CN115334455A - Position determining method and device based on voice recognition and electronic equipment - Google Patents

Abstract

The invention discloses a method, device and electronic equipment for position determination based on speech recognition. The method includes: acquiring a target address book, wherein the target address book contains multiple numbers; configuring a speech template and a robot for each number in the target address book; controlling each robot to make a voice call to the corresponding number; In the case of successful connection, control each robot to determine the location according to the speech template, obtain the location information of the user corresponding to each number, and generate a report. The invention solves the technical problems of low efficiency and waste of manpower caused by the traditional location determination method due to huge data.

Description

Method, device and electronic device for determining position based on speech recognition

technical field

The present invention relates to the field of computers, in particular, to a method, device and electronic equipment for determining a location based on speech recognition.

Background technique

In the existing technology, when it is necessary to determine the location information of the floating personnel, it is necessary to communicate with the floating personnel through the staff, which creates a huge work pressure for the staff. In the case of limited staff, the huge data leads to work efficiency. Low, unable to meet the timeliness.

Contents of the invention

Embodiments of the present invention provide a speech recognition-based location determination method, device, and electronic equipment to at least solve the technical problems of low efficiency and waste of manpower caused by huge data in traditional location determination methods.

According to an aspect of an embodiment of the present invention, a method for determining a location based on speech recognition is provided, including: acquiring a target address book, wherein the target address book contains multiple numbers; configuring each of the numbers in the target address book Speech template and robot; control each of the above-mentioned robots to make a voice call to the corresponding above-mentioned number; in the case that the above-mentioned voice call is successfully connected, control each of the above-mentioned robots to determine the position according to the above-mentioned speech template, and obtain each The location information of the user corresponding to the above number is generated to generate a report.

According to another aspect of the embodiments of the present invention, there is provided a device for determining a location based on speech recognition, including: a first acquisition module, configured to acquire a target address book, wherein the target address book contains multiple numbers; a configuration module, It is used to configure speech templates and robots for each of the above-mentioned numbers in the above-mentioned target address book; the first control module is used to control each of the above-mentioned robots to make voice calls to the corresponding above-mentioned numbers; When the call is successfully connected, each of the above-mentioned robots is controlled to determine the location according to the above-mentioned speech template, obtain the location information of the user corresponding to each of the above-mentioned numbers, and generate the above-mentioned report.

As an optional example, the above-mentioned device further includes: a first obtaining module, configured to obtain a number group whose location needs to be determined before obtaining the target address book; a processing module, configured to set each number in the above-mentioned number group as the current Number, perform the following operations on the above-mentioned current number: obtain the digit of the above-mentioned current number; if the above-mentioned digit is the first value, obtain the value of the first digit of the above-mentioned current number; the value in the above-mentioned first digit is the first In the case of a binary value, the value of the second digit of the above-mentioned current number is obtained; when the value of the above-mentioned second digit is not the third value, the above-mentioned current number is added to the above-mentioned target address book.

As an optional example, the above-mentioned second control module includes: a processing unit, configured to set each of the above-mentioned robots as the current robot, and perform the following operations on the above-mentioned current robot: control the above-mentioned current robot to send the first A speech technique, so as to obtain the affirmative answer or negative answer or other answers given by the above-mentioned user to the above-mentioned first speech technique; obtain the first answer given by the above-mentioned user according to the above-mentioned first speech technique; in the above-mentioned first answer In the case where the above-mentioned affirmative answer is used, according to the above-mentioned speech template, control the above-mentioned robot to issue a second speech, so as to obtain the location information of the above-mentioned user; obtain the second answer issued by the above-mentioned user according to the above-mentioned second speech; The location information is obtained from the second answer, and at the same time, according to the speech template, the robot is controlled to issue a third speech, so that the location determination work is completed, the location information is recorded, and the report is generated.

As an optional example, the above-mentioned device further includes: a third control module, configured to control the above-mentioned robot to issue The above-mentioned third means is to complete the above-mentioned position determination work, record that the above-mentioned position determination work is not successful, and generate the above-mentioned report.

As an optional example, the above-mentioned device further includes: a fourth control module, configured to be activated by the user during the process of controlling the above-mentioned robot to issue the above-mentioned first speech or the above-mentioned second speech or the above-mentioned third speech In the case of interruption, control the above-mentioned robot to suspend sending the above-mentioned first speech or the above-mentioned second speech or the above-mentioned third speech; the fifth control module is used to set a timer, when the above-mentioned timer reaches the first time Under the circumstances, control the above-mentioned robot to continue or reissue the above-mentioned first speech or the above-mentioned second speech or the above-mentioned third speech.

As an optional example, the processing unit further includes: obtaining the affirmative thesaurus and the negative thesaurus in the answer lexicon after obtaining the first answer fed back by the above-mentioned user according to the above-mentioned first speech technique; database, the above-mentioned negative thesaurus and the above-mentioned first answer to carry out feature extraction; calculate the first matching degree between the feature of the above-mentioned first answer and the feature of the above-mentioned affirmative thesaurus, and the feature of the above-mentioned first answer and the above-mentioned negative thesaurus The second matching degree of the feature; when the above-mentioned first matching degree is greater than or equal to the first threshold, and the above-mentioned second matching degree is less than the above-mentioned first threshold, it is determined that the above-mentioned first answer is the above-mentioned positive answer; in the above-mentioned first When the degree of matching is less than the first threshold and the second degree of matching is greater than or equal to the first threshold, it is determined that the first answer is the negative answer; when the first degree of matching is less than the first threshold, the second If the degree of matching between the two is smaller than the above-mentioned first threshold, it is determined that the above-mentioned first answer is the above-mentioned other answer.

As an optional example, the processing unit further includes: obtaining the number of occurrences of the first answer when the first matching degree is less than the threshold and the second matching degree is less than the threshold; or equal to the second threshold, carry out semantic matching to the above-mentioned first answer; in the case that the above-mentioned first answer is matched as affirmative, save the above-mentioned first answer in the above-mentioned affirmative lexicon; If the answer is matched as negative, the above-mentioned first answer is stored in the above-mentioned negative lexicon.

According to yet another aspect of the embodiments of the present invention, there is also provided a storage medium, in which a computer program is stored, wherein, when the computer program is run by a processor, the above method for determining a position based on speech recognition is executed.

According to yet another aspect of the embodiments of the present invention, there is also provided an electronic device, including a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to execute the above speech recognition-based The location determination method.

In the embodiment of the present invention, the acquisition of the target address book is adopted, wherein the above-mentioned target address book contains a plurality of numbers; a speech template and a robot are configured for each of the above-mentioned numbers in the above-mentioned target address book; each of the above-mentioned robots is controlled to correspond to The above-mentioned number makes a voice call; when the above-mentioned voice call is successfully connected, control each of the above-mentioned robots to perform position determination according to the above-mentioned speech template, obtain the location information of the user corresponding to each of the above-mentioned numbers, and generate a report. Because in the above method, through the use of voice recognition technology, the robots are controlled to automatically initiate voice calls in batches at the same time, use real recordings to inquire about location information, and automatically jump to logical questions according to the speech template according to the voice of the answer, and finally recognize the voice of the answer Analyze, obtain the user's location information, and form a report, so as to achieve the purpose of fast and efficient location determination, and then solve the technical problems of low efficiency and waste of manpower caused by the huge data of the traditional location determination method.

Description of drawings

The accompanying drawings described here are used to provide a further understanding of the present invention and constitute a part of the application. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations to the present invention. In the attached picture:

FIG. 1 is a flow chart of an optional speech recognition-based position determination method according to an embodiment of the present invention;

FIG. 2 is a diagram of an optional speech recognition-based position determination method according to an embodiment of the present invention;

Fig. 3 is a diagram of an optional speech recognition-based location determination method according to an embodiment of the present invention;

Fig. 4 is a diagram of an optional speech recognition-based position determination method according to an embodiment of the present invention;

FIG. 5 is a diagram of an optional speech recognition-based location determination method according to an embodiment of the present invention;

Fig. 6 is a schematic structural diagram of an optional speech recognition-based location determination device according to an embodiment of the present invention;

Fig. 7 is a schematic diagram of an optional electronic device according to an embodiment of the present invention.

Detailed ways

In order to enable those skilled in the art to better understand the solutions of the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only It is an embodiment of a part of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts shall fall within the protection scope of the present invention.

It should be noted that the terms "first" and "second" in the description and claims of the present invention and the above drawings are used to distinguish similar objects, but not necessarily used to describe a specific sequence or sequence. It is to be understood that the data so used are interchangeable under appropriate circumstances such that the embodiments of the invention described herein can be practiced in sequences other than those illustrated or described herein. Furthermore, the terms "comprising" and "having", as well as any variations thereof, are intended to cover a non-exclusive inclusion, for example, a process, method, system, product or device comprising a sequence of steps or elements is not necessarily limited to the expressly listed instead, may include other steps or elements not explicitly listed or inherent to the process, method, product or apparatus.

According to the first aspect of the embodiments of the present invention, a method for determining a position based on speech recognition is provided. Optionally, as shown in FIG. 1 , the above method includes:

S102, acquiring a target address book, wherein the target address book contains multiple numbers;

S104, configuring a speech template and a robot for each number in the target address book;

S106, controlling each robot to make a voice call to the corresponding number;

S108, when the voice call is successfully connected, control each robot to determine the position according to the speech template, obtain the position information of the user corresponding to each number, and generate a report.

Optionally, in this embodiment, the target address book contains one or more mobile phone numbers of users who need to determine the location. The speech template includes multiple nodes, and each node corresponds to a paragraph or a sentence. After the robot finishes speaking a paragraph or a sentence, it will automatically jump to the corresponding node according to the user's answer, and the robot will automatically speak the speech corresponding to the node. . The robot is an intelligent outbound robot that can make outbound calls in large quantities at the same time.

Optionally, in this embodiment, the correct number is selected from a plurality of numbers whose positions need to be determined to form a target address book, a speech template and a robot are configured for each number in the target address book, and each number data is sent to To the corresponding robot, it is transmitted to the operator’s digital-to-analog converter through the network, that is, it can make a voice call to the corresponding number through the network communication system. After the user connects to the voice call, each robot will respond to each user according to the speech template. To determine the location, the robot sends out each question in the speech template, records each user's answer to each question, and obtains each user's location information from the answer, which can include specific location information. Generate a corresponding report based on each user's answer, so that the subsequent location determination staff can view it.

Optionally, in this embodiment, through the use of voice recognition technology, the robots are controlled to automatically initiate voice calls in batches at the same time, use real recordings to inquire about location information, and automatically jump to logical questions according to the speech template according to the voice of the answer, and finally answer the question The user's location information is obtained through speech recognition and analysis, and a report is formed, so as to achieve the purpose of fast and efficient location determination, and then solve the technical problems of low efficiency and waste of manpower caused by the huge data of the traditional location determination method.

As an optional example, before obtaining the target address book, it also includes:

Obtain the number group whose location needs to be determined;

Set each number in the number group as the current number, and perform the following operations on the current number:

Get the number of digits in the current number;

In the case that the digit is the first value, get the value of the first digit of the current number;

In the case that the value of the first digit is the second value, obtain the value of the second digit of the current number;

If the value of the second digit is not the third value, the current number is added to the target address book.

Optionally, in this embodiment, the number group whose location needs to be determined is acquired. The number group may include one or more numbers, and there may be wrong numbers. In order to prevent the robot from calling out the wrong number and wasting time, the number group Check all the numbers in , and add the correct number to the target address book. Check each number, for example, check the first number, obtain the digits of the first number, and judge whether the digits of the first number are the first value. The first value can be 11 digits. If the first number is 11 In the case of digits, get the first digit of the first number, and judge whether the first digit of the first number is the second value, the second value can be 1, if the first digit of the first number is 1 Next, get the second digit of the first number, judge whether the second digit of the first number is the third value, the third value can be 1 and 2, if the second digit of the first number is 3 , determine that the first number is the correct number, and add the first number to the target address book. If the digits of the first number are not 11 digits, the first digit of the first number is not 1, and the second digit of the first number When the number of digits is 1 or 2, it is determined that the first number is an incorrect number.

As an optional example, when the voice call is successfully connected, control each robot to determine the location according to the speech template, obtain the location information of the user corresponding to each number, and generate a report including:

Set each robot as the current robot, and perform the following operations on the current robot:

According to the speech template, control the current robot to issue the first speech, so as to obtain the user's affirmative answer or negative answer or other answers for the first speech feedback;

Obtain the user's first answer based on the first speech feedback;

In the case that the first answer is an affirmative answer, according to the speech template, control the robot to issue the second speech to obtain the user's location information;

Obtain the second answer issued by the user according to the second speech;

The location information is obtained from the second answer, and at the same time, according to the speech template, the robot is controlled to issue the third speech to complete the location determination work, record the location information, and generate a report.

Optionally, in this embodiment, multiple robots can make voice calls to the corresponding numbers at the same time. For example, the first robot makes a voice call to the corresponding second number in the target address book. After the user corresponding to the second number is connected, The robot sends out the first utterance in the utterance template according to the utterance template, so as to obtain the user's affirmative answer or negative answer or other answers to the first utterance feedback. The first utterance can be "Hello, we Now I will ask you a few small questions, our call will be recorded, and I hope to get your cooperation!", get the user's first answer to the first speech feedback, the first answer can be an affirmative answer, for example, "Well Um", "OK", can be negative answers "no", "impossible", can be other answers, other answers are not recognized by the robot, or answers that have not been received. If the user receives an affirmative answer of "ummm", the robot will automatically jump to the second script according to the script template. The second script can be "Please tell me your current place of residence in detail" to obtain The user's second answer for the second speech feedback, such as "I am in the E community of the D street in the C district of the city B in the province A", obtains the location information from the second answer "The E community of the D street in the C district of the city B in the province A" , record the location information, and generate a report on the completion of the user location determination work corresponding to the second number. Finally, the robot sends out the second speech in the speech template according to the speech template, which can be "Thank you for answering, I wish you a happy life, goodbye!", so that the position determination work ends. The script template can contain one or more scripts, and set the jump relationship of each script when different answers are received.

As an optional example, the above method further includes:

When the first answer is a negative answer or other answers, according to the speech template, the robot is controlled to issue a third speech, so that the position determination work ends, the position determination work is recorded as unsuccessful, and a report is generated.

Optionally, in this embodiment, when the first answer is a negative answer or the robot does not recognize it, or the robot does not receive the answer, the robot sends out the third sentence in the speech template according to the speech template. technique to bring position determination to an end. Generate a report of unsuccessful location determination efforts.

As an optional example, the above method further includes:

When the robot is interrupted by the user during the process of sending out the first speech, the second speech or the third speech, the control robot stops sending the first speech or the second speech or the third speech;

A timer is set, and when the timer reaches the first time, the robot is controlled to continue or resend the first speech or the second speech or the third speech.

Optionally, in this embodiment, if the robot is interrupted by the user during the process of sending out the first speech, the second speech, or the third speech, the robot pauses immediately, and the duration of the pause is set as the first time , the first time is 1s, 2s, after pausing the first time, continue or reissue the first speech or the second speech or the third speech.

As an optional example, after acquiring the user's first answer based on the first utterance feedback, include:

Obtain the affirmative thesaurus and negative thesaurus in the answer lexicon;

Perform feature extraction on positive thesaurus, negative thesaurus and first answer;

Calculate the first degree of matching between the features of the first answer and the features of the affirmative lexicon, and the second degree of matching between the features of the first answer and the features of the negative lexicon;

When the first degree of matching is greater than or equal to the first threshold and the second degree of matching is less than the first threshold, it is determined that the first answer is an affirmative answer;

When the first degree of matching is less than the first threshold and the second degree of matching is greater than or equal to the first threshold, it is determined that the first answer is a negative answer;

When the first matching degree is smaller than the first threshold and the second matching degree is smaller than the first threshold, it is determined that the first answer is another answer.

Optionally, in this embodiment, an answer lexicon is created, including an affirmative lexicon and a negative lexicon, and each lexicon contains one or more answer words. Perform feature extraction on all answers in the affirmative thesaurus, all answers in the negative thesaurus and the first answer to obtain the features of all the answers in the affirmative thesaurus, the features of all the answers in the negative thesaurus and the first answer Features, calculate the first matching degree of the features of the first answer and the features of the positive lexicon, the second matching degree of the features of the first answer and the features of the negative lexicon, the first matching degree and the second matching degree can be 80%, 90%, the first threshold can be 85%, for example, the first matching degree is 90%, the second matching degree is 50%, that is, the first matching degree is greater than the first threshold, and the second matching degree is smaller than the first threshold, Determine that the first answer is an affirmative answer. If the first matching degree is 50% and the second matching degree is 90%, that is, the first matching degree is less than the first threshold and the second matching degree is greater than the first threshold, determine the first answer. If the first matching degree is 50% and the second matching degree is 40%, that is, the first matching degree is less than the first threshold and the second matching degree is less than the first threshold, it is determined that the first answering language is other. language.

As an optional example, the above method further includes:

When the first degree of matching is less than the threshold and the second degree of matching is less than the threshold, the number of occurrences of the first answer is obtained;

When the number of occurrences is greater than or equal to the second threshold, perform semantic matching on the first answer;

In the case that the first answer is matched as affirmative, the first answer is stored in the affirmative lexicon;

In the case that the first answer is matched as negative, the first answer is stored in a negative lexicon.

Optionally, in this embodiment, when the first degree of matching is less than the threshold and the second degree of matching is less than the threshold, the number of occurrences of the first answer is obtained. If it is recorded that the first answer has appeared 20 times, the second threshold It can be 10 times. When the number of occurrences is 20 times greater than the second threshold of 10 times, perform semantic matching on the first answer, and if the first answer is affirmative after semantic matching, save the first answer in affirmative Thesaurus, when the first answer is negative, save the first answer in the negative thesaurus. If it is not identified as affirmative or negative after semantic matching, the staff will be prompted to perform manual matching.

Optionally, in this embodiment, the present invention relates to a location determination method based on speech recognition. In order to ensure the fast and efficient location determination work, a batch outbound call technology is innovatively introduced to solve the problem of a large number of personnel and a small number of staff; innovative The use of speech templates to ensure that the answers automatically called out to ask location questions have a certain logic orientation, and all fixed and repetitive questions are asked once; at the same time, a knowledge base is established based on the content of the answers to facilitate learning and analysis, and conduct in-depth Repeated dialing of matching and invalid numbers.

1. The present invention proposes a visual speech template editing function. By editing the speech question and answer logic online, the fixed and repeated questions are solidified in the form of a template, and different question and answer paths can be selected according to the content of the answer. After being answered, an automatic voice broadcast can be made, and the next question path can be selected according to the content of the answer;

2. The present invention proposes the use of VoIP (Voice over Internet Protocol IP-based voice transmission) technology to develop batch number outbound capabilities. Through the multi-threaded intelligent outbound robot, a large number of outbound inquiries can be made at the same time, and the answering status can be recorded at the same time , screen abnormal numbers, and make a second call for those who did not answer;

3. The present invention proposes the function of establishing a knowledge base for learning. Through the reply content collected in the communication process, self-learning and active learning can be performed, and the question-and-answer logic can be improved to make the location determination work more accurate and efficient.

In order to meet the requirements of different speech skills in different scenarios, the scene speech template method is innovatively designed, which supports visual template editing and design. Each template consists of multiple nodes, and the nodes can be defined according to the content of the answer. Association logic between nodes.

First of all, it is necessary to determine the work content for the position of a specific scene and analyze it, sort out a speech template, and carry out logical analysis, logically label the speech, and organize it according to the template form of node-condition-content-result-jump to form A speech template for a specific scene.

Determine the work content based on the location, analyze and sort out the scene speech template as shown in Table 1 below:

(Table 1)

Once you have a scene speech template, you can perform a visual system configuration. First, you must complete the customization of the nodes, and at the same time define the template for the new node and the visual presentation form. Because there are many types of nodes, if you want to uniformly customize the configuration, you need to sort out the attribute model of the nodes, as shown in Table 2 below:

(Table 2)

The node logic setting is to connect the query logic between each node in series. The core problem lies in the construction of the customer reply model. Based on the reply model, the node logical relationship can be directly generated. The steps are as follows:

1. Define the answer model, the definition of the node reply model is shown in Table 3 below:

(table 3)

2. Exhaustively list the content of the answers, because the user will have many kinds of answers to the question of a node, and it is necessary to cover the answers as much as possible, so as to ensure that there is a corresponding path to execute the logic when the user answers, so that the location is determined Work more fluently and smoothly. There is a general principle for exhaustive enumeration, which is defined as: affirmative answer*2+negative answer*2+exception*2, taking node 2 as an example to illustrate, the question of node 2 is "Excuse me, are you in province A now?" Is it?", an exhaustive example of the answers is shown in Table 4 below:

(Table 4)

Numbering Response name Response type 1 in province A Affirmative answer 2 Yes Affirmative answer 3 not in province A negative answer 4 No negative answer 5 unknown exception answer 6 no response exception answer

3. Improve the answer lexicon. The answer lexicon is the knowledge base of the answer, which can assist the robot to quickly identify the answer content that contains and has similar meanings, and improve the intelligence level of the robot. The knowledge base is a process of gradual improvement. 2 Exhaustively enumerate the content of the answers to improve the answer lexicon, as an example, as shown in Table 5 below:

(table 5)

4. The connection forms the logic, and connects the user reply result and the corresponding reply logic result with a line.

When the script template is edited, you can configure the outbound task. The outbound task uses VOIP technology innovatively. Because it is a packaged application service, it can conveniently support multiple services at the same time. The intelligent location determination system sends data to a single number through a robot (background service) through a visual interface, and transmits it to the operator's digital-to-analog converter through the network, that is, it can make a voice call through the network communication system. After the answer is converted by the operator's analog-to-digital converter, it is sent to the intelligent position determination system. Through the above process, the whole process of position determination is realized. In order to serve batch calls, it is necessary to import batches of mobile phone numbers in advance. At the same time, each application service can be independently controlled, and automatic question-answer path selection can be performed according to the script template and answer voice. The specific detailed flow chart is shown in Figure 2.

The batch import of mobile phone numbers that need to determine the location adopts the common Excel import method, but adds the mobile phone number verification rule: the mobile phone number has 11 digits, the first digit is 1, and the second digit can be 3, 4, Any one of 5, 7, 8, and 9, the third to eleventh digits can be arbitrarily selected from 0-9, and the decision logic is shown in Figure 3.

After the batch import is judged, there is a verification result as shown in Table 6 below. After the batch import of mobile phone numbers is completed, a task can be created to realize automatic dialing.

(Table 6)

serial number cellphone number check result 1 1358657342* Y 2 12586573423 N 3 1358657342 N 4 1358657342A N 5 … …

There are several key configuration steps for task automatic dialing, which are distributed as speech template, number and number of robots. After the configuration is completed and the task is started, the intelligent position determination of the robot can be realized. The specific working method needs to be automatically identified in combination with the answer. , the specific steps are shown in Figure 4.

When the task is started, the multi-threading of the robot is started to determine the position. In this process, the audio comparison is directly used instead of being translated into text, and the answer lexicon is converted into voice and directly compared with the user's answer voice. The accuracy rate is improved and the recognition efficiency is improved; the specific automatic recognition process is shown in Figure 5.

Taking node 2 as an example to illustrate, the question and answer lexicon of node 2 is shown in Table 7 below:

(Table 7)

1. Perform feature extraction on the answer lexicon, as shown in Table 8 below:

(Table 8)

Answer thesaurus feature extraction I'm here wo zai a yes shi a I am not in province A wo bu zai a sheng No mei you

2. Play the node phrase "Excuse me, are you in province A now?";

3. Receive the answer signal: "I am here";

4. Response feature extraction: "wo zai a";

5. Feature matching: the matching degree of the answer feature "wo zai a" and the answer lexicon "wo zai a" is as high as 100%, which is greater than the matching threshold of 85%, and it is judged as "in province A";

6. Record the knowledge base, because it matches the thesaurus, the answer will not be recorded in the knowledge base.

By adopting an innovative method, the automatic recognition of the answer is realized, which not only improves the accuracy rate but also improves the recognition efficiency, and can ensure the accuracy and timeliness of the position determination work.

For local dialects and special language expression methods, the intelligent voice position determination system has been compatible, but there are still some omissions. In this way, these voice expressions can be recorded in the knowledge base and continuously improved through knowledge learning. The knowledge base learning process must first accumulate the knowledge base, and then it can be confirmed and improved through automatic knowledge learning and artificial knowledge learning.

The accumulation of knowledge base is divided into two parts, one part is the pre-organized thesaurus, and the other part is obtained by analyzing the recordings of unknown problems. Analyzing unknown question recordings is divided into automatic knowledge learning and artificial knowledge learning. Through the study of unknown questions, the passed answers can be added to the thesaurus of the corresponding questions, and the failed answers will be marked as ignored. Therefore, it is very important to sort out the unknown problems. The attribute model of the sorted unknown problems is shown in Table 9 below:

(Table 9)

Numbering Attributes illustrate 1 Response content The voice of the user's answer is converted into text 2 Number of hits The number of times the answer was repeated in the location determination exercise 3 record time The time when the answer was recorded 4 phone number The phone number where the location needs to be determined 5 Number of user replies Number of replies from users 6 call result answering status 7 speech template Referenced utterance template name 8 dial time dialed time 9 call time The actual length of the overall call 10 Call Recording Recording of the entire call, presented as a dialogue 11 operate 3 actions: view, learn and ignore

Artificial knowledge learning is based on the content of the answer, learning the artificial knowledge base, and the learning process is as follows:

1. According to the speech template, sort in reverse order according to the number of hits. Usually, the value of learning the answer with a large number of hits is higher;

2. Find the answer that needs to be learned manually, and judge whether it is necessary to manually create the knowledge base by listening to the recorded dialogue and obtaining the context before and after. If not, choose to ignore, and the answer will enter the ignored list to exclude the unknown Question list; if it is necessary to create an artificial knowledge base, go to the next step;

3. If the number of answers hits is relatively large, it can be added to the knowledge base after manual judgment. You only need to select Learning in the operation, and you can choose to match the speech template, node, condition, content, answer name and answer word library. For example, the answer "Yes" is added to the answer lexicon after manual matching, as marked in red.

Automatic knowledge learning is based on the content of the answer, keyword information learning, and automatic knowledge base node matching, which improves efficiency compared with manual knowledge learning. The automatic learning algorithm is rule matching. Through the setting of rules and thresholds, the system can automatically match whether the answer needs to be added to the corresponding knowledge base. The matching logic of rules and thresholds is shown in Table 10 below:

(Table 10)

Numbering rule Rules Threshold 1 contains matches Determine whether the answer contains the answer lexicon 100% 2 feature matching Determine the degree of matching between the characteristics of the answer and the characteristics of the lexicon 95% 3 hit match The number of times the judgment answer hits 50 4 semantic matching Determine whether the semantic judgment of the answer is affirmative or negative 100%

Taking the answer "um" as an example to illustrate, the specific recognition steps are as follows:

1. Obtain the answer "um", and perform feature extraction as "ng";

2. According to the rules, match item by item. Once the rules are met, the match will jump out. By matching "um" and the answer lexicon, "um" is not in the answer lexicon, and it does not match the inclusion match; by matching "ng" and the lexicon The feature, "ng" does not match the feature of the lexicon, and it does not meet the feature matching; by matching the number of hits of "um", it is 40 hits, which does not meet the hit rules; by matching the semantics of "um", it means yes or affirmation, There are also doubts, not 100% match positive or negative, so it does not meet the semantic match;

3. Through rule matching calculation, if no rules are matched, the answer "um" will not be added to the knowledge base. Because the number of hits is relatively large, it will prompt for manual matching.

To sum up, through manual and automatic knowledge learning, the answer lexicon is continuously improved to ensure the completeness and accuracy of the dialogue.

It should be noted that for the foregoing method embodiments, for the sake of simple description, they are expressed as a series of action combinations, but those skilled in the art should know that the present invention is not limited by the described action sequence. Because of the present invention, certain steps may be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification belong to preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.

According to another aspect of the embodiment of the present application, there is also provided a device for determining a location based on speech recognition, as shown in FIG. 6 , including:

The first obtaining module 602 is used to obtain the target address book, wherein the target address book contains multiple numbers;

Configuration module 604, configured to configure speech templates and robots for each number in the target address book;

The first control module 606 is used to control each robot to make a voice call to the corresponding number;

The second control module 608 is used to control each robot to determine the location according to the speech template when the voice call is successfully connected, obtain the location information of the user corresponding to each number, and generate a report.

Optionally, in this embodiment, the target address book contains one or more mobile phone numbers of users who need to determine the location. The speech template includes multiple nodes, and each node corresponds to a paragraph or a sentence. After the robot finishes speaking a paragraph or a sentence, it will automatically jump to the corresponding node according to the user's answer, and the robot will automatically speak the speech corresponding to the node. . The robot is an intelligent outbound robot that can make outbound calls in large quantities at the same time.

Optionally, in this embodiment, the correct number is selected from a plurality of numbers whose positions need to be determined to form a target address book, a speech template and a robot are configured for each number in the target address book, and each number data is sent to To the corresponding robot, it is transmitted to the operator’s digital-to-analog converter through the network, that is, it can make a voice call to the corresponding number through the network communication system. After the user connects to the voice call, each robot will respond to each user according to the speech template. To determine the location, the robot sends out each question in the speech template, records each user's answer to each question, and obtains each user's location information from the answer, which can include specific location information. Generate a corresponding report based on each user's answer, so that the subsequent location determination staff can view it.

Optionally, in this embodiment, through the use of voice recognition technology, the robots are controlled to automatically initiate voice calls in batches at the same time, use real recordings to inquire about location information, and automatically jump to logical questions according to the speech template according to the voice of the answer, and finally answer the question The user's location information is obtained through speech recognition and analysis, and a report is formed, so as to achieve the purpose of fast and efficient location determination, and then solve the technical problems of low efficiency and waste of manpower caused by the huge data of the traditional location determination method.

As an optional example, the above-mentioned device also includes:

The first obtaining module is used to obtain the number group whose location needs to be determined before obtaining the target address book;

The processing module is used to set each number in the number group as the current number, and perform the following operations on the current number:

Get the digit of the current number; if the digit is the first value, get the value of the first digit of the current number;

In the case that the value of the first digit is the second value, obtain the value of the second digit of the current number;

If the value of the second digit is not the third value, the current number is added to the target address book.

Optionally, in this embodiment, the number group whose location needs to be determined is acquired. The number group may include one or more numbers, and there may be wrong numbers. In order to prevent the robot from calling out the wrong number and wasting time, the number group Check all the numbers in , and add the correct number to the target address book. Check each number, for example, check the first number, obtain the digits of the first number, and judge whether the digits of the first number are the first value. The first value can be 11 digits. If the first number is 11 In the case of digits, get the first digit of the first number, and judge whether the first digit of the first number is the second value, the second value can be 1, if the first digit of the first number is 1 Next, get the second digit of the first number, judge whether the second digit of the first number is the third value, the third value can be 1 and 2, if the second digit of the first number is 3 , determine that the first number is the correct number, and add the first number to the target address book. If the digits of the first number are not 11 digits, the first digit of the first number is not 1, and the second digit of the first number When the number of digits is 1 or 2, it is determined that the first number is an incorrect number.

As an optional example, the second control module includes:

The processing unit is used to set each robot as the current robot, and perform the following operations on the current robot:

According to the speech template, control the current robot to issue the first speech, so as to obtain the user's affirmative answer or negative answer or other answers for the first speech feedback;

Obtain the user's first answer based on the first speech feedback;

In the case that the first answer is an affirmative answer, according to the speech template, control the robot to issue the second speech to obtain the user's location information;

Obtain the second answer issued by the user according to the second speech;

The location information is obtained from the second answer, and at the same time, according to the speech template, the robot is controlled to issue the third speech to complete the location determination work, record the location information, and generate a report.

Optionally, in this embodiment, multiple robots can make voice calls to the corresponding numbers at the same time. For example, the first robot makes a voice call to the corresponding second number in the target address book. After the user corresponding to the second number is connected, The robot sends out the first utterance in the utterance template according to the utterance template, so as to obtain the user's affirmative answer or negative answer or other answers to the first utterance feedback. The first utterance can be "Hello, we Now I will ask you a few small questions, our call will be recorded, and I hope to get your cooperation!", get the user's first answer to the first speech feedback, the first answer can be an affirmative answer, for example, "Well Um", "OK", can be negative answers "no", "impossible", can be other answers, other answers are not recognized by the robot, or answers that have not been received. If the user receives an affirmative answer of "ummm", the robot will automatically jump to the second script according to the script template. The second script can be "Please tell me your current place of residence in detail" to obtain The user's second answer for the second speech feedback, such as "I am in the E community of the D street in the C district of the city B in the province A", obtains the location information from the second answer "The E community of the D street in the C district of the city B in the province A" , record the location information, and generate a report on the completion of the user location determination work corresponding to the second number. Finally, the robot sends out the second speech in the speech template according to the speech template, which can be "Thank you for answering, I wish you a happy life, goodbye!", so that the position determination work ends. The script template can contain one or more scripts, and set the jump relationship of each script when different answers are received.

As an optional example, the above-mentioned device also includes:

The third control module is used to control the robot to issue a third speech according to the speech template when the first answer is a negative answer or other answers, so that the position determination work ends and the position determination work is recorded as unsuccessful , to generate a report.

Optionally, in this embodiment, when the first answer is a negative answer or the robot does not recognize it, or the robot does not receive the answer, the robot sends out the third sentence in the speech template according to the speech template. technique to bring position determination to an end. Generate a report of unsuccessful location determination efforts.

As an optional example, the above-mentioned device also includes:

The fourth control module is used to control the robot to suspend the sending of the first speech or the second speech or the third speech in the case of being interrupted by the user during the process of controlling the robot to send the first speech or the second speech or the third speech or the third speech;

The fifth control module is used to set a timer, and when the timer reaches the first time, control the robot to continue or reissue the first speech, the second speech or the third speech.

Optionally, in this embodiment, if the robot is interrupted by the user during the process of sending out the first speech, the second speech, or the third speech, the robot pauses immediately, and the duration of the pause is set as the first time , the first time is 1s, 2s, after pausing the first time, continue or reissue the first speech or the second speech or the third speech.

As an optional example, the processing unit also includes:

Obtain the positive thesaurus and the negative thesaurus in the answer lexicon after obtaining the first answer fed back by the user according to the first speech technique;

Perform feature extraction on positive thesaurus, negative thesaurus and first answer;

Calculate the first degree of matching between the features of the first answer and the features of the affirmative lexicon, and the second degree of matching between the features of the first answer and the features of the negative lexicon;

When the first degree of matching is greater than or equal to the first threshold and the second degree of matching is less than the first threshold, it is determined that the first answer is an affirmative answer;

When the first degree of matching is less than the first threshold and the second degree of matching is greater than or equal to the first threshold, it is determined that the first answer is a negative answer;

When the first matching degree is smaller than the first threshold and the second matching degree is smaller than the first threshold, it is determined that the first answer is another answer.

Optionally, in this embodiment, an answer lexicon is created, including an affirmative lexicon and a negative lexicon, and each lexicon contains one or more answer words. Perform feature extraction on all answers in the affirmative thesaurus, all answers in the negative thesaurus and the first answer to obtain the features of all the answers in the affirmative thesaurus, the features of all the answers in the negative thesaurus and the first answer Features, calculate the first matching degree of the features of the first answer and the features of the positive lexicon, the second matching degree of the features of the first answer and the features of the negative lexicon, the first matching degree and the second matching degree can be 80%, 90%, the first threshold can be 85%, for example, the first matching degree is 90%, the second matching degree is 50%, that is, the first matching degree is greater than the first threshold, and the second matching degree is smaller than the first threshold, Determine that the first answer is an affirmative answer. If the first matching degree is 50% and the second matching degree is 90%, that is, the first matching degree is less than the first threshold and the second matching degree is greater than the first threshold, determine the first answer. If the first matching degree is 50% and the second matching degree is 40%, that is, the first matching degree is less than the first threshold and the second matching degree is less than the first threshold, it is determined that the first answering language is other. language.

As an optional example, the processing unit also includes:

When the first degree of matching is less than the threshold and the second degree of matching is less than the threshold, the number of occurrences of the first answer is obtained;

When the number of occurrences is greater than or equal to the second threshold, perform semantic matching on the first answer;

In the case that the first answer is matched as affirmative, the first answer is stored in the affirmative lexicon;

In the case that the first answer is matched as negative, the first answer is stored in a negative lexicon.

Optionally, in this embodiment, when the first degree of matching is less than the threshold and the second degree of matching is less than the threshold, the number of occurrences of the first answer is obtained. If it is recorded that the first answer has appeared 20 times, the second threshold It can be 10 times. When the number of occurrences is 20 times greater than the second threshold of 10 times, perform semantic matching on the first answer, and if the first answer is affirmative after semantic matching, save the first answer in affirmative Thesaurus, when the first answer is negative, save the first answer in the negative thesaurus. If it is not identified as affirmative or negative after semantic matching, the staff will be prompted to perform manual matching.

For other examples of this embodiment, refer to the above examples, and details are not repeated here.

FIG. 7 is a structural block diagram of an optional electronic device according to an embodiment of the present application. As shown in FIG. 704 and memory 706 complete mutual communication through communication bus 708, wherein,

memory 706, for storing computer programs;

When the processor 702 is used to execute the computer program stored on the memory 706, the following steps are implemented:

Obtain the target address book, wherein the target address book contains multiple numbers;

Configure speech templates and robots for each number in the target address book;

Control each robot to make a voice call to the corresponding number;

When the voice call is successfully connected, control each robot to determine the location according to the speech template, obtain the location information of the user corresponding to each number, and generate a report.

Optionally, in this embodiment, the aforementioned communication bus may be a PCI (Peripheral Component Interconnect, Peripheral Component Interconnect Standard) bus, or an EISA (Extended Industry Standard Architecture, Extended Industry Standard Architecture) bus, or the like. The communication bus can be divided into an address bus, a data bus, a control bus, and the like. For ease of representation, only one thick line is used in FIG. 7 , but it does not mean that there is only one bus or one type of bus. The communication interface is used for communication between the electronic device and other devices.

The memory may include RAM, and may also include non-volatile memory (non-volatile memory), for example, at least one magnetic disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

As an example, the memory 706 may include, but is not limited to, the first acquisition module 602 , the configuration module 604 , the first control module 606 and the second control module 608 in the speech recognition-based location determination apparatus. In addition, it may also include but not limited to other module units in the above request processing apparatus, which will not be repeated in this example.

Above-mentioned processor can be general-purpose processor, can include but not limited to: CPU (Central Processing Unit, central processing unit), NP (Network Processor, network processor) etc.; Can also be DSP (DigitalSignal Processing, digital signal processor), ASIC (Application Specific Integrated Circuit, Application Specific Integrated Circuit), FPGA (Field-Programmable Gate Array, Field Programmable Gate Array) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

Optionally, for specific examples in this embodiment, reference may be made to the examples described in the foregoing embodiments, and details are not repeated in this embodiment.

Those of ordinary skill in the art can understand that the structure shown in Figure 7 is only for illustration, and the device implementing the above-mentioned method for determining a position based on speech recognition may be a terminal device, and the terminal device may be a smart phone (such as an Android phone, an iOS phone, etc.) , Tablet PC, PDA, and mobile Internet devices (Mobile Internet Devices, MID), PAD and other terminal equipment. FIG. 7 does not limit the structure of the above-mentioned electronic equipment. For example, the electronic device may also include more or less components than those shown in FIG. 7 (such as a network interface, a display device, etc.), or have a different configuration from that shown in FIG. 7 .

Those skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing hardware related to the terminal device through a program, and the program can be stored in a computer-readable storage medium, and the storage medium can be Including: flash disk, ROM, RAM, magnetic disk or optical disk, etc.

According to yet another aspect of the embodiments of the present invention, a computer-readable storage medium is also provided, and a computer program is stored in the computer-readable storage medium, wherein, when the computer program is run by a processor, the above-mentioned voice-based Steps in the location determination method are identified.

Optionally, in this embodiment, those skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing hardware related to the terminal device through a program, and the program can be stored in a In the computer-readable storage medium, the storage medium may include: a flash disk, a read-only memory (Read-Only Memory, ROM), a random access device (Random Access Memory, RAM), a magnetic disk or an optical disk, and the like.

The serial numbers of the above embodiments of the present invention are for description only, and do not represent the advantages and disadvantages of the embodiments.

If the integrated units in the above embodiments are realized in the form of software function units and sold or used as independent products, they can be stored in the above computer-readable storage medium. Based on this understanding, the essence of the technical solution of the present invention or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium. Several instructions are included to make one or more computer devices (which may be personal computers, servers or network devices, etc.) execute all or part of the steps of the methods described in the various embodiments of the present invention.

In the above-mentioned embodiments of the present invention, the descriptions of each embodiment have their own emphases, and for parts not described in detail in a certain embodiment, reference may be made to relevant descriptions of other embodiments.

In the several embodiments provided in this application, it should be understood that the disclosed client can be implemented in other ways. Wherein, the device embodiments described above are only illustrative, for example, the division of the units is only a logical function division, and there may be other division methods in actual implementation, for example, multiple units or components can be combined or can be Integrate into another system, or some features may be ignored, or not implemented. In another point, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces, and the indirect coupling or communication connection of units or modules may be in electrical or other forms.

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 may be distributed to multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, each unit may exist separately physically, or two or more units may be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

The above is only a preferred embodiment of the present invention, it should be pointed out that, for those of ordinary skill in the art, without departing from the principle of the present invention, some improvements and modifications can also be made, and these improvements and modifications can also be made. It should be regarded as the protection scope of the present invention.

Claims (10)

1. A method for determining position based on speech recognition, comprising: Obtaining a target address book, wherein the target address book contains multiple numbers; Configuring a speech template and a robot for each of the numbers in the target address book; Controlling each of the robots to make a voice call to the corresponding number; When the voice call is successfully connected, each of the robots is controlled to determine the location according to the speech template, to obtain the location information of the user corresponding to each of the numbers, and to generate a report. 2. The method according to claim 1, characterized in that, before obtaining the target address book, the method further comprises: Obtain the number group whose location needs to be determined; Set each number in the number group as the current number, and perform the following operations on the current number: Obtain the digits of the current number; In the case that the digit is the first value, obtain the value of the first digit of the current number; In the case that the value of the first digit is the second value, obtain the value of the second digit of the current number; If the value of the second digit is not the third value, adding the current number to the target address book. 3. The method according to claim 1, characterized in that, when the voice call is successfully connected, each of the robots is controlled to determine the position according to the speech template to obtain each The location information of the user corresponding to the number, generating the report includes: Set each robot as the current robot, and perform the following operations on the current robot: According to the utterance template, control the current robot to issue a first utterance, so as to obtain an affirmative answer or a negative answer or other answers from the user for the first utterance feedback; Obtaining the first answer fed back by the user according to the first speech; When the first answer is the affirmative answer, according to the speech template, control the robot to issue a second speech, so as to obtain the location information of the user; Obtaining a second answer issued by the user according to the second speech technique; The location information is obtained from the second answer, and at the same time, according to the speech template, the robot is controlled to issue a third speech, so that the work of determining the location is completed, the location information is recorded, and the speech is generated. Report. 4. method according to claim 3, is characterized in that, described method also comprises: In the case that the first answer is the negative answer or the other answer, according to the speech template, control the robot to issue the third speech, so that the position determination work ends , logging that the location determination job was unsuccessful, generating the report. 5. method according to claim 3, is characterized in that, described method also comprises: When the process of controlling the robot to issue the first speech or the second speech or the third speech is interrupted by the user, control the robot to suspend sending the first speech script or said second script or said third script; A timer is set, and when the timer reaches the first time, the robot is controlled to continue or reissue the first speech or the second speech or the third speech. 6. The method according to claim 3, characterized in that, after obtaining the first answer fed back by the user according to the first speech technique, the method comprises: Obtain the affirmative thesaurus and negative thesaurus in the answer lexicon; Carrying out feature extraction to the positive thesaurus, the negative thesaurus and the first answer; calculating the first degree of matching between the features of the first answer and the features of the affirmative thesaurus, and the second matching degree between the features of the first answer and the features of the negative thesaurus; When the first matching degree is greater than or equal to a first threshold and the second matching degree is smaller than the first threshold, determining that the first answer is the affirmative answer; When the first matching degree is less than the first threshold and the second matching degree is greater than or equal to the first threshold, determining that the first answer is the negative answer; If the first matching degree is smaller than the first threshold and the second matching degree is smaller than the first threshold, it is determined that the first answer is the other answer. 7. The method according to claim 6, further comprising: If the first degree of matching is less than the threshold and the second degree of matching is less than the threshold, acquiring the number of occurrences of the first answer; In the case where the number of occurrences is greater than or equal to a second threshold, perform semantic matching on the first answer; In the case that the first answer is matched as affirmative, storing the first answer in the affirmative lexicon; If the first answer is negatively matched, the first answer is stored in the negative lexicon. 8. A position determination device based on speech recognition, characterized in that it comprises: The first obtaining module is used to obtain a target address book, wherein the target address book contains multiple numbers; A configuration module, configured to configure a speech template and a robot for each of the numbers in the target address book; The first control module is used to control each of the robots to make voice calls to the corresponding numbers; The second control module is configured to control each of the robots to perform position determination according to the speech template to obtain the position information of the user corresponding to each of the numbers when the voice call is successfully connected, Generate report. 9. A computer-readable storage medium, the computer-readable storage medium stores a computer program, characterized in that, when the computer program is executed by a processor, the computer program according to any one of claims 1 to 7 is executed. described method. 10. An electronic device comprising a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to execute any one of claims 1 to 7 through the computer program. the method described.

Images