CN116151832A

CN116151832A - Interactive wind control system and method

Info

Publication number: CN116151832A
Application number: CN202310429485.5A
Authority: CN
Inventors: 岳汉
Original assignee: Alipay Hangzhou Information Technology Co Ltd
Current assignee: Alipay Hangzhou Information Technology Co Ltd
Priority date: 2023-04-18
Filing date: 2023-04-18
Publication date: 2023-05-23
Anticipated expiration: 2043-04-18
Also published as: CN116151832B

Abstract

The specification discloses an interactive air control system and method. The system comprises: an operation node deployed on an external network of a designated institution and an execution node deployed on an internal network of the designated institution; wherein: the operation node trains a man-machine interaction model for voice risk prompt based on a risk prompt interaction strategy of a target wind control scene, and submits model data of the trained man-machine interaction model to the execution node according to a cross-network data transmission channel pre-agreed with the appointed mechanism. The execution node deploys the man-machine interaction model in a target cloud application according to the received model data, wherein the target cloud application is called in an internal network by the appointed mechanism, and is used for calling an appointed user outwards, and interacting with the appointed user based on the man-machine interaction model in calling outwards so as to finish voice risk prompt aiming at the target wind control scene.

Description

Interactive wind control system and method

Technical Field

This document belongs to the technical field of wind control, and particularly relates to an interactive wind control system and method.

Background

Traditional wind control mode that relates to trade mainly focuses on identity check and verification, to the operation of paying by oneself because of being influenced by illegal personnel "is difficult to accomplish earlier-stage risk identification, and after the user is influenced by long-time degree of depth" in addition, still probably changes to other channels after the transaction is intercepted by the system and accomplishes the payment, and the wind control degree of difficulty increases by a wide margin. For this reason, interactive air control systems have been created that are "user-centric". The interactive pneumatic control is a safety service for prompting risks to users in proper scenes and at proper occasions, and guides the users to actively identify and treat risks while realizing risk education and counseling. Among them, outbound prompting by a robot "phone wake" is a common implementation of interactive wind control.

Currently, outbound risk cues performed based on robots rely on deep learning human-computer interaction techniques. For institutions requiring deployment, the design conditions of related algorithms are not necessarily met, and due to strict supervision requirements on own user data, the data cannot be shared to cooperate with other institutions. Therefore, a technical scheme capable of cooperatively deploying interactive wind control among institutions on the premise of user data protection is needed currently.

Disclosure of Invention

The embodiment of the specification provides an interactive wind control system and method, which can develop cooperative deployment of interactive wind control service between institutions on the premise of user data protection.

For the above purpose, the embodiments of the present specification are implemented as follows:

in a first aspect, an interactive wind control system is provided, comprising: an operation node deployed on an external network of a designated institution and an execution node deployed on an internal network of the designated institution; wherein:

the operation node trains a man-machine interaction model for voice risk prompt based on a risk prompt interaction strategy of a target wind control scene, and submits model data of the trained man-machine interaction model to the execution node according to a cross-network data transmission channel pre-agreed with the appointed mechanism;

the execution node deploys the man-machine interaction model in a target cloud application according to the received model data, wherein the target cloud application is called in an internal network by the appointed mechanism, and is used for calling an appointed user outwards, and interacting with the appointed user based on the man-machine interaction model in calling outwards so as to finish voice risk prompt aiming at the target wind control scene.

In a second aspect, an interactive wind control system is provided, comprising:

training a man-machine interaction model for voice risk prompt on the basis of a risk prompt interaction strategy of a target wind control scene in an external network of a designated mechanism, and submitting model data of the trained man-machine interaction model to an internal network of the designated mechanism according to a cross-network data transmission channel pre-agreed with the designated mechanism;

and deploying the man-machine interaction model in a target cloud application according to the received model data in an internal network of the appointed mechanism, wherein the target cloud application is called in the internal network by the appointed mechanism, is used for calling an appointed user outwards, and interacts with the appointed user based on the man-machine interaction model in calling conversation outwards so as to finish voice risk prompt aiming at the target wind control scene.

The scheme of the specification provides a cooperation mode aiming at a specified mechanism with interactive wind control requirements but without deep learning algorithm capability. Specifically, a target cloud application for completing interactive wind control is deployed in an internal network of a designated mechanism, and the target cloud application can be called in the internal network by a designated mechanism structure to carry out outbound call on a designated user with possible risk, and interact with the designated user according to a man-machine interaction model in the internal network in outbound call conversation, so that voice risk prompt for a wind control scene is completed. On the basis of the above, an external organization further provided with deep learning algorithm capability deploys operation nodes in the internal network of the designated organization, and deploys execution nodes in the internal network of the designated organization. The operation node fully utilizes the deep learning algorithm capability of an external mechanism, is responsible for training a man-machine interaction model for voice risk prompt according to a risk prompt interaction strategy of a wind control scene, and submits model data of the trained man-machine interaction model to an execution node of an internal network according to a cross-network data transmission channel pre-agreed with a designated mechanism; and mirroring the man-machine interaction model in the internal network of the designated mechanism according to the received model data by the execution node load, thereby ensuring that the man-machine interaction model in the internal network obtains a mature and reliable algorithm. Based on the scheme of the specification, after the man-machine interaction model is deployed in the internal network, the appointed mechanism can call the target cloud application in the internal network to provide voice risk prompt of outbound for the user, the execution node of the external mechanism is completely sealed in the internal network of the appointed mechanism in the whole process, and the data of the appointed mechanism cannot go out of the domain, so that the information of the appointed mechanism can be prevented from being exposed to the external mechanism, and the safety of user data is ensured.

Drawings

The accompanying drawings, which are included to provide a further understanding of the specification, illustrate and explain the exemplary embodiments of the present specification and their description, are not intended to limit the specification unduly. In the drawings:

fig. 1 is a schematic diagram of a first structure of an interactive air control system according to an embodiment of the present disclosure.

Fig. 2 is a schematic diagram of a second structure of the interactive air control system according to the embodiment of the present disclosure.

Fig. 3 is a schematic diagram of an interface related to an outbound platform in an interactive wind control system according to an embodiment of the present disclosure.

Fig. 4 is a schematic flow chart of an interactive air control according to an embodiment of the present disclosure.

Fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present disclosure.

Detailed Description

For the purposes, technical solutions and advantages of this document, the technical solutions of this specification will be clearly and completely described below with reference to specific embodiments of this specification and corresponding drawings. It will be apparent that the embodiments described are only some, but not all, of the embodiments of this document. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are intended to be within the scope of the present disclosure.

As mentioned above, the present robot-based implementation of outbound risk cues is highly dependent on deep learning human-computer interaction techniques. For institutions requiring deployment, the design conditions of related algorithms are not necessarily met, and due to strict supervision requirements on own user data, the data cannot be shared to cooperate with other institutions.

In view of this, this document aims to propose a technical solution suitable for collaborative development of interactive pneumatic control between institutions on the premise of user data protection.

In one aspect, one embodiment of the present description provides an interactive air control system. Wherein, fig. 1 is a schematic structural diagram of an interactive wind control system, including: an operating node 110 deployed at an external network of a designated organization and an executing node 120 deployed at an internal network of the designated organization.

In this embodiment, the designation mechanism refers to an authority responsible for executing risk prompt to the user. The designated institution may be a bank, a security center, an organization, etc., and is not specifically limited herein. In addition, the legal compliance of the appointed mechanism is mastered with the user data, and some information of the user, in particular the contact mode of the user, can be extracted.

As described above, the specification mechanism has no algorithm design condition of man-machine interaction technique although grasping user data, and therefore, it is necessary to cooperate with an external mechanism to provide algorithm support by the external mechanism.

For the algorithm mechanism, the present embodiment splits the algorithm mechanism into the operation node 110 and the execution node 120 according to the operation logic. Wherein:

the operation node of the external network of the designated organization is responsible for training a man-machine interaction model for voice risk prompt based on a risk prompt interaction strategy of a target wind control scene, and submits model data (i.e., algorithm data of the model) of the trained man-machine interaction model to the execution node 120 according to a cross-network data transmission channel pre-agreed with the designated organization.

As an exemplary introduction, the target wind-controlled scenario may be a wind-controlled scenario for a telecommunication illegal activity, a wind-controlled scenario for a financial illegal activity, etc., which is not specifically limited herein. In addition, the cross-network data transmission channel is a channel in which the operation node 110 transmits data unidirectionally to the execution node 120. That is, the user data of the authority cannot be transmitted to the external network without leaving the domain in the internal network.

And the execution node of the network in the designated mechanism is responsible for deploying the trained man-machine interaction model in the target cloud application according to the received model data.

The target cloud application is called in the internal network by a designation mechanism, is used for calling out a designated user with potential risk, and interacts with the designated user based on a deployed man-machine interaction model in the calling-out call, so that voice risk prompt aiming at a target wind control scene is completed.

As an exemplary introduction, the target cloud application of the present embodiment may invoke an outbound platform of a telecommunications carrier according to a session initiation protocol (Session Initiation Protocol, SIP) to outbound a specified user for a specified organization-specific telecommunications number using the outbound platform.

In practical application, the target cloud application can also make outbound availability judgment before outbound the designated user. For example, whether the electrical signal code of the appointed user is available or not, whether the outbound platform is normal or not, whether the preset outbound time period is out of date or not is judged, and the like. And if the availability judgment is passed, the target cloud application recalls the outbound platform to outbound the appointed user. For the wind control scene with larger user scale, the availability judgment can reduce the invalid resource occupation (such as the occupation of an outbound interface) caused by the outbound platform to a certain extent, so that the outbound platform can better serve the interactive wind control.

Further, if the specified user selects To answer the outbound call, the target cloud application may obtain the input content of the specified user in the outbound call through an automatic Speech recognition technology (Automatic Speech Recognition, ASR) in the outbound call, determine the response content corresponding To the input content based on the deployed man-machine interaction model, and then convert the response content into audio data according To a Text To Speech (TTS) technology To play the specified user.

Specifically, the human-computer interaction model of the present embodiment may include an intention recognition model for recognizing a conversation intention (also referred to as speaking) corresponding to the specified user input content; correspondingly, after the target cloud application determines the call intention, the call intention related call content can be matched from a call library preset for the target wind control scene, and the matched call content is used as response content to be played through a TTS technology.

In addition, the target cloud application of the embodiment may also generate a question content based on user information of the specified user and a preset question template locally recorded by the specified mechanism in the outbound call, and play the question content to the specified user through a TTS technology.

In the practical application, the target cloud application can actively guide the appointed user to interact by playing the questioning content through voice. For example, in outbound passes, the target cloud application initiates interactions of the first sentence in a questioning manner in the first round. In addition, the target cloud application can also verify the identity and operation of the specified user in a questioning manner. Taking identity verification as an example, the question template for identity verification may be "you are +user name +my person", and correspondingly, the target cloud application may extract name information of the specified user from user data of the specified institution, and import the name information into the question template of "you are +user name +my person" to generate question content for identity verification, for example, "you are Zhang Sanmy person".

In addition, the target cloud application may record outbound result data (such as whether the outbound is on or not, whether the designated user accepts the prompt, etc.) of the outbound call, so that the designated mechanism performs subsequent processing, which is not described herein again by way of example.

Based on the above, it can be seen that: the system of the embodiment of the specification deploys a target cloud application for completing interactive wind control in an internal network of a designated organization, wherein the target cloud application can be called in the internal network by the structure of the designated organization to make an outbound call to a designated user with possible risk, and interacts with the designated user according to a man-machine interaction model in the internal network in an outbound call conversation, so that voice risk prompt for a wind control scene is completed. On the basis of the above, an external organization further provided with deep learning algorithm capability deploys operation nodes in the internal network of the designated organization, and deploys execution nodes in the internal network of the designated organization. The operation node fully utilizes the deep learning algorithm capability of an external mechanism, is responsible for training a man-machine interaction model for voice risk prompt according to a risk prompt interaction strategy of a wind control scene, and submits model data of the trained man-machine interaction model to an execution node of an internal network according to a cross-network data transmission channel pre-agreed with a designated mechanism; and mirroring the man-machine interaction model in the internal network of the designated mechanism according to the received model data by the execution node load, thereby ensuring that the man-machine interaction model in the internal network obtains a mature and reliable algorithm. Based on the scheme of the specification, after the man-machine interaction model is deployed in the internal network, the appointed mechanism can call the target cloud application in the internal network to provide voice risk prompt of outbound for the user, the execution node of the external mechanism is completely sealed in the internal network of the appointed mechanism in the whole process, and the data of the appointed mechanism cannot go out of the domain, so that the information of the appointed mechanism can be prevented from being exposed to the external mechanism, and the safety of user data is ensured.

The following describes the interactive wind control system of the present embodiment in detail in connection with an actual application scenario.

The application scene aims at an electronic payment service to construct an interactive wind control system for prompting the payment risk of illegal telecommunication activities by voice.

Referring to fig. 2, the interactive wind control system mainly comprises an operation node and an execution node at the side of an external organization. The operation node is further subdivided into a strategy operation center, a model training center and a local site management center; the executing nodes are subdivided with software as a service platform (Software as a Service, SAAS) and model engines that provide interactive wind control.

In the application scene, the execution node of the external mechanism is locally deployed in the internal network of the appointed mechanism and is in an intranet closed state. Meanwhile, the operation nodes of the external institutions are managed by the external institutions, and can be directly deployed on the external network of the appointed institutions.

Specifically, in the internal network of the designated organization, the SAAS is deployed with an interactive wind-controlled cloud application cluster, and entity nodes in each region of the designated organization can call cloud applications in the internal network through service interfaces of the SAAS. It should be appreciated that the process of invoking the cloud application is done in the internal network, which can ensure that its own data does not go out of domain. Meanwhile, an operation node in an external network is mainly responsible for centrally managing risk prompt interaction strategies, model algorithms and the like, and a local site management center utilizes a cross-network data transmission channel to carry out unidirectional data transmission, dynamically updates strategies for SAAS and model data for a model engine, so that the wind control capability of the system is ensured not to be attenuated.

The reason for this division is as follows:

1) The software of the cloud application and the model engine belongs to a relatively stable part, and the software cannot be easily changed unless software function upgrading or bug repairing is carried out, and the part is deployed in an internal network of a designated mechanism, so that the maintenance cost is relatively controllable; in addition, in the running process of the software, user data of a designated mechanism is needed to be used, and the internal network deployed in the designated mechanism can ensure that the user data does not leave the domain.

2) The strategy and model algorithm need high-frequency adjustment in the evolution process of risk attack and defense, the part of the strategy and model algorithm is configured in an external network of a designated organization, and centralized management is carried out by the external organization, so that the algorithm capability of the external organization can be fully utilized to improve the risk prevention and control level and improve the operation efficiency. In addition, the strategy, model data and the like produced after adjustment can be transmitted to the cloud application cluster of the network in the appointed mechanism in a mode of issuing the configuration file, and the problem that the appointed mechanism leaks user data does not exist because the configuration file is issued in one-way transmission.

The following describes the nodes and functions involved in an interactive air control system:

1. cross-network data transmission channel

In the application scene, the internal network of the appointed mechanism is taken as a whole local area network, and the entity nodes of each area of the appointed mechanism are connected in series and isolated from a public network. For the purpose of issuing strategies, model data and the like, a network communication special line, namely a cross-network data transmission channel, is reserved between the internal networks of the external institutions and the designated institutions.

The operation node transmits local strategies, model data and the like to the execution node of the internal network of the appointed mechanism through the cross-network data transmission channel, so that the daily operation efficiency is improved. Meanwhile, the cross-network data transmission channel can ensure the safety of the data transmission through data encryption and authority control so as to meet the supervision regulations of the specified institutions for data transmission.

Specifically, the cross-network data transmission channel can be used as a transfer platform for data transmission, is configured with a dedicated storage space, and is responsible for operation and maintenance by a designated mechanism. In the data issuing process, the operation node can submit the strategy, the model data and the corresponding deployment execution parameters (optional) to a cross-network data transmission channel; then, accessing the storage space into an external network through a cross-network data transmission channel, and writing the received strategy, model data and deployment execution parameters into the storage space; after the writing is completed, the cross-network data transmission channel further switches the storage space to be connected with the internal network. And then, the execution node reads the strategy, the model data and the deployment execution parameters from the storage space in the internal network, updates the read strategy into the SASS according to the indication of the deployment execution parameters, and updates the read model data into the model engine according to the indication of the deployment execution parameters, so that the corresponding human-computer interaction model is mirrored. It should be noted that, in practical application, the man-machine interaction model is not unique, and each man-machine interaction model in the model engine can be invoked by cloud application of the SASS.

It should be understood that, based on the design of the cross-network data transmission channel, the external mechanism and the designated mechanism can be isolated from each other through the storage space, so as to ensure the security of the data issuing process.

2. Model policy operation

In the application scene, the operation of the strategy and the model adopts a centralized management mode, and the unified operation is carried out by an external organization. The technical staff responsible for the strategy and the model algorithm carries out strategy adjustment and corresponding model training by combining the current risk situation in a strategy operation center and a model training center of an external mechanism, and issues model data obtained after strategy and training through a cross-network data transmission channel.

Referring to fig. 2, the issued data flow is mainly divided into the following steps:

1) After model training is completed, configuration issuing operation is carried out through the local site management center, so that configuration information is packaged into a configuration file, and the configuration file is encrypted and then sent to a cross-network data transmission channel for transfer. To this end, the configuration information has been submitted by the external institution to the internal network of the designated institution.

2) And carrying out data decryption on the cross-network data transmission channel, submitting the configuration file to a cloud application cluster of the SAAS in a route distribution mode, loading the configuration file by the cloud application cluster, and finally finishing upgrading of cluster policy configuration.

It should be understood that the configuration file belongs to unidirectional transmission in the above process, so that the operation efficiency of the policy model is ensured, and sensitive information of a user is not involved.

3. The appointed mechanism executes interactive wind control

In this application scenario, the SASS can provide services to entity nodes deployed by a designated organization in various regions. The appointed mechanism relates to the data flow direction and is divided into the following steps:

1) When the appointed mechanism receives the early warning task and needs to carry out outbound warning on part of appointed users suspected to have risks, the outbound task is uploaded to the system background through a local operating system.

2) And the system background responds to the outbound task, transmits outbound information to the interactive wind-controlled operation application cluster, and executes strategy risk analysis and model recommendation.

3) After the recommendation of the interactive wind control operation system is finished, the local telephone traffic system of the appointed mechanism is called back through a special line, the interactive wind control operation system interacts with a local telephone operator through an SIP protocol, finally, the special telephone number of the appointed mechanism is used for calling out, a user is touched, and finally, tasks such as risk prompt, education and the like are completed through a recommended man-machine interaction model.

It should be understood that, in the above process, the dedicated telecommunication number of the designated institution is used to make the outbound, so that the authority of the designated institution can be utilized to improve the success rate of outbound, and achieve a better risk prompting effect.

4. Outbound system docking

In the application scene, the system can be in butt joint with the outgoing private line of each entity node of the appointed mechanism through the running engine, so that outgoing is carried out through the local exclusive telecommunication number.

Specifically, the interactive interface related to the outbound platform of the runtime engine is divided into three stages of before, during and after calling according to the illustration of fig. 3. Wherein:

1) Before calling, the interactive wind control system receives the outbound command, and after finishing the internal flow, firstly calls an outbound platform availability interface to judge whether the outbound platform can execute outbound tasks at the moment, if the availability is passed, the outbound task is initialized, and meanwhile, the text of the first round of speech operation is transmitted to the outbound platform. Next, the outbound platform invokes tts algorithm services, converts text to audio, and initiates first round interactions by the operator touching the user.

2) In the call, after the user finishes answering, the answer content audio is transmitted to an outbound platform, at the moment, the outbound platform calls asr algorithm service, converts the audio stream into text, then calls interactive wind-controlled cloud application, acquires the next round of phone operation, and touches the user again through an operator. And (5) repeating the steps until the cloud application of the interactive wind control notifies the outbound platform to finish.

3) After calling, the outbound platform receives the user hang-up message and transmits the hang-up message to the cloud application of the interactive wind control in a callback mode so as to execute the post logic of the platform. Of course, in practical applications, the user may also reject the outbound call directly, in which case the outbound call platform will call the reject message back to the cloud application of the interactive air control.

The description of the interfaces and functions involved in the outbound procedure described above is shown in the following table, as an exemplary introduction:

corresponding to the system shown in fig. 2, the present disclosure also provides an interactive wind control method. Wherein, fig. 4 is a flowchart of an interactive wind control method according to an embodiment of the present disclosure, including:

s402, training a man-machine interaction model for voice risk prompt on the basis of a risk prompt interaction strategy of a target wind control scene in an external network of a designated institution, and submitting model data of the trained man-machine interaction model to an internal network of the designated institution according to a cross-network data transmission channel pre-agreed with the designated institution.

S404, a man-machine interaction model is deployed in a target cloud application according to the received model data in an internal network of a designated mechanism, wherein the target cloud application is called in the internal network by the designated mechanism, is used for calling a designated user outwards, and interacts with the designated user based on the man-machine interaction model in calling outwards to finish voice risk prompt aiming at a target wind control scene.

The method of the embodiment of the specification provides a cooperation mode aiming at a specified mechanism with interactive wind control requirements but without deep learning algorithm capability. Specifically, a target cloud application for completing interactive wind control is deployed in an internal network of a designated mechanism, and the target cloud application can be called in the internal network by a designated mechanism structure to carry out outbound call on a designated user with possible risk, and interact with the designated user according to a man-machine interaction model in the internal network in outbound call conversation, so that voice risk prompt for a wind control scene is completed. On the basis of the above, an external organization further provided with deep learning algorithm capability deploys operation nodes in the internal network of the designated organization, and deploys execution nodes in the internal network of the designated organization. The operation node fully utilizes the deep learning algorithm capability of an external mechanism, is responsible for training a man-machine interaction model for voice risk prompt according to a risk prompt interaction strategy of a wind control scene, and submits model data of the trained man-machine interaction model to an execution node of an internal network according to a cross-network data transmission channel pre-agreed with a designated mechanism; and mirroring the man-machine interaction model in the internal network of the designated mechanism according to the received model data by the execution node load, thereby ensuring that the man-machine interaction model in the internal network obtains a mature and reliable algorithm. Based on the scheme of the specification, after the man-machine interaction model is deployed in the internal network, the appointed mechanism can call the target cloud application in the internal network to provide voice risk prompt of outbound for the user, the execution node of the external mechanism is completely sealed in the internal network of the appointed mechanism in the whole process, and the data of the appointed mechanism cannot go out of the domain, so that the information of the appointed mechanism can be prevented from being exposed to the external mechanism, and the safety of user data is ensured.

Optionally, the target cloud is applied in the outbound call, obtains input content of the specified user in the outbound call based on an automatic voice recognition technology, determines response content corresponding to the input content based on the deployed man-machine interaction model, and plays the response content to the specified user in a voice mode based on a text-to-voice technology.

Optionally, the target cloud is applied in the outbound call, and further generates a question content based on the user information of the specified user and a preset question template recorded locally by the specified mechanism, and plays the question content to the specified user through voice based on the text-to-voice technology.

Optionally, the target cloud application invokes an outbound platform of a telecommunications carrier based on a session initiation protocol to outbound the specified user using a telecommunications code specific to the specified organization.

Optionally, the target cloud application makes an outbound availability determination before outbound the specified user, wherein the availability determination includes at least one of:

judging whether the electric signal code of the appointed user is available or not;

judging whether the outbound platform is normal or not;

judging whether the preset outbound time period is expired.

Optionally, the target cloud application further sends outbound result data of the outbound call to the designated mechanism, and the designated mechanism records the received outbound result data.

Optionally, the operation node submits deployment execution parameters corresponding to the model data to the execution node based on the cross-network data transmission channel;

the cross-network data transmission channel is configured with a designated storage space, and after receiving the model data and the corresponding deployment execution parameters sent by the operation node, the cross-network data transmission channel accesses the storage space to the external network, writes the model data and the deployment execution parameters into the storage space, and switches the storage space to be connected with the internal network after the writing is completed; and after the storage space is connected with the internal network, the execution node reads the model data and the deployment execution parameters from the storage space in the internal network, and deploys the model data into the target cloud application according to the deployment execution parameters.

Optionally, the designated mechanism is deployed with a software-as-a-service platform in the internal network, and the target cloud application is provided to the designated mechanism for calling by the software-as-a-service platform in a subscription mode.

Optionally, the target wind control scenario belongs to an electronic payment service, and the target cloud application is used for prompting payment risk of illegal telecommunication activities through voice.

Obviously, the interactive wind control method in the embodiment of the present disclosure takes the interactive wind control system shown in fig. 1 as an execution body, so that the corresponding function of the interactive wind control system can be realized, and detailed descriptions thereof are omitted.

Fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present disclosure. Referring to fig. 5, at the hardware level, the electronic device includes a processor, and optionally an internal bus, a network interface, and a memory. The Memory may include a Memory, such as a Random-Access Memory (RAM), and may further include a non-volatile Memory (non-volatile Memory), such as at least 1 disk Memory. Of course, the electronic device may also include hardware required for other services.

The processor, network interface, and memory may be interconnected by an internal bus, which may be an ISA (Industry Standard Architecture ) bus, a PCI (Peripheral Component Interconnect, peripheral component interconnect standard) bus, or EISA (Extended Industry Standard Architecture ) bus, among others. The buses may be classified as address buses, data buses, control buses, etc. For ease of illustration, only one bi-directional arrow is shown in FIG. 5, but not only one bus or type of bus.

And the memory is used for storing programs. In particular, the program may include program code including computer-operating instructions. The memory may include memory and non-volatile storage and provide instructions and data to the processor.

The processor reads the corresponding computer program from the nonvolatile memory to the memory and then runs the computer program to form the interactive wind control system on a logic level. The processor is used for executing the programs stored in the memory and is specifically used for executing the following operations:

training a man-machine interaction model for voice risk prompt on the basis of a risk prompt interaction strategy of a target wind control scene in an external network of a designated mechanism, and submitting model data of the man-machine interaction model after training to an internal network of the designated mechanism according to a cross-network data transmission channel pre-agreed with the designated mechanism.

The method suggested by the embodiment shown in fig. 5 of the present specification can be applied to a processor or implemented by the processor. The processor may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the above method may be performed by integrated logic circuits of hardware in a processor or by instructions in the form of software. The processor may be a general-purpose processor, including a central processing unit (Central Processing Unit, CPU), a network processor (Network Processor, NP), etc.; but also digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), field programmable gate arrays (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. The various methods, steps, and logic blocks disclosed in one or more embodiments of the present description may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of a method disclosed in connection with one or more embodiments of the present disclosure may be embodied directly in a hardware decoding processor or in a combination of hardware and software modules in a decoding processor. The software modules may be located in a random access memory, flash memory, read only memory, programmable read only memory, or electrically erasable programmable memory, registers, etc. as well known in the art. The storage medium is located in a memory, and the processor reads the information in the memory and, in combination with its hardware, performs the steps of the above method.

The electronic device may also perform the method shown in fig. 4, which is not described in detail herein.

Of course, in addition to the software implementation, the electronic device in this specification does not exclude other implementations, such as a logic device or a combination of software and hardware, that is, the execution subject of the following process is not limited to each logic unit, but may also be hardware or a logic device.

The present specification embodiment also proposes a computer-readable storage medium storing one or more programs.

Wherein the one or more programs include instructions, which when executed by a portable electronic device comprising a plurality of application programs, enable the portable electronic device to perform the method of the embodiment of fig. 4, and in particular to:

The electronic device in the embodiment of the specification provides a cooperation mode aiming at a specified mechanism with interactive wind control requirements but without deep learning algorithm capability. Specifically, a target cloud application for completing interactive wind control is deployed in an internal network of a designated mechanism, and the target cloud application can be called in the internal network by a designated mechanism structure to carry out outbound call on a designated user with possible risk, and interact with the designated user according to a man-machine interaction model in the internal network in outbound call conversation, so that voice risk prompt for a wind control scene is completed. On the basis of the above, an external organization further provided with deep learning algorithm capability deploys operation nodes in the internal network of the designated organization, and deploys execution nodes in the internal network of the designated organization. The operation node fully utilizes the deep learning algorithm capability of an external mechanism, is responsible for training a man-machine interaction model for voice risk prompt according to a risk prompt interaction strategy of a wind control scene, and submits model data of the trained man-machine interaction model to an execution node of an internal network according to a cross-network data transmission channel pre-agreed with a designated mechanism; and mirroring the man-machine interaction model in the internal network of the designated mechanism according to the received model data by the execution node load, thereby ensuring that the man-machine interaction model in the internal network obtains a mature and reliable algorithm. Based on the scheme of the specification, after the man-machine interaction model is deployed in the internal network, the appointed mechanism can call the target cloud application in the internal network to provide voice risk prompt of outbound for the user, the execution node of the external mechanism is completely sealed in the internal network of the appointed mechanism in the whole process, and the data of the appointed mechanism cannot go out of the domain, so that the information of the appointed mechanism can be prevented from being exposed to the external mechanism, and the safety of user data is ensured.

In summary, the foregoing description is only a preferred embodiment of the present invention and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, improvement, or the like, which is within the spirit and principles of one or more embodiments of the present disclosure, is intended to be included within the scope of one or more embodiments of the present disclosure.

The system, apparatus, module or unit set forth in the above embodiments may be implemented in particular by a computer chip or entity, or by a product having a certain function. One typical implementation is a computer. In particular, the computer may be, for example, a personal computer, a laptop computer, a cellular telephone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

Computer readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. Computer-readable media, as defined herein, does not include transitory computer-readable media (transmission media), such as modulated data signals and carrier waves.

It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or apparatus that comprises the element.

In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments. In particular, for system embodiments, since they are substantially similar to method embodiments, the description is relatively simple, as relevant to see a section of the description of method embodiments.

Claims

1. An interactive air control system, comprising: an operation node deployed on an external network of a designated institution and an execution node deployed on an internal network of the designated institution; wherein:

2. The system of claim 1, wherein the system further comprises a controller configured to control the controller,

the target cloud is applied to the outbound call, acquires input content of the appointed user in the outbound call based on an automatic voice recognition technology, determines response content corresponding to the input content based on the deployed man-machine interaction model, and plays the response content to the appointed user in a voice mode based on a text-to-voice technology.

3. The system of claim 2, wherein the system further comprises a controller configured to control the controller,

and the target cloud is applied to the outbound call, generates questioning contents based on the user information of the appointed user locally recorded by the appointed mechanism and a preset questioning template, and plays the questioning contents to the appointed user through voice based on the text-to-voice technology.

4. The system of claim 1, wherein the system further comprises a controller configured to control the controller,

the target cloud application calls an outbound platform of a telecom operator based on a session initiation protocol to outbound the specified user by using a telecommunication number specific to the specified organization.

5. The system of claim 4, wherein the system further comprises a controller configured to control the controller,

the target cloud application performs outbound availability judgment before outbound the specified user, wherein the availability judgment comprises at least one of the following:

judging whether the outbound platform is normal or not;

judging whether the preset outbound time period is expired.

6. The system of any one of claims 1-5, wherein,

and the target cloud application also records outbound result data of the outbound call.

7. The system of any one of claims 1-5, wherein,

the operation node submits deployment execution parameters corresponding to the model data to the execution node based on the cross-network data transmission channel;

8. The system of any one of claims 1-5, wherein,

the specified mechanism is provided with a software-as-a-service platform in the internal network, and the target cloud application is provided for the specified mechanism to be called in a subscription mode by the software-as-a-service platform.

9. The system of any one of claims 1-5, wherein,

the target wind control scene belongs to an electronic payment service, and the target cloud application is used for prompting payment risks of illegal telecommunication activities through voice.

10. An interactive wind control method is characterized in that,