CN112559721B

CN112559721B - Method, device, equipment, medium and program product for adjusting man-machine dialogue system

Info

Publication number: CN112559721B
Application number: CN202011566521.5A
Authority: CN
Inventors: 付波; 褚晓梅
Original assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Current assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Priority date: 2020-12-25
Filing date: 2020-12-25
Publication date: 2023-10-20
Anticipated expiration: 2040-12-25
Also published as: CN112559721A

Abstract

The disclosure provides a method, a device, an electronic device, a computer readable storage medium and a computer program product for adjusting a man-machine conversation system, which relate to the technical field of artificial intelligence, in particular to the fields of big data, intelligent recommendation and natural language processing. The implementation scheme is as follows: acquiring a dialogue log associated with a man-machine dialogue system; based on the dialog log, a node state jump map is created. The node state jump graph defines a plurality of dialog nodes and jump relationships between the plurality of dialog nodes, at least one of the plurality of dialog nodes representing a respective interaction page of the human-machine dialog system. And calculating weight parameters of at least one dialogue node based on the node state jump graph, and executing corresponding adjustment operation on the interaction page represented by the at least one dialogue node based on the weight parameters of the at least one dialogue node.

Description

Method, device, equipment, medium and program product for adjusting man-machine dialogue system

Technical Field

The present disclosure relates to the field of artificial intelligence, and in particular to the field of big data, intelligent recommendation, and natural language processing. In particular, the present disclosure provides a method, apparatus, electronic device, computer readable storage medium and computer program product for tuning a human-machine dialog system.

Background

Artificial intelligence is the discipline of studying the process of making a computer mimic certain mental processes and intelligent behaviors (e.g., learning, reasoning, thinking, planning, etc.) of a person, both hardware-level and software-level techniques. Artificial intelligence hardware technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing, etc.: the artificial intelligence software technology mainly comprises a computer vision technology, a voice recognition technology, a natural language processing technology, a machine learning/deep learning technology, a big data processing technology, a knowledge graph technology and the like.

At present, man-machine conversation systems based on artificial intelligence in the market are presented. In some dialogue systems that interact with users in a node click jump mode, a developer can complete the development of a man-machine dialogue system by simply configuring dialogue nodes. After the human-machine interaction system is put into use, how to evaluate and adjust the human-machine interaction system is particularly important.

In the related art, there are also limitations on the evaluation scheme of the human-computer dialogue system and there is also a great room for improvement in the evaluation effect.

The approaches described in this section are not necessarily approaches that have been previously conceived or pursued. Unless otherwise indicated, it should not be assumed that any of the approaches described in this section qualify as prior art merely by virtue of their inclusion in this section. Similarly, the problems mentioned in this section should not be considered as having been recognized in any prior art unless otherwise indicated.

Disclosure of Invention

The present disclosure provides a method, apparatus, electronic device, computer-readable storage medium, and computer program product for tuning a human-machine dialog system.

According to an aspect of the present disclosure, there is provided a method for adjusting a human-machine conversation system, including: acquiring a dialogue log associated with a man-machine dialogue system; based on the dialog log, a node state jump map is created. The node state jump graph defines a plurality of dialog nodes and jump relationships between the plurality of dialog nodes, at least one of the plurality of dialog nodes representing a respective interaction page of the human-machine dialog system. And calculating weight parameters of at least one dialogue node based on the node state jump graph, and executing corresponding adjustment operation on the interaction page represented by the at least one dialogue node based on the weight parameters of the at least one dialogue node.

According to another aspect of the present disclosure, there is provided an adjusting apparatus of a human-machine conversation system, including: and an acquisition unit configured to acquire a conversation log associated with the man-machine conversation system. And a creating unit configured to create a node state jump graph based on the dialogue log. The node state jump graph defines a plurality of dialog nodes and jump relationships between the plurality of dialog nodes, at least one of the plurality of dialog nodes representing a respective interaction page of the human-machine dialog system. The system comprises a calculation unit, an adjustment unit and a control unit, wherein the calculation unit is configured to calculate weight parameters of at least one dialogue node based on a node state jump graph, and the adjustment unit is configured to execute corresponding adjustment operation on an interaction page represented by the at least one dialogue node based on the weight parameters of the at least one dialogue node.

According to another aspect of the present disclosure, there is provided a computer apparatus comprising: memory, a processor, and a computer program stored on the memory. Wherein the processor is configured to execute a computer program to implement the steps of the method of tuning a human-machine dialog system described above.

According to another aspect of the present disclosure, a non-transitory computer readable storage medium having a computer program stored thereon is provided. Wherein the computer program when executed by the processor implements the steps of the method for adjusting a human-machine interaction system described above.

According to another aspect of the present disclosure, a computer program product is provided, including a computer program. Wherein the computer program when executed by the processor implements the steps of the method for adjusting a human-machine interaction system described above.

The technical scheme provided by the embodiment of the disclosure has the beneficial technical effects that at least:

the dialogue flow of the man-machine dialogue system is evaluated and adjusted from the global angle, and the dialogue system is automatically adjusted according to the man-machine dialogue data, so that the labor cost of operation is saved.

It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the disclosure, nor is it intended to be used to limit the scope of the disclosure. Other features of the present disclosure will become apparent from the following specification.

Drawings

The accompanying drawings illustrate exemplary embodiments and, together with the description, serve to explain exemplary implementations of the embodiments. The illustrated embodiments are for exemplary purposes only and do not limit the scope of the claims. Throughout the drawings, identical reference numerals designate similar, but not necessarily identical, elements.

FIG. 1 illustrates a schematic diagram of an exemplary system in which various methods described herein may be implemented, in accordance with an embodiment of the present disclosure;

FIG. 2 illustrates a flow chart of a method of tuning a human-machine conversation system in accordance with an embodiment of the present disclosure;

FIG. 3 illustrates a flowchart of example operations for creating a node state jump map in the method of FIG. 2, according to embodiments of the present disclosure;

FIG. 4 illustrates a schematic diagram of example operations for creating a node state jump map in the method of FIG. 3, according to embodiments of the present disclosure;

FIG. 5 illustrates a schematic diagram of a structure of a node state jump graph, according to an embodiment of the present disclosure;

FIG. 6 illustrates a flowchart of example operations for calculating weight parameters of a dialog node in the method of FIG. 2, according to embodiments of the present disclosure;

FIG. 7 illustrates a schematic diagram of an example operation of establishing a transition probability distribution between a plurality of dialog nodes for the node state transition diagram of FIG. 5 according to the method of FIG. 6, in accordance with an embodiment of the present disclosure;

FIG. 8 illustrates a flowchart of example operations for modeling a decision process in the method of FIG. 6, according to embodiments of the present disclosure;

FIG. 9 illustrates a schematic diagram of example operations for modeling a decision process in the method of FIG. 8, in accordance with embodiments of the present disclosure;

FIG. 10 illustrates a flowchart of example operations for calculating a contribution value of a dialog node in the method of FIG. 6, according to embodiments of the present disclosure;

FIG. 11 illustrates a flowchart of example operations for adjusting a dialog node in the method of FIG. 2, according to embodiments of the present disclosure;

FIG. 12 illustrates a flowchart of example operations for locating a dialog node in the method of FIG. 11 that requires an adjustment operation, in accordance with an embodiment of the present disclosure;

13A-13C illustrate a schematic diagram of an example system being adjusted by an adjustment method of a human-machine conversation system, according to an embodiment of the present disclosure;

FIG. 14 illustrates a flowchart of example operations for obtaining a dialog log for a dialog system in the method of FIG. 2, according to embodiments of the present disclosure;

FIG. 15 illustrates a flowchart of overall operation of a method of tuning a human-machine conversation system, in accordance with an embodiment of the present disclosure;

FIG. 16 illustrates a block diagram of a human-machine dialog system adjustment device, according to an embodiment of the disclosure;

Fig. 17 illustrates a block diagram of an exemplary electronic device that can be used to implement embodiments of the present disclosure.

Detailed Description

Exemplary embodiments of the present disclosure are described below in conjunction with the accompanying drawings, which include various details of the embodiments of the present disclosure to facilitate understanding, and should be considered as merely exemplary. Accordingly, one of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope of the present disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.

In the present disclosure, the use of the terms "first," "second," and the like to describe various elements is not intended to limit the positional relationship, timing relationship, or importance relationship of the elements, unless otherwise indicated, and such terms are merely used to distinguish one element from another. In some examples, a first element and a second element may refer to the same instance of the element, and in some cases, they may also refer to different instances based on the description of the context.

The terminology used in the description of the various illustrated examples in this disclosure is for the purpose of describing particular examples only and is not intended to be limiting. Unless the context clearly indicates otherwise, the elements may be one or more if the number of the elements is not specifically limited. Furthermore, the term "and/or" as used in this disclosure encompasses any and all possible combinations of the listed items.

In the related art, evaluation and adjustment of the man-machine conversation system are based on a conversion rate of a target link or a next hop rate of a conversation node to evaluate interaction of a user with the man-machine conversation system. However, the schemes in the related art cannot determine the effect of other dialog nodes on the conversion effect of the system, or can only measure the stay preference of the user on the man-machine dialog system. Furthermore, in the related art, the optimization of the man-machine conversation system consumes a lot of manpower to adjust the conversation nodes, and if the external environment changes, a lot of manpower needs to be continuously input to re-optimize.

In order to solve the above problems in the related art, the present disclosure provides the following technical solutions of the adjustment method, based on the man-machine conversation log in big data, to evaluate and automatically adjust the man-machine conversation system from the global perspective.

Embodiments of the present disclosure will be described in detail below with reference to the accompanying drawings.

Fig. 1 illustrates a schematic diagram of an exemplary system 100 in which various methods and apparatus described herein may be implemented, in accordance with an embodiment of the present disclosure. Referring to fig. 1, the system 100 includes one or more client devices 101, 102, 103, 104, 105, and 106, a server 120, and one or more communication networks 110 coupling the one or more client devices to the server 120. Client devices 101, 102, 103, 104, 105, and 106 may be configured to execute one or more applications.

In an embodiment of the present disclosure, the server 120 may run one or more services or software applications that enable execution of the tuning method of the man-machine conversation system of the present disclosure.

In some embodiments, server 120 may also provide other services or software applications that may include non-virtual environments and virtual environments. In some embodiments, these services may be provided as web-based services or cloud services, for example, provided to users of client devices 101, 102, 103, 104, 105, and/or 106 under a software as a service (SaaS) model.

In the configuration shown in fig. 1, server 120 may include one or more components that implement the functions performed by server 120. These components may include software components, hardware components, or a combination thereof that are executable by one or more processors. A user operating client devices 101, 102, 103, 104, 105, and/or 106 may in turn utilize one or more client applications to interact with server 120 to utilize the services provided by these components. It should be appreciated that a variety of different system configurations are possible, which may differ from system 100. Accordingly, FIG. 1 is one example of a system for implementing the various methods described herein and is not intended to be limiting.

A user may interact with the human-machine dialog system using client devices 101, 102, 103, 104, 105, and/or 106. The client device may provide an interface that enables a user of the client device to interact with the client device. The client device may also output information to the user via the interface. Although fig. 1 depicts only six client devices, those skilled in the art will appreciate that the present disclosure may support any number of client devices.

Client devices 101, 102, 103, 104, 105, and/or 106 may include various types of computer devices, such as portable handheld devices, general purpose computers (such as personal computers and laptop computers), workstation computers, wearable devices, gaming systems, thin clients, various messaging devices, sensors or other sensing devices, and the like. These computer devices may run various types and versions of software applications and operating systems, such as Microsoft Windows, apple iOS, UNIX-like operating systems, linux, or Linux-like operating systems (e.g., google Chrome OS); or include various mobile operating systems such as Microsoft Windows Mobile OS, iOS, windows Phone, android. Portable handheld devices may include cellular telephones, smart phones, tablet computers, personal Digital Assistants (PDAs), and the like. Wearable devices may include head mounted displays and other devices. The gaming system may include various handheld gaming devices, internet-enabled gaming devices, and the like. The client device is capable of executing a variety of different applications, such as various Internet-related applications, communication applications (e.g., email applications), short Message Service (SMS) applications, and may use a variety of communication protocols.

Network 110 may be any type of network known to those skilled in the art that may support data communications using any of a number of available protocols, including but not limited to TCP/IP, SNA, IPX, etc. For example only, the one or more networks 110 may be a Local Area Network (LAN), an ethernet-based network, a token ring, a Wide Area Network (WAN), the internet, a virtual network, a Virtual Private Network (VPN), an intranet, an extranet, a Public Switched Telephone Network (PSTN), an infrared network, a wireless network (e.g., bluetooth, WIFI), and/or any combination of these and/or other networks.

The server 120 may include one or more general purpose computers, special purpose server computers (e.g., PC (personal computer) servers, UNIX servers, mid-end servers), blade servers, mainframe computers, server clusters, or any other suitable arrangement and/or combination. The server 120 may include one or more virtual machines running a virtual operating system, or other computing architecture that involves virtualization (e.g., one or more flexible pools of logical storage devices that may be virtualized to maintain virtual storage devices of the server). In various embodiments, server 120 may run one or more services or software applications that provide the functionality described below.

The computing units in server 120 may run one or more operating systems including any of the operating systems described above as well as any commercially available server operating systems. Server 120 may also run any of a variety of additional server applications and/or middle tier applications, including HTTP servers, FTP servers, CGI servers, JAVA servers, database servers, etc.

In some implementations, server 120 may include one or more applications to analyze and consolidate data feeds and/or event updates received from users of client devices 101, 102, 103, 104, 105, and 106. Server 120 may also include one or more applications to display data feeds and/or real-time events via one or more display devices of client devices 101, 102, 103, 104, 105, and 106.

In some implementations, the server 120 may be a server of a distributed system or a server that incorporates a blockchain. The server 120 may also be a cloud server, or an intelligent cloud computing server or intelligent cloud host with artificial intelligence technology. The cloud server is a host product in a cloud computing service system, so as to solve the defects of large management difficulty and weak service expansibility in the traditional physical host and virtual private server (VPS, virtual Private Server) service.

The system 100 may also include one or more databases 130. In some embodiments, these databases may be used to store data and other information. For example, one or more of databases 130 may be used to store information such as audio files and video files. The data store 130 may reside in a variety of locations. For example, the data store used by the server 120 may be local to the server 120, or may be remote from the server 120 and may communicate with the server 120 via a network-based or dedicated connection. The data store 130 may be of different types. In some embodiments, the data store used by server 120 may be a database, such as a relational database. One or more of these databases may store, update, and retrieve the databases and data from the databases in response to the commands.

In some embodiments, one or more of databases 130 may also be used by applications to store application data. The databases used by the application may be different types of databases, such as key value stores, object stores, or conventional stores supported by the file system.

The system 100 of fig. 1 may be configured and operated in various ways to enable application of the various methods and apparatus described in accordance with the present disclosure.

Fig. 2 shows a flowchart of a method of tuning a man-machine conversation system, according to an embodiment of the present disclosure. As shown in fig. 2, the method 200 for adjusting a human-machine conversation system includes:

at step 210, a conversation log associated with a human-machine conversation system is obtained. For example, the human-machine dialog system may be a dialog marketing robot that employs a pattern of node click jumps and/or voice command jumps for dialog interactions with a user. Illustratively, the dialog log may include the number of user visits to the human-machine dialog system, which may include, but is not limited to, the number of individual visitors (UV), the number of individual IPs (IP), the number of visits (VV), the amount of Page Views (PV), etc.

At step 220, a node state jump map is created based on the dialog log. The node state jump graph defines a plurality of dialog nodes and jump relationships between the plurality of dialog nodes, at least one of the plurality of dialog nodes representing a respective interaction page of the human-machine dialog system. Illustratively, a user may jump from an interaction page represented by one dialog node to an interaction page represented by a next dialog node during interaction with the man-machine dialog system.

In step 230, weight parameters for at least one dialog node are calculated based on the node state jump map. In some exemplary embodiments, the node state jump map as created in step 220 may represent a behavior profile of user interactions with the human-machine dialog system. Alternatively, the node state jump graph may represent a sequential decision model, and the model may be used to solve for the weight parameters of each dialog node. The weight parameter may represent an impact value or contribution of the dialog node to the decision objective.

In step 240, a corresponding adjustment operation is performed on the interaction page represented by the at least one dialog node based on the weight parameter of the at least one dialog node. For example, the dialogue nodes may be adjusted such that the interaction page or the corresponding interaction portal represented by the dialogue node with the high weight parameter is presented to the user first, thereby increasing the probability of letting the user access the interaction page or the interaction portal.

In summary, the method provided by the embodiment of the present disclosure generates the node state jump map based on the interaction situation between the user and the interaction page represented by each dialogue node in the man-machine dialogue system. And the weight parameter of each dialogue node in the global estimated node state jump diagram is further optimized by executing corresponding adjustment operation on the interaction page according to the weight parameter of the dialogue node, so that the probability of the user accessing the interaction page represented by the dialogue node with high weight parameter or the corresponding interaction entrance is higher. In addition, the embodiment of the disclosure utilizes the node state jump diagram to carry out global evaluation, and considers the conversion condition of each dialogue node to the target link under the global condition, so that the evaluation result and the optimization scheme are more objective, and the dialogue nodes can be adjusted and optimized from the whole angle.

Fig. 3 illustrates a flowchart of example operations for creating a node state jump map (step 220) in the method 200 of fig. 2, according to embodiments of the present disclosure. As shown in fig. 3, step 220 further includes steps 310 and 320 based on the dialog log.

At step 310, the number of individual users accessing each interaction page of the human-machine dialog system and the operations performed by the individual users for each interaction page may be counted based on the dialog log. In some exemplary embodiments, an independent visitor count (UV) for each interaction page may be counted. The operation performed by the independent user for each interaction page may be, for example, an operation performed by the user to access a target link (also referred to as "conversion" in this context) on the interaction page, or an operation performed by the user to leave the current page without any interaction on the interaction page.

In step 320, a node state jump graph may be generated based on the number of independent users accessing each interactive page of the man-machine conversation system and the operations performed by the independent users for each interactive page, where the jump relationships between the plurality of conversation nodes and the plurality of conversation nodes are generated. In some exemplary embodiments, the skip relationship between dialog nodes may include a skip direction (i.e., from which dialog node to skip) and a UV number to skip from a current dialog node to a next dialog node.

Fig. 4 illustrates a schematic diagram of example operations for creating a node state jump map in the method of fig. 3, according to embodiments of the present disclosure. As shown in fig. 4, the node state jump diagram 400 includes dialog nodes 410, 420, 430, 440, 450, and 460. Dialog nodes 410 and 420 represent interaction pages 410' and 420', respectively, in human-machine dialog system 400 '. Likewise, dialog nodes 430 through 460 each represent a corresponding interaction page (not shown). According to an exemplary embodiment of the present disclosure, when the user clicks the button "apply credit card" or speaks "apply credit card" on the interactive page 410', the dialog node 410 will jump to the dialog node 420. The human-machine dialog system 400 'presents the interactive page 420' represented by the dialog node 420 to the user.

According to some example embodiments, edge 412 connecting dialog nodes 410 and 420 may represent a jump from dialog node 410 to dialog node 420. Illustratively, the UV number of the user jumping from the interaction page 410 'to the interaction page 420' can be obtained by counting the user's behavior on the interaction page 410' based on the man-machine interaction log, thereby generating the edge 412 representing the jump from the interaction node 410 to the interaction node 420. Further, the overall node state jump graph 400 may be generated by counting the UV numbers 414 that jump from the dialog node 410 to the dialog node 430, and the UV numbers 422, 424, and 426 that jump from the dialog node 420 to 440, 450, and 460, respectively.

In summary, according to the embodiments of the present disclosure, by generating the node state jump map to describe the dialogue node representing the interaction page and the edge representing the interaction behavior of the user, the node state jump map can abstract the behavior of the user on the interaction page into the dialogue node jump map, so as to fully reflect the interaction behavior of the user and the man-machine dialogue system, and provide a basis for global evaluation optimization.

Fig. 5 shows a schematic diagram of the structure of a node state jump diagram 500 according to an embodiment of the present disclosure. As shown in fig. 5, a node state jump graph 500 may be generated by the process shown in fig. 4, including session nodes 510, 520, 530 and session nodes 541-548. According to example operations as shown in fig. 4, the number of individual user accesses from session node 510 to session node 541 may be counted based on the session log. As shown in fig. 5, the number of accesses from session node 510 to session node 541 is 1890, i.e., there is a behavior in which 1890 individual users jump to session node 541 on session node 510.

Further exemplary, there are 6 behaviors of dialog node 542, corresponding to: 3 individual users from dialog node 542 to dialog node 541, 18 individual users from dialog node 542 to dialog node 530, 51 individual users from dialog node 542 to dialog node 545, 112 individual users from dialog node 542 to dialog node 543, 30 individual users from dialog node 542 to dialog node 546, and 105 individual users from dialog node 542 to dialog node 547. Similarly, the overall node state jump graph 500 is generated by counting the user's behavior on each dialog node.

In some demonstrative embodiments, the plurality of dialog nodes in node state jump diagram 500 may include a transformation node 520, transformation node 520 representing a user's behavior to access the target link on one or more interaction pages of the human-machine dialog system.

In other exemplary embodiments, the plurality of dialog nodes in the node state jump diagram 500 may also include an initial node 510 and an end node 530. The initial node 510 may represent the behavior of a user entering the human-machine interaction system from that node, and the end node 530 may represent the behavior of a user exiting directly without a jump operation in the other dialog nodes.

In summary, on the basis of each dialog node representing the corresponding interaction page, the node state jump diagram may more completely describe the behavior distribution of the user in the human-computer dialog system by adding the conversion node 520, the initial node 510 and the end node 530 to the node state jump diagram.

Fig. 6 illustrates a flowchart of example operations for calculating weight parameters for an opposite node (step 230) in the method 200 of fig. 2, according to an embodiment of the present disclosure. As shown in fig. 6, step 230 further includes steps 610 through 630.

At step 610, a probability of a hop between a plurality of dialog nodes may be calculated based on the hop relationship between the plurality of dialog nodes. For example, the UV number of each session node may be summed and divided by the sum of the UV number of the next session node, respectively, to obtain the probability of a jump from that session node to the next session node.

At step 620, a decision process model may be built based on the probability of a jump between the plurality of dialog nodes. Illustratively, the decision process model may represent a model used to evaluate the impact of each dialog node on the user's behavior of accessing the target link.

At step 630, a contribution value of each of the at least one dialog node to the conversion node may be calculated as a weight parameter of each of the at least one dialog node based on the decision process model.

In summary, by establishing the decision process model based on the node state jump graph and the jump probability, the contribution value of each dialogue node to the user access target link behavior can be calculated, so that the dialogue nodes can be evaluated from the whole.

Fig. 7 illustrates a schematic diagram of example operations 700 for establishing a transition probability distribution between a plurality of dialog nodes for the node state transition diagram 500 of fig. 5 according to the method of fig. 6, in accordance with an embodiment of the present disclosure. As shown in fig. 7, the access numbers of all the individual users of the dialog node 542 in fig. 5 may be summed to obtain a total (3+18+51+30+112+105) =319. Based on the summation result 319, then, a probability of a jump between the dialog node 542 to the other dialog node can be calculated. Further, the hop probabilities for the dialog node 542 to hop to the dialog nodes 530, 541, 543, 545, 546 and 547 can be calculated to be 0.056, 0.009, 0.351, 0.160, 0.094, and 0.329, respectively.

Illustratively, the probability of a jump from dialog node 542 to 545 in example operation 700 may represent that the probability of a user selecting to jump to the interaction page represented by dialog node 545 on the interaction page represented by dialog node 542 when interacting with the human-machine dialog system is 0.160. The probability of a jump between the conversation node 545 and the conversion node 520 may represent a probability of 0.170 of a user's behavior (i.e., "conversion") to access the target link on the interaction page represented by the conversation node 545. Similarly, the dialog nodes 545, as well as all other dialog nodes shown in fig. 5, may be calculated and result in a probability distribution of transitions between a plurality of dialog nodes corresponding to the node state transition diagram 500 shown in fig. 5, which is not described in detail herein.

Fig. 8 illustrates a flowchart of example operations for modeling the decision process (step 620) in the method of fig. 6, according to embodiments of the present disclosure. As shown in fig. 8, step 620 further includes steps 810 through 840.

At step 810, a reward function for each of a plurality of dialog nodes may be obtained, the reward function for each of the plurality of dialog nodes being a desire for a reward value that the decision body can obtain at a next jump time when the decision body is at that dialog node.

At step 820, an attenuation coefficient may be obtained, the attenuation coefficient acting on the bonus function to obtain an attenuated bonus.

In step 830, a benefit function for each of the plurality of dialog nodes may be obtained, the benefit function being a summation of the attenuated rewards for all jump times from the current dialog node onwards.

In step 840, a model that measures the contribution of at least one dialog node to the conversion node may be built as a decision process model based on the reward function, the decay factor, the benefit function, and the probability of a jump between the plurality of dialog nodes.

Fig. 9 illustrates a schematic diagram of example operations for modeling a decision process in the method of fig. 8, according to an embodiment of the present disclosure. As shown in fig. 9, the decision process model 900 may include a decision body 910, a decision environment 920, a dialog node 930, a reward function 930 'corresponding to the dialog node 930, a dialog node 940 for the next time instance, and a reward function 940' and a jump behavior 950.

In some exemplary embodiments, decision body 910 may be a user and decision environment 920 may be a human-machine dialog system. Assuming the decision process model has markov properties, it can be determined by R _t A bonus function 930' representing a dialog node 930, and passing S _t Indicating that decision body 910 is at dialog node 930 at time t. Decision body 910 can interact with decision environment 920 at time t and enter next dialog node 940 at next time t+1, i.e. S, through selection behavior 950 _t+1 . Decision context 920 feeds back behavior 950 and rewards decision body 910 with function 940', i.e., R _t+1 . The goal of decision body 910 may beIs a function that maximizes the cumulative collection of rewards.

Illustratively, a reward function R at a dialog node s _S At a certain time t, the prize that can be obtained at the next time t+1 is:

R _S ＝E[R _t+1 |S _t ＝s]

the decay factor gamma may represent, for example, a value proportion of a future reward of the dialog node 930 at the current time t. Benefit function G _t The sum of rewards with decay coefficient y for all jump times from time t onward for the current dialog node 930 may be represented and may satisfy:

and further can obtain the contribution value v of the node 930 to the conversion node _s The method comprises the following steps:

v _s ＝E[G _t |S _t ＝s]

in summary, by the bonus function R _t Attenuation coefficient gamma, gain function G _t And the probability of jump, a decision process model 900 may be built. Since the decision process model 900 can measure the contribution v of at least one dialog node, such as dialog node 930, to the conversion node _s The contribution of dialog nodes to conversion nodes can be evaluated globally by solving the decision process model 900.

Fig. 10 illustrates a flowchart of example operations for calculating a contribution value of each of at least one dialog node to a conversion node in the method of fig. 6 (step 630) in accordance with an embodiment of the present disclosure. As shown in fig. 10, step 630 further includes steps 1010 through 1030.

At step 1010, a prize value may be assigned to each of the plurality of dialog nodes. Illustratively, the prize value may represent a transient prize that the decision-maker is able to obtain when entering the next dialog node.

In step 1020, a desire for a benefit function for each of a plurality of dialog nodes in the decision process model may be calculated based on the benefit function, a probability of a jump between the dialog node and other dialog nodes, and a reward value for the other dialog nodes.

In step 1030, the expectations of the respective benefit functions of the at least one dialog node may be taken as the respective contribution values of the at least one dialog node to the conversion node.

In some exemplary embodiments, the contribution value v _s Can be further rewritten as:

by means of the probability of a jump between the dialog nodes, it is further possible to obtain:

Where s' denotes a dialogue node other than the dialogue node s, p _ss′ The probability of a jump from session node s to session node s' is indicated.

Illustratively, the initial benefit of the conversion node may be set to 10 and the initial benefit of the end node to-3. Further, the contribution value v(s) of each dialogue node can be obtained by iterative calculation.

In summary, since the contribution value v(s) of each dialog node to the conversion node reflects the contribution degree of a certain dialog node to the final conversion node in the overall node state jump diagram, the contribution value obtained according to the calculation decision process model can be used as the weight parameter of the dialog node. By evaluating the position of the dialog node and the influence on the conversion node by means of the weight parameter, the dialog node can be better evaluated globally.

Fig. 11 illustrates a flowchart of example operations for adjusting a dialog node (step 240) in the method of fig. 2, according to embodiments of the present disclosure. As shown in fig. 11, step 240 further includes step 1110 and step 1120.

At step 1110, a dialog node that needs to be subjected to an adjustment operation may be searched for from the at least one dialog node based on the weight parameter of the at least one dialog node. Illustratively, the dialogue nodes can be ranked by the height of the weight parameter, and the dialogue nodes with changed ranking results are searched.

At step 1120, the dialog nodes that require adjustment may be cropped or reset such that the interaction page represented by the dialog nodes that require adjustment is removed or re-laid out accordingly.

In summary, the interaction page represented by the dialog node may be adjusted and optimized according to the weighting parameters of the dialog node solved by the decision process model. Because the weight parameters of the dialogue nodes represent the evaluation from the global dialogue nodes, the adjustment and optimization based on the weight parameters can ensure that the man-machine dialogue system is in an optimized state and keep a higher conversion effect.

Fig. 12 illustrates a flowchart of example operations for locating a dialog node that requires an adjustment operation in the method of fig. 11 (step 1110), in accordance with an embodiment of the present disclosure. As shown in fig. 12, step 1110 further includes steps 1210 through 1250.

At step 1210, a current first order of at least one dialog node may be obtained.

At step 1220, the at least one dialog node may be ranked according to a size of a weight parameter of the at least one dialog node.

At step 1230, a second ranking of at least one dialog node after the ranking may be obtained.

At step 1240, the first order bit and the second order bit of the at least one dialog node may be compared.

At step 1250, a dialog node that requires an adjustment operation may be determined based on the result of the comparison.

In summary, by sorting the dialogue nodes and comparing the sequence positions of the dialogue nodes before and after sorting, for example, the node with poor conversion effect of the dialogue node in the current man-machine dialogue system, or the node with better conversion effect but with rear sorting, etc. can be found. Thus, by adjusting these dialog nodes, the system can be guaranteed to be in a better high-conversion state.

In some exemplary embodiments, clipping the dialog nodes requiring the adjustment operation may include clipping the dialog node having the smallest second rank among the dialog nodes requiring the adjustment operation such that the interaction page represented by the dialog node having the smallest second rank is removed from the human-machine dialog system accordingly.

Because the weight parameter of the dialogue node represents the contribution value of the dialogue node to the conversion node, cutting out the dialogue node with the lowest contribution value can improve the conversion rate of the man-machine dialogue system on the whole.

13A-13C illustrate a schematic diagram of an example system being adjusted by an adjustment method of a human-machine conversation system, according to embodiments of the present disclosure. As shown in fig. 13A, the pre-adaptation human-machine dialog system 1300A includes dialog nodes 1310, 1320, 1330, 1340, and the like.

In some demonstrative embodiments, interaction pages 1310', 1320', 1330' and 1340, respectively, represented by at least one of dialog nodes 1310, 1320, 1330 and 1340, each interaction page includes at least one interaction portal, each interaction portal pointing to a next interaction page represented by a corresponding one of the at least one dialog nodes. Illustratively, dialog node 1310 may include interaction portals 1312, 1314, and 1316, and dialog node 1330 may include interaction portals 1332, 1334, 1336, and 1338. Illustratively, the interaction portal 1312 may point to the next interaction page 1320' represented by the dialog node 1320. Interaction portal 1338 can point to the next interaction page 1340' represented by dialog node 1340.

Illustratively, resetting the dialog node for which an adjustment operation is required may include:

for each of the dialog nodes that require an adjustment operation: and resetting at least one interactive entry in the dialogue node according to the ordering of the dialogue node corresponding to the interactive page pointed by at least one interactive entry in the interactive page represented by the dialogue node, so that the at least one interactive entry is presented according to the ordering result from top to bottom in the interactive page represented by the dialogue node.

It should be appreciated that the human-machine dialog system 1300A shown in fig. 13A is merely illustrative, and that the dialog system 1300A may include any number of dialog nodes.

Fig. 13B is a result 1300B of calculating and ordering respective weight parameters for each of at least one dialog node 1310, 1320, 1330, 1340, etc. in the human-machine dialog system 1300A in fig. 13A, according to a method of an embodiment of the disclosure. Like reference numerals in fig. 13B denote like elements as in fig. 13A, and are not repeated here.

As shown in fig. 13B, dialog node weight parameter ranking result 1300B may include weight parameters corresponding to dialog nodes 1310 through 1340. Illustratively, 1350 represents the weight parameters of dialog node 1340 and 1360 represents the weight parameters of dialog node 1320.

Illustratively, for the human-machine conversation system 1300A, the conversation nodes 1310 through 1340 appear in sequence. After calculating the weight parameters of each dialog node, the result of reordering the dialog nodes is shown in 1300B. The order of the dialog nodes 1310 to 1340 ordered according to the weight parameter size is changed to: from top to bottom, session nodes 1340, 1310, 1330, and 1320 are in turn.

Fig. 13C is a schematic diagram of the human-machine conversation system 1300C after the human-machine conversation system 1300A is adjusted according to the weight parameter ordering result 1300B of the conversation node. Like reference numerals in fig. 13C to those in fig. 13A and 13B denote like elements, and are not repeated here.

As shown in fig. 13C, since dialog node 1320 has the lowest weight parameter in weight parameter ranking 1300B, which represents that node 1320 has the lowest contribution to the transformation, dialog node 1320 may be tailored in the adjustment such that interaction page 1320' represented by dialog node 1320 is no longer presented to the user.

Illustratively, because dialog node 1340 has the highest weight parameter in weight parameter ranking 1300B, indicating that its contribution to the conversion node is the highest, interaction portal 1338 pointing to dialog node 1340 may be adjusted such that interaction portal 1338 is at the uppermost position on its corresponding interaction page. Dialogue node 1330C represents a dialogue node that is reset by dialogue node 1330 according to the ranking 1300B of the weighting parameters, where the interaction portal 1338 pointing to the dialogue node 1340 with the highest weighting parameter is located at the top of all interaction portals in the interaction page represented by dialogue node 1330C, so as to ensure that the probability of a user interacting in the interaction page through interaction portal 1338 is the highest.

In summary, by globally evaluating the weight parameter of each dialogue node, the dialogue nodes of the whole man-machine dialogue system can be adjusted and optimized according to the weight parameter, so that the probability that the dialogue nodes with higher weight parameters in the adjusted and optimized man-machine dialogue system are accessed by the user is higher.

Fig. 14 illustrates a flowchart of example operations for obtaining a dialog log (step 210) of a dialog system in the method of fig. 2, in accordance with an embodiment of the present disclosure. As shown in fig. 14, step 210 further includes steps 1410 through 1430.

In step 1410, dialog log data from a human-machine dialog system is collected. For example, the dialog log data collection may be a conventional data collection method in the related art. The dialog log data may include interactive behavior data of the user and data unrelated to the interactions.

In step 1420, the user's interaction with the human-machine dialog system is counted. For example, the user's jump behavior at the current dialog node may be counted. For example, when there is a user jumping to the next session node at the current session node, the data corresponding to the jumping behavior may be counted. Alternatively, the data of the behavior may be counted when the user exits the human-machine interaction system without any action. Alternatively, when a user accesses a target link existing on a dialogue node, the data of the access target link may be counted.

In step 1430, data other than interactive behavior in the dialog log data is deleted. For example, the collected dialog log data may be purged to remove data therein that is not relevant to the interaction behavior.

In summary, according to the method of the embodiment of the disclosure, the conversation log is obtained by screening out irrelevant data in the man-machine conversation data, so as to ensure that the interaction data of each conversation node is considered, and further the man-machine conversation log including accurate global information can be obtained.

Fig. 15 shows a flowchart of the overall operation of a method 1500 of tuning a man-machine conversation system, according to an embodiment of the present disclosure. As shown in fig. 15, the man-machine conversation system adjustment method 1500 includes a log collection step 1510, a log cleansing step 1520, a step 1530 of creating a node state jump map, and a conversation node weight parameter estimation step 1540. Then, a determination is made as to whether the current human-machine dialog system is in an optimal state, via decision 1550. If the dialog node to be adjusted does not exist after the comparison of the weight parameters of the dialog node, which indicates that the current system is in the optimal state, step 1590 is entered. Step 1590 queries other human-machine dialog systems and triggers a new round of log collection process after a period of time. If the dialog node to be adjusted exists after the comparison of the weight parameters of the dialog node, which indicates that the current system is not in the optimal state, the process proceeds to step 1560 of adjusting the dialog node. Illustratively, steps 1510-1560 may be the corresponding steps as described in fig. 2-14 of the present disclosure, which are not repeated herein.

In some exemplary embodiments, step 1580 updates a dialog log of the human-machine dialog system at regular intervals 1570, and repeats steps of the disclosed method based on the updated dialog log.

In summary, the method 1500 for adjusting a man-machine conversation system can automatically perform log collection, conversation node evaluation, and conversation node adjustment and optimization on the current man-machine conversation system after a new man-machine conversation is experienced. Therefore, the adjustment method 1500 can automatically adjust and optimize the man-machine conversation system in an unattended manner at regular intervals, thereby saving the labor cost of operation.

Fig. 16 shows a block diagram of a human-machine dialog system adjustment device in accordance with an embodiment of the disclosure. As shown in fig. 16, the adjustment device 1600 of the human-machine interaction system may include: an acquisition unit 1610 configured to acquire a dialogue log associated with the human-machine dialogue system 1600; a creating unit 1620 configured to create a node state jump graph based on the dialogue log, the node state jump graph defining a jump relationship between a plurality of dialogue nodes and a plurality of dialogue nodes, at least one of the plurality of dialogue nodes representing a respective interaction page of the human-machine dialogue system; a calculating unit 1630 configured to calculate a weight parameter of at least one dialogue node based on the node state jump map; and an adjustment unit 1640 configured to perform a corresponding adjustment operation on the interaction page represented by the at least one dialog node based on the weight parameter of the at least one dialog node.

In some exemplary embodiments, the creation unit 1620 may include: a statistics subunit 1622 configured to count, based on the dialogue log, the number of individual users accessing each interaction page of the human-machine dialogue system 1600 and the operations performed by the individual users for each interaction page; and a generating subunit 1624 configured to generate a plurality of dialog nodes and a skip relation between the plurality of dialog nodes based on the number of independent users accessing each interaction page of the human-machine dialog system 1600 and the operations performed by the independent users for each interaction page, thereby obtaining a node state skip graph.

In other exemplary embodiments, the plurality of dialog nodes may include a transformation node that may represent the behavior of a user accessing a target link on one or more interactive pages of the human-machine dialog system 1600.

In other exemplary embodiments, the computing unit 1630 may include: a first calculating subunit 1632 configured to calculate a skip probability between the plurality of dialogue nodes based on the skip relationship between the plurality of dialogue nodes; a building sub-unit 1634 configured to build a decision process model based on the probability of hops between the plurality of dialog nodes; and a second calculation subunit 1636 configured to calculate, based on the decision process model, a contribution value of each of the at least one dialog node to the conversion node as a weight parameter of each of the at least one dialog node.

In other exemplary embodiments, the adjustment unit 1640 may include: a searching subunit 1642 configured to search for a dialogue node that needs to perform an adjustment operation from the at least one dialogue node based on the weight parameter of the at least one dialogue node; and an adjustment subunit 1644 configured to clip or reset the dialog nodes requiring adjustment operations so that the interaction pages represented by the dialog nodes requiring adjustment operations are removed or rearranged accordingly.

In summary, the apparatus 1600 can evaluate and score each session node in the man-machine session system from the global, and adjust and optimize the man-machine session system by using the weight parameter, so as to ensure that the man-machine session system can be in an optimal state with a higher target conversion condition.

It should be understood that the various units and sub-units of the apparatus 1600 shown in fig. 16 may correspond to the various steps in the method 200 described with reference to fig. 2-14. Thus, the operations, features, and advantages described above with respect to method 200 are equally applicable to apparatus 1600 and the units and sub-units that it includes, and are not repeated for brevity.

Although specific functions are discussed above with reference to specific units, it should be noted that the functions of the various units discussed herein may be divided into multiple units and/or at least some of the functions of the multiple units may be combined into a single unit. The particular unit performing the action discussed herein includes the particular unit itself performing the action, or alternatively the particular unit invoking or otherwise accessing another component or unit performing the action (or performing the action in conjunction with the particular unit). Thus, a particular element performing an action may include the particular element performing the action itself and/or another element performing the action that the particular element invokes or otherwise accesses.

Various techniques may be described herein in the general context of software hardware elements or program modules. The various units, sub-units described above may be implemented in hardware or in hardware in combination with software and/or firmware. For example, the units, sub-units may be implemented as computer program code/instructions configured to be executed in one or more processors and stored in a computer readable storage medium. Alternatively, these modules may be implemented as hardware logic/circuitry. For example, one or more of the units, sub-units, may be implemented together in a system on a chip (SoC). The SoC may include an integrated circuit chip (which includes one or more components of a processor (e.g., a Central Processing Unit (CPU), microcontroller, microprocessor, digital Signal Processor (DSP), etc.), memory, one or more communication interfaces, and/or other circuitry), and may optionally execute received program code and/or include embedded firmware to perform functions.

Examples of such a computer device, non-transitory computer readable storage medium, and computer program product are described below in conjunction with fig. 17. Fig. 17 illustrates a block diagram of an exemplary electronic device that can be used to implement embodiments of the present disclosure.

Referring to fig. 17, a block diagram of an electronic device 1700 that may be a server or client of the present disclosure, which is an example of a hardware device that may be applied to aspects of the present disclosure, will now be described. Electronic devices are intended to represent various forms of digital electronic computer devices, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the disclosure described and/or claimed herein.

As shown in fig. 17, the apparatus 1700 includes a computing unit 1701 that can perform various appropriate actions and processes according to a computer program stored in a Read Only Memory (ROM) 1702 or a computer program loaded from a storage unit 1708 into a Random Access Memory (RAM) 1703. In the RAM 1703, various programs and data required for the operation of the device 1700 may also be stored. The computing unit 1701, the ROM 1702, and the RAM 1703 are connected to each other via a bus 1704. An input/output (I/O) interface 1705 is also connected to the bus 1704.

Various components in device 1700 are connected to I/O interface 1705, including: input unit 1706, output unit 1707, storage unit 1708, and communication unit 1709. Input unit 1706 may be any type of device capable of inputting information to device 1700, input unit 1706 may receive input numeric or character information and generate key signal inputs related to user settings and/or function control of an electronic device, and may include, but is not limited to, a mouse, a keyboard, a touch screen, a trackpad, a trackball, a joystick, a microphone, and/or a remote control. The output unit 1707 may be any type of device capable of presenting information and may include, but is not limited to, a display, speakers, video/audio output terminals, vibrators, and/or printers. The storage unit 1708 may include, but is not limited to, a magnetic disk, an optical disk. The communication unit 1709 allows the device 1700 to exchange information/data with other devices via a computer network, such as the internet, and/or various telecommunications networks, and may include, but is not limited to, modems, network cards, infrared communication devices, wireless communication transceivers and/or chipsets, such as bluetooth (TM) devices, 1302.11 devices, wiFi devices, wiMax devices, cellular communication devices, and/or the like.

The computing unit 1701 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of computing unit 1701 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various computing units running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 1701 performs the various methods and processes described above, such as method 200. For example, in some embodiments, the method 200 may be implemented as a computer software program tangibly embodied on a machine-readable medium, such as the storage unit 1708. In some embodiments, part or all of the computer program may be loaded and/or installed onto device 1700 via ROM 1702 and/or communication unit 1709. One or more of the steps of the method 200 described above may be performed when the computer program is loaded into the RAM 1703 and executed by the computing unit 1701. Alternatively, in other embodiments, the computing unit 1701 may be configured to perform the method 200 by any other suitable means (e.g., by means of firmware).

Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuit systems, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems On Chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.

Program code for carrying out methods of the present disclosure may be written in any combination of one or more programming languages. These program code may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus such that the program code, when executed by the processor or controller, causes the functions/operations specified in the flowchart and/or block diagram to be implemented. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.

In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), 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.

To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and pointing device (e.g., a mouse or trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic input, speech input, or tactile input.

The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), and the internet.

The computer system may include a client and a server. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.

It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps recited in the present disclosure may be performed in parallel, sequentially or in a different order, provided that the desired results of the disclosed aspects are achieved, and are not limited herein.

Although embodiments or examples of the present disclosure have been described with reference to the accompanying drawings, it is to be understood that the foregoing methods, systems, and apparatus are merely exemplary embodiments or examples, and that the scope of the present invention is not limited by these embodiments or examples but only by the claims following the grant and their equivalents. Various elements of the embodiments or examples may be omitted or replaced with equivalent elements thereof. Furthermore, the steps may be performed in a different order than described in the present disclosure. Further, various elements of the embodiments or examples may be combined in various ways. It is important that as technology evolves, many of the elements described herein may be replaced by equivalent elements that appear after the disclosure.

Claims

1. A method for tuning a human-machine dialog system, comprising:

acquiring a dialogue log associated with the man-machine dialogue system;

creating a node state jump graph based on the dialog log, the node state jump graph defining a plurality of dialog nodes and jump relationships between the plurality of dialog nodes, at least one dialog node of the plurality of dialog nodes representing a respective interaction page of the human-machine dialog system;

Calculating weight parameters of the at least one dialogue node based on the node state jump diagram; and

based on the weight parameter of the at least one dialogue node, performing a corresponding adjustment operation on the interaction page represented by the at least one dialogue node,

wherein the plurality of dialog nodes includes a transformation node representing a user's behavior of accessing a target link on one or more interactive pages of the human-machine dialog system,

wherein said calculating weight parameters of said at least one dialogue node based on said node state jump map comprises:

calculating a skip probability among the plurality of dialogue nodes based on the skip relation among the plurality of dialogue nodes;

establishing a decision process model based on the skip probabilities among the plurality of dialogue nodes; and

calculating a contribution value of each of the at least one dialog node to the conversion node as a weight parameter of each of the at least one dialog node based on the decision process model,

wherein said calculating, based on said decision process model, a contribution value of each of said at least one dialog node to said conversion node comprises:

Assigning a prize value to each of the plurality of dialog nodes;

calculating a desire for a benefit function of a dialog node based on the benefit function corresponding to each of a plurality of dialog nodes in the decision process model, the probability of a jump between the dialog node and other dialog nodes, and a reward value for the other dialog nodes; and is also provided with

The expectation of the respective benefit function of the at least one dialog node is taken as the contribution value of the respective at least one dialog node to the conversion node,

wherein, based on the weight parameter of the at least one dialogue node, performing a corresponding adjustment operation on the interaction page represented by the at least one dialogue node includes:

and adjusting the at least one dialogue node so that the interactive pages represented by the at least one dialogue node are presented to the user according to the order of the weight parameters.

2. The method of claim 1, wherein the creating a node state jump map based on the dialog log comprises:

based on the dialogue log, counting the number of independent users accessing each interaction page of the man-machine dialogue system and the operation of the independent users aiming at each interaction page; and

And generating the plurality of dialogue nodes and the jump relation among the plurality of dialogue nodes based on the number of independent users accessing each interaction page of the man-machine dialogue system and the operation of the independent users aiming at each interaction page, so as to obtain the node state jump diagram.

3. The method of claim 1, wherein the establishing a decision process model based on the probability of hops between the plurality of dialog nodes comprises:

obtaining a reward function of each of the plurality of dialogue nodes, wherein the reward function of each of the plurality of dialogue nodes is a desire of a reward value that a decision body can obtain at a next jump time when the decision body is at the dialogue node;

obtaining an attenuation coefficient, wherein the attenuation coefficient acts on the rewarding function to obtain attenuated rewards;

obtaining a benefit function of each of the plurality of dialogue nodes, wherein the benefit function is a summation result of attenuated rewards at all jump moments from the current dialogue node; and is also provided with

And establishing a model for measuring the contribution of the at least one dialogue node to the conversion node as the decision process model according to the reward function, the attenuation coefficient, the benefit function and the jump probability among the dialogue nodes.

4. The method of claim 1, wherein the adjusting the at least one dialog node such that the interactive pages represented by the at least one dialog node are presented to the user in order of the height of the weight parameter comprises:

searching dialogue nodes needing to be subjected to adjustment operation from the at least one dialogue node based on weight parameters of the at least one dialogue node; and

and cutting or resetting the dialogue node needing to be subjected to the adjustment operation so that the interaction page represented by the dialogue node needing to be subjected to the adjustment operation is correspondingly removed or rearranged.

5. The method of claim 4, wherein the searching for the dialog node from the at least one dialog node for which an adjustment operation is required based on the weight parameter of the at least one dialog node comprises:

obtaining a current first sequence of the at least one dialogue node;

sorting the at least one dialogue node according to the magnitude of the weight parameter of the at least one dialogue node;

obtaining a second sequence of the at least one dialogue node after the sequencing;

comparing the first order bit and the second order bit of the at least one dialogue node;

And determining the dialogue node needing to be subjected to adjustment operation based on the comparison result.

6. The method of claim 5, wherein the clipping the dialog node requiring the adjustment operation comprises:

and cutting out the dialogue node with the smallest second order bit in the dialogue nodes needing adjustment operation, so that the interaction page represented by the dialogue node with the smallest second order bit is correspondingly removed from the man-machine dialogue system.

7. The method according to claim 5, wherein the method comprises,

wherein the interaction pages represented by the at least one dialog node each comprise at least one interaction portal, each interaction portal pointing to a next interaction page represented by a respective one of the at least one dialog node,

wherein, the resetting the dialogue node needing to perform the adjustment operation includes:

for each of the session nodes requiring an adjustment operation:

and resetting the at least one interactive portal in the dialogue node according to the ordering of the dialogue node corresponding to the interactive page pointed by the at least one interactive portal in the interactive page represented by the dialogue node, so that the at least one interactive portal is presented according to the ordering result from top to bottom in the interactive page represented by the dialogue node.

8. The method of claim 1, wherein the obtaining a dialog log associated with the human-machine dialog system comprises:

collecting dialogue log data from the man-machine dialogue system;

counting the interaction behavior of the user and the man-machine dialogue system; and

and deleting the data except the interactive behavior in the dialogue log data.

9. The method of any one of claims 1 to 8, further comprising:

updating the dialogue log of the man-machine dialogue system at regular intervals; and is also provided with

The steps of the method are repeated based on the updated dialog log.

10. An adjustment device of a man-machine conversation system, comprising:

an acquisition unit configured to acquire a dialogue log associated with the man-machine dialogue system;

a creating unit configured to create a node state jump graph, based on the dialog logs, the node state jump graph defining a plurality of dialog nodes and jump relationships between the plurality of dialog nodes, at least one dialog node of the plurality of dialog nodes representing a respective interaction page of the human-machine dialog system;

a calculation unit configured to calculate a weight parameter of the at least one dialogue node based on the node state jump map; and

An adjustment unit configured to perform a corresponding adjustment operation on the interaction page represented by the at least one dialog node based on the weight parameter of the at least one dialog node,

assigning a prize value to each of the plurality of dialog nodes;

11. The apparatus of claim 10, wherein the creating unit comprises:

a statistics subunit configured to, based on the dialogue log, count the number of independent users accessing each interaction page of the man-machine dialogue system and operations performed by the independent users for each interaction page; and

and the generation subunit is configured to generate the plurality of dialogue nodes and the jump relation among the plurality of dialogue nodes based on the number of independent users accessing each interaction page of the man-machine dialogue system and the operation of the independent users for each interaction page, so as to obtain the node state jump diagram.

12. The apparatus of claim 10, wherein the adjusting unit comprises:

a searching subunit configured to search for a dialogue node that needs to perform an adjustment operation from the at least one dialogue node based on the weight parameter of the at least one dialogue node; and

and the adjustment subunit is configured to clip or reset the dialogue node needing to be subjected to adjustment operation, so that the interaction page represented by the dialogue node needing to be subjected to adjustment operation is removed or rearranged accordingly.

13. A computer device, comprising:

a memory, a processor and a computer program stored on the memory,

wherein the processor is configured to execute the computer program to implement the steps of the method of any of claims 1-9.

14. A non-transitory computer readable storage medium having stored thereon a computer program, wherein the computer program when executed by a processor implements the steps of the method of any of claims 1-9.