implement the steps of the method of any of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110479849.1A CN113192505B (en) | 2021-04-30 | 2021-04-30 | Session information generation method and device and service robot |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110479849.1A CN113192505B (en) | 2021-04-30 | 2021-04-30 | Session information generation method and device and service robot |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113192505A true CN113192505A (en) | 2021-07-30 |
| CN113192505B CN113192505B (en) | 2023-05-02 |
Family
ID=76982978
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110479849.1A Active CN113192505B (en) | 2021-04-30 | 2021-04-30 | Session information generation method and device and service robot |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113192505B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116319630A (en) * | 2023-03-07 | 2023-06-23 | 北京奇艺世纪科技有限公司 | Message reply method and device, electronic equipment and readable storage medium |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170195486A1 (en) * | 2015-12-30 | 2017-07-06 | Shanghai Xiaoi Robot Technology Co., Ltd. | Intelligent customer service systems, customer service robots, and methods for providing customer service |
| US20180182380A1 (en) * | 2016-12-28 | 2018-06-28 | Amazon Technologies, Inc. | Audio message extraction |
| CN109979457A (en) * | 2019-05-29 | 2019-07-05 | 南京硅基智能科技有限公司 | A method of thousand people, thousand face applied to Intelligent dialogue robot |
| CN110188177A (en) * | 2019-05-28 | 2019-08-30 | 北京搜狗科技发展有限公司 | Talk with generation method and device |
| CN110377720A (en) * | 2019-07-26 | 2019-10-25 | 中国工商银行股份有限公司 | The more wheel exchange methods of intelligence and system |
| CN111813900A (en) * | 2019-04-10 | 2020-10-23 | 北京猎户星空科技有限公司 | Multi-turn conversation processing method and device, electronic equipment and storage medium |
| CN111881271A (en) * | 2020-07-08 | 2020-11-03 | 泰康保险集团股份有限公司 | Method and device for realizing automatic conversation |
| US20210110822A1 (en) * | 2019-10-11 | 2021-04-15 | Tata Consultancy Services Limited | Conversational systems and methods for robotic task identification using natural language |
-
2021
- 2021-04-30 CN CN202110479849.1A patent/CN113192505B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170195486A1 (en) * | 2015-12-30 | 2017-07-06 | Shanghai Xiaoi Robot Technology Co., Ltd. | Intelligent customer service systems, customer service robots, and methods for providing customer service |
| US20180182380A1 (en) * | 2016-12-28 | 2018-06-28 | Amazon Technologies, Inc. | Audio message extraction |
| CN111813900A (en) * | 2019-04-10 | 2020-10-23 | 北京猎户星空科技有限公司 | Multi-turn conversation processing method and device, electronic equipment and storage medium |
| CN110188177A (en) * | 2019-05-28 | 2019-08-30 | 北京搜狗科技发展有限公司 | Talk with generation method and device |
| CN109979457A (en) * | 2019-05-29 | 2019-07-05 | 南京硅基智能科技有限公司 | A method of thousand people, thousand face applied to Intelligent dialogue robot |
| CN110377720A (en) * | 2019-07-26 | 2019-10-25 | 中国工商银行股份有限公司 | The more wheel exchange methods of intelligence and system |
| US20210110822A1 (en) * | 2019-10-11 | 2021-04-15 | Tata Consultancy Services Limited | Conversational systems and methods for robotic task identification using natural language |
| CN111881271A (en) * | 2020-07-08 | 2020-11-03 | 泰康保险集团股份有限公司 | Method and device for realizing automatic conversation |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116319630A (en) * | 2023-03-07 | 2023-06-23 | 北京奇艺世纪科技有限公司 | Message reply method and device, electronic equipment and readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113192505B (en) | 2023-05-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110377720B (en) | Intelligent multi-round interaction method and system | |
| CN108446322A (en) | A kind of implementation method and device of intelligent Answer System | |
| CN112202978A (en) | Intelligent outbound call system, method, computer system and storage medium | |
| US8417523B2 (en) | Systems and methods for interactively accessing hosted services using voice communications | |
| US10698934B1 (en) | Interface between a virtual personal assistant and an interactive voice response system | |
| CN109325091A (en) | Update method, device, equipment and the medium of points of interest attribute information | |
| US11727923B2 (en) | System and method for virtual conversations | |
| CN112131358A (en) | Scene flow structure and intelligent customer service system applied by same | |
| CN109582760A (en) | A kind of response flow engine system applied in Intelligent dialogue system | |
| CN112507092A (en) | Implementation method for completing work order intelligent filling function based on AI text robot | |
| CN110266900A (en) | Recognition methods, device and the customer service system that client is intended to | |
| CN111462726B (en) | Method, device, equipment and medium for answering out call | |
| CN114186036A (en) | Dialogue processing method, device, computer equipment and storage medium | |
| CN115982331A (en) | Information interaction method, device and equipment in session scene | |
| CN112035630A (en) | Dialogue interaction method, device, equipment and storage medium combining RPA and AI | |
| CN113192505B (en) | Session information generation method and device and service robot | |
| CN112801721B (en) | Information processing method, information processing device, electronic equipment and storage medium | |
| CN111666388B (en) | Dialogue data processing method, device, computer equipment and storage medium | |
| CN112231027A (en) | Task type session configuration method and system | |
| CN116860938A (en) | Voice question-answering construction method, device and medium based on large language model | |
| CN110047473A (en) | A kind of man-machine collaboration exchange method and system | |
| CN104468926A (en) | Method and device for controlling contact persons in mobile terminal | |
| CN111556096B (en) | Information pushing method, device, medium and electronic equipment | |
| CN111930917B (en) | Conversation process mining method and device, computer equipment and storage medium | |
| CN113782022B (en) | Communication method, device, equipment and storage medium based on intention recognition model |
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 |