Intelligent outbound method and device based on AI algorithm

Abstract

The invention provides an intelligent outbound method and device based on an AI algorithm. The method comprises the following steps: firstly, on the basis of a machine learning model, carrying out multi-dimensional portrait construction on collected past behavior data of a user, and generating a figure portrait; and then, based on the dimension of the figure portrait, performing multi-dimensional analysis on the outbound content to generate an outbound content portrait. Then, on the basis of a preset matching degree threshold value, the dimensions of the figure portrait and the outbound content portrait are matched, and the outbound matching degree is generated; and finally, based on the outbound matching degree, synthesizing the outbound content into an echo for voice feedback, and completing intelligent calling of the to-be-outbound person. According to the embodiment of the invention, the method achieves the targeted call-out of the to-be-called person, can guarantee higher recognition precision, improves the recognition accuracy and communication efficiency through the semantic segmentation technology and the voiceprint recognition noise shielding technology, and finally improves the order forming rate.

Description

Technical Field

[0001] The present disclosure relates to the field of network marketing technology, and in particular to an intelligent outbound calling method and device based on an AI algorithm. Background Art

[0002] With the rapid development of artificial intelligence technology, intelligent outbound call systems have gradually become an important tool for improving customer interaction efficiency. Traditional outbound call systems usually rely on preset scripts and processes when processing large-scale customer data and making automated calls, lacking flexibility and personalized customer responses. This approach not only leads to poor user experience, but also has problems of low efficiency, especially when facing diverse customer needs, it is difficult to provide accurate services. By combining artificial intelligence algorithms, especially natural language processing, speech recognition, personalized recommendations and other technologies, it is not only possible to achieve accurate matching of customer needs, but also improve the effect of customer interaction and overall operational efficiency.

[0003] In the existing technical solution, outbound call tasks are managed by collaborating with the middleware server cluster and the outbound call server. Although this solution realizes the automated management of outbound call tasks by artificial intelligence, its main problem is that it cannot achieve accurate matching of outbound call content and outbound call objects. Specifically, outbound call content is often fixed, lacking targeted personalized recommendations or intelligent push based on user historical behavior. In addition, the efficiency of the system is also constrained by the speed of data processing and task scheduling. It cannot respond to changes in the needs of different customers in real time and still relies on the traditional sequential calling mechanism. This makes it difficult for this technical solution to meet the needs of the intelligent outbound call platform for efficient, accurate and personalized services.

[0004] Therefore, one or more methods are needed to solve the above problems.

[0005] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention

[0006] The present disclosure shows an intelligent outbound calling method and device based on AI algorithm.

[0007] In a first aspect, the present disclosure provides an intelligent outbound calling method based on an AI algorithm, the method comprising:

[0008] Step 1: Use machine learning models to construct multi-dimensional portraits of collected user past behavior data to build character portraits;

[0009] Step 2: Based on the dimensions of the character portrait, a multi-dimensional analysis is performed on the outbound call content to generate an outbound call content portrait;

[0010] Step 3, setting a preset matching degree threshold, and matching the dimensions of the character portrait and the outbound call content portrait by the preset matching degree threshold to generate an outbound call matching degree;

[0011] Step 4, according to the outbound call matching degree, the outbound call content is synthesized into a reply for voice feedback, thereby completing the intelligent call for the outbound caller.

[0012] In an exemplary embodiment of the present disclosure, a multi-dimensional portrait is constructed by using a machine learning model to collect the user's past behavior data, including:

[0013] Based on the user's unique identifier, text data is generated by preprocessing the user's previous behavior data;

[0014] Based on the machine learning model, feature extraction is performed on the text data to generate feature words, and high-dimensional word vectors are generated by performing high-dimensional transformation on the feature words;

[0015] Based on the time when the behavior occurred, the weight coefficient is generated by weighting the user's activity in the behavior;

[0016] Based on the weight coefficient, weighting the high-dimensional word vector to generate a weighted word vector;

[0017] Based on the weighted word vector, the dimensional preference scores of the character portraits are calculated by using a cosine similarity algorithm to generate preference scores;

[0018] Based on the preference scores, generating character portrait dimensions by clustering and grading the user's preferences on the dimensions;

[0019] Based on the character portrait dimensions, a character portrait is generated by extracting and analyzing the user's features.

[0020] In an exemplary embodiment of the present disclosure, a multi-dimensional analysis is performed on the outbound call content, including:

[0021] Based on the match between the outbound call content and the user, the outbound call content dimension is generated by calculating the intersection of the outbound call content and the character portrait dimension;

[0022] Based on the outbound call content dimension, an outbound call content portrait is generated by extracting and analyzing the features of the outbound call information.

[0023] In an exemplary embodiment of the present disclosure, matching the dimensions of the character portrait and the outbound call content portrait includes:

[0024] Based on a preset matching degree threshold, a key score of the character is generated by calculating the matching degree of the key dimension in the character portrait;

[0025] Based on a preset matching degree threshold, a content key score is generated by calculating the matching degree of the key dimension in the outbound call content portrait;

[0026] A dimension score is generated by calculating the product of the character key score and the content key score, and an outbound call matching degree is generated by summarizing the dimension scores.

[0027] In an exemplary embodiment of the present disclosure, voice feedback is performed by synthesizing the outbound call content into a reply, including:

[0028] When a communication connection is established, an input voice is generated by picking up a voice signal in the communication connection;

[0029] Based on the long speech segmentation technology, the input speech is segmented to generate an input short speech set, and the short speech in the input short speech set is recognized to generate the intention of the intended person;

[0030] Based on the intention of the intended person, the outbound call content is matched by the outbound call matching degree to generate reply content;

[0031] Based on speech synthesis technology, the reply content is speech synthesized to generate speech feedback, thereby completing the intelligent call for the outbound caller.

[0032] In an exemplary embodiment of the present disclosure, picking up a voice signal in a communication connection includes:

[0033] Based on the preset background channel model, the voiceprint identification vector is generated by training the voice of the outbound caller;

[0034] When a communication connection is established, a voiceprint feature vector is generated by extracting the voiceprint feature of the speech signal in the communication connection;

[0035] Based on a preset distance threshold, when the spatial distance between the voiceprint identification vector and the voiceprint feature vector is not less than the preset distance threshold, determining that the voice signal is noise;

[0036] Based on a preset distance threshold, when the spatial distance between the voiceprint identification vector and the voiceprint feature vector is less than the preset distance threshold, the voice signal is determined to be the voice of the caller, and the input voice is generated.

[0037] In an exemplary embodiment of the present disclosure, segmenting the input speech includes:

[0038] Based on a preset silence model, a continuous voice signal in the input voice is recognized by a deep learning algorithm to generate a silent voice;

[0039] Taking the silent speech as a segmentation point, the input speech is segmented to generate an input short speech set.

[0040] In a second aspect, the present disclosure shows an intelligent outbound calling device based on an AI algorithm, the device comprising:

[0041] The character portrait module is used to construct a character portrait by using a machine learning model to construct a multi-dimensional portrait of the collected user's past behavior data;

[0042] The outbound call content portrait module is used to generate an outbound call content portrait by performing multi-dimensional analysis on the outbound call content;

[0043] A matching module, used to set a preset matching degree threshold, and match the dimensions of the character portrait and the outbound call content portrait by the preset matching degree threshold to generate an outbound call matching degree;

[0044] The outbound call module is used to complete intelligent calls to outbound callers by synthesizing the outbound call content into a reply for voice feedback.

[0045] In a third aspect, the present disclosure shows an electronic device, comprising: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to execute the method described in any of the above aspects.

[0046] In a fourth aspect, the present disclosure shows a non-temporary computer-readable storage medium, which, when instructions in the storage medium are executed by a processor of an electronic device, enables the electronic device to execute the method described in any of the above aspects.

[0047] In a fifth aspect, the present disclosure shows a computer program product. When instructions in the computer program product are executed by a processor of an electronic device, the electronic device is enabled to execute the method described in any of the above aspects.

[0048] The technical solution provided by the present disclosure may have the following beneficial effects:

[0049] Based on the embodiment of the present disclosure, first, according to the machine learning model, a character portrait is generated by constructing a multi-dimensional portrait of the collected user's previous behavior data. Afterwards, based on the dimensions of the character portrait, a multi-dimensional analysis of the outbound call content is performed to generate an outbound call content portrait. Then, based on the preset matching threshold, the outbound call matching degree is generated by matching the dimensions of the character portrait and the outbound call content portrait. Finally, based on the outbound call matching degree, the outbound call content is synthesized into a reply for voice feedback to complete the intelligent call for the outbound caller. Therefore, the embodiment of the present disclosure provides an intelligent outbound call system, which accurately matches the character portrait with the outbound call content portrait. Targeted outbound calls to outbound callers are achieved, which can ensure higher recognition accuracy, and the recognition accuracy and communication efficiency are improved through semantic segmentation technology and voiceprint recognition to shield noise technology, thereby ultimately improving the order conversion rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a flowchart of the steps of an intelligent outbound calling method based on an AI algorithm disclosed in the present invention;

[0051] Figure 2 It is a structural block diagram of an intelligent outbound calling device based on AI algorithm disclosed in the present invention;

[0052] Figure 3 is a block diagram of an electronic device disclosed herein;

[0053] Figure 4 is a block diagram of a computer-readable storage medium of the present disclosure. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present disclosure to clearly and completely describe the technical solutions in the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, rather than all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0055] Reference Figure 1 , shows a flowchart of a step of an intelligent outbound calling method based on an AI algorithm disclosed in the present invention, the method can be applied to an electronic device, wherein the method can specifically include the following steps:

[0056] Step S101, construct a character portrait by using a machine learning model to construct a multi-dimensional portrait of the collected user's past behavior data.

[0057] Step S102, based on the dimensions of the character portrait, generate an outbound call content portrait by performing a multi-dimensional analysis on the outbound call content.

[0058] Step S103, setting a preset matching degree threshold, and matching the dimensions of the character portrait and the outbound call content portrait by using the preset matching degree threshold to generate an outbound call matching degree.

[0059] Step S104, according to the outbound call matching degree, the outbound call content is synthesized into a reply for voice feedback, thereby completing the intelligent call for the outbound caller.

[0060] This intelligent outbound calling method based on AI algorithm achieves accurate matching by making multi-dimensional portraits of people and outbound calling content, enhancing the pertinence of outbound calling, and improving communication efficiency and order conversion rate. At the same time, it improves the accuracy of speech recognition through speech segmentation technology, and combines the voiceprint recognition method to identify the speaker, which can effectively block noise and further improve recognition accuracy. This solution not only optimizes the execution effect of outbound calling tasks, but also improves the overall performance of the system in terms of speech processing, ensuring the accuracy and efficiency of information during outbound calling.

[0061] Next, an intelligent outbound calling method based on an AI algorithm in an embodiment of the present disclosure will be further described.

[0062] In step S101, a multi-dimensional portrait may be constructed based on the collected past behavior data of the user using a machine learning model to construct a character portrait.

[0063] In some optional embodiments of this example, constructing a multi-dimensional portrait of the user's past behavior data includes the following steps:

[0064] The first step is data collection and processing. That is, the user's past behavior data (such past behavior data includes but is not limited to browsing records, click records, sharing records, etc.) is collected according to the user's unique identifier (for example, user ID, etc.), and the collected past behavior data is cleaned and pre-processed to remove duplicates to obtain text data.

[0065] The second step is feature extraction and quantification. That is, the word frequency and inverse document frequency are used to extract features from the above text data to obtain feature words in the text data. After that, a machine learning model (the Word2vec model is used in this example) is used to convert the extracted feature words into high dimensions to obtain high-dimensional word vectors.

[0066] The third step is to calculate the weight coefficient. That is, through the weight formula W i =α·log(1+Activity i )+β·TimeDecay i Calculate the weight coefficient. i represents the user's activity on behavior i, TimeDecay iIt represents the time decay factor (used to consider the impact of the time when the behavior occurs on the weight), and α and β are specific parameter values ​​preset by the administrator.

[0067] The fourth step is to introduce the weight coefficient W into the high-dimensional word vector, weight the features, and obtain the weighted word vector.

[0068] The fifth step is to calculate the degree of preference. That is, the cosine similarity is used to calculate the preference score of each dimension. Specifically, the preference score is calculated using the formula:

[0069]

[0070] Among them, Vector i Represents the word vector of the i-th feature word, CategoryVector j represents the center vector of the jth preference category (obtained through the clustering algorithm), n is the number of feature words, and k is the number of preference categories.

[0071] Step 6: Construct a multi-dimensional character portrait. Based on the preference scores of the above users, cluster and grade the users in each preference category to form character portrait dimensions. Then, extract and analyze the user's features to complete the character portrait construction of the user's preference degree.

[0072] Preferably, the character portrait can be visualized using a radar chart or a heat map. Also, an update cycle is set, and after the update cycle is reached, the character portrait is updated and iterated.

[0073] In step S102, based on the dimensions of the character portrait, a multi-dimensional analysis of the outbound call content may be performed to generate an outbound call content portrait.

[0074] In some optional embodiments of this example, the outbound call content is generated based on the degree of matching with the user's character portrait. Specifically, firstly, the character portrait dimensions of all users are counted to construct a dimension union. Then, the intersection related to the outbound call content is found out, so as to construct the outbound call content dimension.

[0075] Then, based on the outbound call content dimension, the features of the outbound call information to be sent are extracted and analyzed to generate an outbound call content portrait.

[0076] In step S103, a preset matching degree threshold may be set, and the dimensions of the character portrait and the outbound call content portrait may be matched using the preset matching degree threshold to generate an outbound call matching degree.

[0077] In some optional embodiments of this example, first, multiple dimensions in the outbound call content portrait are obtained, and these dimensions are collectively referred to as key dimensions. At the same time, the matching degree of each key dimension in the outbound call content portrait is calculated to obtain a content key score.

[0078] Afterwards, each dimension in the character portrait is compared with the corresponding key dimension in the outbound call content portrait, and the matching degree of each key dimension in the character portrait is calculated to obtain the character key score.

[0079] Then, according to the pre-set matching threshold, the product of the above-mentioned content key score, character key score and preset matching threshold is calculated to obtain the dimension score of the final character portrait in the outbound call content dimension.

[0080] Finally, the dimension scores of all dimensions are summarized and sorted to obtain the matching degree between the persona and the outbound call content.

[0081] In step S104, according to the outbound call matching degree, the outbound call content can be synthesized into a reply for voice feedback to complete the intelligent call to the outbound caller.

[0082] In some optional embodiments of this example, after obtaining the matching degree between the character and the outbound call content, the task of the outbound call content can be set.

[0083] Assume that the target of this task is set to personnel waiting for outbound calls.

[0084] When establishing a communication connection with the person to be called, first, the voice signal sent by the person to be called is picked up to generate input language. In the process of picking up the voice signal, it is necessary to use the Gaussian mixture general background channel model to fit the human voice system through unsupervised learning (the fitting process is to construct different specific person models according to the voice characteristics of different users). The specific picking process is as follows:

[0085] The first step is the data preparation stage, which involves preparing speech data containing 100 to 200 speakers.

[0086] In the second step, a clustering algorithm (K-Means clustering algorithm is used in this example) is used to preliminarily group the data, 1024 classes are set (i.e., components of the Gaussian mixture general background channel model), and 5 iterations are performed to obtain the center and variance information of the class.

[0087] The third step is to further optimize the clustering results obtained by the above K-Means clustering algorithm through the expectation maximization (EM) algorithm to improve the accuracy of the model.

[0088] The fourth step is to build a specific person model for each speaker based on the optimized voiceprint identification vector data.

[0089] In the fifth step, the system performs speech recognition and uses the trained background model to identify whether the current input voice belongs to the person waiting for an outbound call.

[0090] In the sixth step, the system extracts voiceprint features from the received voice data. These voiceprint features reflect the speaker's personalized pronunciation pattern. Then, the system constructs a voiceprint feature vector based on these features for further recognition and comparison.

[0091] In the seventh step, the system will identify the voice signal with the most common voiceprint features, assuming that this signal comes from the target person of the call, that is, the voice signal of the person to be called out.

[0092] In the eighth step, the system will continue to collect voice data and continuously extract voiceprint feature vectors, and then input these feature vectors into the pre-trained background channel model (i.e., the specific person model trained by the Gaussian mixture general background channel model), and compare them with the voiceprint identification vector.

[0093] The ninth step is to set a distance threshold, and based on this, calculate the spatial distance between the voiceprint feature vector of the current caller and the voiceprint identification vector of the person to be called.

[0094] When the spatial distance between the voiceprint identification vector and the voiceprint feature vector is not less than the preset distance threshold, the speech signal is determined to be noise and is discarded.

[0095] When the spatial distance between the voiceprint identification vector and the voiceprint feature vector is less than a preset distance threshold, the voice signal is determined to be the voice of the caller and the input voice is generated.

[0096] Afterwards, a silence model is established based on the long speech segmentation technology (this example uses the Mode1Vad technology), and a deep learning algorithm is used to recognize the language of each frame of the continuous speech signal in the input speech to identify the silent language.

[0097] Then, the continuous speech signal in the input speech is divided into a plurality of speech segments by taking the silent speech as the segmentation point, thereby obtaining an input short speech set consisting of a plurality of input short speech segments.

[0098] Then, each of the segmented input short voices is independently recognized to obtain one-to-one corresponding recognition results, and these recognition results are logically combined to obtain the intention of the intended person to be called out.

[0099] Finally, according to the intention of the intended person, the outbound call content is matched through the above outbound call matching degree to obtain the reply content that meets the wishes of the outbound caller. And according to the speech synthesis technology, the reply content is speech synthesized to generate speech feedback to complete the intelligent call for the outbound caller.

[0100] In this way, the intelligent outbound call system can accurately match the character portrait with the outbound call content portrait, and make targeted outbound calls to outbound callers, ensuring higher recognition accuracy.

[0101] At the same time, through semantic segmentation technology and voiceprint recognition noise shielding technology, the recognition accuracy and communication efficiency are improved on the other hand, and ultimately the order conversion rate is improved through a two-pronged approach.

[0102] It should be noted that, for the method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that the present disclosure is not limited by the order of the actions described, because according to the present disclosure, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all optional embodiments, and the actions involved are not necessarily required by the present disclosure.

[0103] Reference Figure 2 , shows a structural block diagram of an intelligent outbound calling device 200 based on an AI algorithm of the present disclosure, the device comprising:

[0104] The character portrait module 210 is used to construct a character portrait by using a machine learning model to construct a multi-dimensional portrait of the collected user's past behavior data;

[0105] The outbound call content portrait module 220 is used to generate an outbound call content portrait by performing multi-dimensional analysis on the outbound call content;

[0106] The matching module 230 is used to set a preset matching degree threshold, and match the dimensions of the character portrait and the outbound call content portrait by the preset matching degree threshold to generate an outbound call matching degree;

[0107] The outbound call module 240 is used to complete the intelligent call to the outbound caller by synthesizing the outbound call content into a reply for voice feedback.

[0108] The character portrait module 210 is provided with:

[0109] The data collection and processing module 211 is used to collect, clean and remove duplicate data of customers' past behaviors;

[0110] A feature extraction module 212, used for extracting features from text data;

[0111] A weight coefficient calculation module 213 is used to calculate the weight coefficient for weighted processing of features;

[0112] A preference degree calculation module 214 is used to calculate the preference score of each dimension;

[0113] The character portrait building module 215 is used to build a character portrait according to the preference score.

[0114] The outbound call content portrait module 220 is provided with:

[0115] A dimension matching module 221, used to match the outbound call content with the union of the character portrait dimensions;

[0116] The outbound call content portrait construction module 222 is used to construct an outbound call content portrait based on the dimension matching results.

[0117] The matching module 230 is provided with:

[0118] A character portrait matching score calculation module 231 is used to calculate the matching degree of the key dimension in the character portrait;

[0119] The outbound call content portrait matching score calculation module 232 is used to calculate the matching degree of key dimensions in the outbound call content portrait.

[0120] The outbound call module 240 is provided with:

[0121] A voice pickup module 241, used to pick up voice data in the communication connection;

[0122] A voice segmentation module 242, used for segmenting the picked-up voice data;

[0123] A speech recognition module 243, used for independently recognizing the segmented speech data;

[0124] The speech synthesis module 244 is used to synthesize the outbound call content into speech and provide speech feedback to the outbound caller.

[0125] The intelligent outbound calling device 200 based on AI algorithm significantly improves the accuracy and efficiency of outbound calls through accurate user portrait construction, intelligent content matching and efficient voice interaction technology.

[0126] First, the character portrait module 210 can construct a multi-dimensional user portrait based on the user's historical behavior data, making the outbound call content more personalized and customized, thereby improving user participation and satisfaction.

[0127] Secondly, the matching module 230 ensures that each outbound call is highly consistent with user needs by calculating the matching degree between the user portrait and the outbound call content portrait, thereby reducing invalid outbound calls and improving the success rate of outbound calls.

[0128] Furthermore, the outbound call module 240 combines speech recognition and speech synthesis technology to achieve natural and smooth voice interaction, so that users can obtain more real and immediate feedback, thereby improving the interactive experience.

[0129] Finally, the efficient collaboration of the voice pickup module 241, the voice segmentation module 242, the voice recognition module 243, and the voice synthesis module 244 ensures the accuracy and rapid response of the outbound calling process and avoids manual operation errors.

[0130] Overall, the device optimizes the entire outbound call process through the synergy of multiple modules, making it more personalized, efficient and automated. It is widely used in the fields of intelligent customer service, marketing promotion, etc., and improves the overall effect and user experience of the outbound call system.

[0131] Optionally, an embodiment of the present disclosure further provides an electronic device, comprising: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, the various processes of the above-mentioned method embodiment are implemented and the same technical effect can be achieved. To avoid repetition, it will not be described here.

[0132] The embodiment of the present disclosure also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, each process of the above method embodiment is implemented, and the same technical effect can be achieved. To avoid repetition, it is not repeated here. The computer-readable storage medium is, for example, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0133] Figure 3 800 is a block diagram of an electronic device 800 shown in the present disclosure. For example, the electronic device 800 may be a mobile phone, a computer, a digital broadcast terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, etc.

[0134] Reference Figure 3 , the electronic device 800 may include one or more of the following components: a processing component 802 , a memory 804 , a power component 806 , a multimedia component 808 , an audio component 810 , an input / output (I / O) interface 812 , a sensor component 814 , and a communication component 816 .

[0135] The processing component 802 generally controls the overall operation of the electronic device 800, such as operations associated with display, phone calls, data communications, camera operations, and recording operations. The processing component 802 may include one or more processors 820 to execute instructions to complete all or part of the steps of the above method. In addition, the processing component 802 may include one or more modules to facilitate the interaction between the processing component 802 and other components. For example, the processing component 802 may include a multimedia module to facilitate the interaction between the multimedia component 808 and the processing component 802.

[0136] The memory 804 is configured to store various types of data to support operations on the device 800. Examples of such data include instructions for any application or method operating on the electronic device 800, contact data, phone book data, messages, images, videos, etc. The memory 804 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0137] The power supply component 806 provides power to the various components of the electronic device 800. The power supply component 806 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to the electronic device 800.

[0138] The multimedia component 808 includes a screen that provides an output interface between the electronic device 800 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundaries of the touch or slide action, but also detect the duration and pressure associated with the touch or slide operation. In some embodiments, the multimedia component 808 includes a front camera and / or a rear camera. When the device 800 is in an operating mode, such as a shooting mode or a video mode, the front camera and / or the rear camera may receive external multimedia data. Each front camera and the rear camera may be a fixed optical lens system or have a focal length and optical zoom capability.

[0139] The audio component 810 is configured to output and / or input audio signals. For example, the audio component 810 includes a microphone (MIC), and when the electronic device 800 is in an operating mode, such as a call mode, a recording mode, and a voice recognition mode, the microphone is configured to receive an external audio signal. The received audio signal can be further stored in the memory 804 or sent via the communication component 816. In some embodiments, the audio component 810 also includes a speaker for outputting audio signals.

[0140] I / O interface 812 provides an interface between processing component 802 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include but are not limited to: home button, volume button, start button, and lock button.

[0141] The sensor assembly 814 includes one or more sensors for providing various aspects of status assessment for the electronic device 800. For example, the sensor assembly 814 can detect the open / closed state of the device 800, the relative positioning of components, such as the display and keypad of the electronic device 800, and the sensor assembly 814 can also detect the position change of the electronic device 800 or a component of the electronic device 800, the presence or absence of contact between the user and the electronic device 800, the orientation or acceleration / deceleration of the electronic device 800, and the temperature change of the electronic device 800. The sensor assembly 814 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. The sensor assembly 814 may also include an optical sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 814 may also include an accelerometer, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

[0142] The communication component 816 is configured to facilitate wired or wireless communication between the electronic device 800 and other devices. The electronic device 800 can access a wireless network based on a communication standard, such as WiFi, a carrier network (such as 2G, 3G, 4G or 5G), or a combination thereof. In an exemplary embodiment, the communication component 816 receives a broadcast signal or broadcast operation information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 816 also includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.

[0143] In an exemplary embodiment, the electronic device 800 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above methods.

[0144] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 804 including instructions, and the instructions can be executed by a processor 820 of an electronic device 800 to perform the above method. For example, the non-transitory computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.

[0145] Figure 4 19 is a block diagram of a computer-readable storage medium 1900 shown in the present disclosure. For example, the computer-readable storage medium 1900 may be provided as a server.

[0146] Reference Figure 4 , the computer-readable storage medium 1900 includes a processing component 1922, which further includes one or more processors, and a memory resource represented by a memory 1932 for storing instructions, such as an application, that can be executed by the processing component 1922. The application stored in the memory 1932 may include one or more modules, each corresponding to a set of instructions. In addition, the processing component 1922 is configured to execute instructions to perform the above method.

[0147] The computer readable storage medium 1900 may also include a power supply component 1926 configured to perform power management of the computer readable storage medium 1900, a wired or wireless network interface 1950 configured to connect the computer readable storage medium 1900 to a network, and an input / output (I / O) interface 1958. The computer readable storage medium 1900 may operate based on an operating system stored in the memory 1932, such as Windows Server™, Mac OS X™, Unix™, Linux™, FreeBSD™ or the like.

[0148] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0149] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present disclosure, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a disk, or an optical disk), and includes a number of instructions for enabling a terminal (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present disclosure.

[0150] The embodiments of the present disclosure are described above in conjunction with the accompanying drawings, but the present disclosure is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present disclosure, ordinary technicians in this field can also make many forms without departing from the scope of protection of the purpose of the present disclosure and the claims, all of which are within the protection of the present disclosure.

[0151] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the present disclosure can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this disclosure.

[0152] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0153] In the embodiments provided in the present disclosure, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0154] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0155] In addition, each functional unit in each embodiment of the present disclosure may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0156] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present disclosure, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present disclosure. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, ROM, RAM, magnetic disks or optical disks.

[0157] The above is only a specific implementation of the present disclosure, but the protection scope of the present disclosure is not limited thereto. Any person skilled in the art who is familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present disclosure, which should be included in the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be based on the protection scope of the claims.

Claims

1. An intelligent outbound calling method based on AI algorithm, characterized in that: The method comprises: Step 1: Use machine learning models to construct multi-dimensional portraits of collected user past behavior data to build character portraits; Step 2: Based on the dimensions of the character portrait, a multi-dimensional analysis is performed on the outbound call content to generate an outbound call content portrait; Step 3, setting a preset matching degree threshold, and matching the dimensions of the character portrait and the outbound call content portrait by the preset matching degree threshold to generate an outbound call matching degree; Step 4, according to the outbound call matching degree, the outbound call content is synthesized into a reply for voice feedback, thereby completing the intelligent call for the outbound caller.

2. The method according to claim 1, characterized in that By using machine learning models, we build multi-dimensional portraits of the collected user behavior data, including: Based on the user's unique identifier, text data is generated by preprocessing the user's previous behavior data; Based on the machine learning model, feature extraction is performed on the text data to generate feature words, and high-dimensional word vectors are generated by performing high-dimensional transformation on the feature words; Based on the time when the behavior occurred, the weight coefficient is generated by weighting the user's activity in the behavior; Based on the weight coefficient, weighting the high-dimensional word vector to generate a weighted word vector; Based on the weighted word vector, the dimensional preference scores of the character portraits are calculated by using a cosine similarity algorithm to generate a preference score; Based on the preference scores, generating character portrait dimensions by clustering and grading the user's preferences on the dimensions; Based on the character portrait dimensions, a character portrait is generated by extracting and analyzing the user's features.

3. The method according to claim 1, characterized in that Through multi-dimensional analysis of outbound call content, including: Based on the match between the outbound call content and the user, the outbound call content dimension is generated by calculating the intersection of the outbound call content and the character portrait dimension; Based on the outbound call content dimension, an outbound call content portrait is generated by extracting and analyzing the features of the outbound call information.

4. The method according to claim 1, characterized in that: By matching the dimensions of the character portrait and the outbound call content portrait, including: Based on a preset matching degree threshold, a key score of the character is generated by calculating the matching degree of the key dimension in the character portrait; Based on a preset matching degree threshold, a content key score is generated by calculating the matching degree of the key dimension in the outbound call content portrait; A dimension score is generated by calculating the product of the character key score and the content key score, and an outbound call matching degree is generated by summarizing the dimension scores.

5. The method according to claim 1, characterized in that Provide voice feedback by synthesizing the outbound call content into a reply, including: When a communication connection is established, an input voice is generated by picking up a voice signal in the communication connection; Based on the long speech segmentation technology, the input speech is segmented to generate an input short speech set, and the short speech in the input short speech set is recognized to generate the intention of the intended person; Based on the intention of the intended person, the outbound call content is matched by the outbound call matching degree to generate reply content; Based on speech synthesis technology, the reply content is speech synthesized to generate speech feedback, thereby completing the intelligent call for the outbound caller.

6. The method according to claim 5, characterized in that By picking up the voice signal in the communication connection, including: Based on the preset background channel model, the voiceprint identification vector is generated by training the voice of the outbound caller; When a communication connection is established, a voiceprint feature vector is generated by extracting the voiceprint feature of the speech signal in the communication connection; Based on a preset distance threshold, when the spatial distance between the voiceprint identification vector and the voiceprint feature vector is not less than the preset distance threshold, determining that the voice signal is noise; Based on a preset distance threshold, when the spatial distance between the voiceprint identification vector and the voiceprint feature vector is less than the preset distance threshold, the voice signal is determined to be the voice of the caller, and the input voice is generated.

7. The method according to claim 5, characterized in that The input speech is segmented, including: Based on a preset silence model, a continuous voice signal in the input voice is recognized by a deep learning algorithm to generate a silent voice; Taking the silent speech as a segmentation point, the input speech is segmented to generate an input short speech set.

8. An intelligent outbound calling device based on AI algorithm, characterized in that: include: The character portrait module is used to construct a character portrait by using a machine learning model to construct a multi-dimensional portrait of the collected user's past behavior data; The outbound call content portrait module is used to generate an outbound call content portrait by performing multi-dimensional analysis on the outbound call content; A matching module, used to set a preset matching degree threshold, and match the dimensions of the character portrait and the outbound call content portrait by the preset matching degree threshold to generate an outbound call matching degree; The outbound call module is used to complete intelligent calls to outbound callers by synthesizing the outbound call content into a reply for voice feedback.

9. An electronic device, characterized in that: include: A processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the computer program implements the method according to any one of claims 1 to 7 when executed by the processor.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images