WO2019174072A1 - 智能机器人培训方法、装置、计算机设备及存储介质 - Google Patents

智能机器人培训方法、装置、计算机设备及存储介质 Download PDF

Info

Publication number
WO2019174072A1
WO2019174072A1 PCT/CN2018/081504 CN2018081504W WO2019174072A1 WO 2019174072 A1 WO2019174072 A1 WO 2019174072A1 CN 2018081504 W CN2018081504 W CN 2018081504W WO 2019174072 A1 WO2019174072 A1 WO 2019174072A1
Authority
WO
WIPO (PCT)
Prior art keywords
search
target
request
presentation
basic
Prior art date
Application number
PCT/CN2018/081504
Other languages
English (en)
French (fr)
Inventor
张政
Original Assignee
平安科技（深圳）有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 平安科技（深圳）有限公司 filed Critical 平安科技（深圳）有限公司
Publication of WO2019174072A1 publication Critical patent/WO2019174072A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/30—Semantic analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/20—Natural language analysis
    • G06F40/205—Parsing
    • G06F40/216—Parsing using statistical methods
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue

Definitions

  • the present application relates to the field of artificial intelligence technologies, and in particular, to an intelligent robot training method, apparatus, computer equipment, and storage medium.
  • AI Artificial Intelligence
  • Artificial intelligence is a technology that studies how to create smart devices or intelligent systems to simulate human intelligence activities.
  • Intelligent robots are devices that use artificial intelligence. With the rapid development of artificial intelligence technology, intelligent robots are also developing rapidly.
  • the embodiment of the present application provides an intelligent robot training method, device, computer equipment and storage medium to solve the problem that the traditional professional training cannot meet the individualized needs of the personnel to be trained.
  • an embodiment of the present application provides a smart robot training method, including:
  • the target presentation is displayed while playing the target handout voice data associated with the target presentation.
  • an intelligent robot training apparatus including:
  • a mode selection module configured to obtain a mode selection request, and enter the training presentation interface if the mode selection request is a training speech mode request;
  • a search request obtaining module configured to acquire a search request, and acquire at least one search keyword based on the search request
  • a target search module configured to search a training database based on at least one of the search keywords, and acquire interrelated target presentations and target handout voice data;
  • an embodiment of the present application provides a computer device, including a memory, a processor, and computer readable instructions stored in the memory and executable on the processor, where the processor executes the computer The following steps are implemented when reading the instruction:
  • Obtaining a search request acquiring at least one search keyword based on the search request
  • the target presentation is displayed while playing the target handout voice data associated with the target presentation.
  • the embodiment of the present application provides one or more non-volatile readable storage media storing computer readable instructions, when the computer readable instructions are executed by one or more processors, such that the one or Multiple processors perform the following steps:
  • Obtaining a search request acquiring at least one search keyword based on the search request
  • the target presentation is displayed while playing the target handout voice data associated with the target presentation.
  • FIG. 1 is a flowchart of an implementation of an intelligent robot training method provided in Embodiment 1 of the present application;
  • step S20 in FIG. 1 is a flowchart of an implementation of step S20 in FIG. 1;
  • FIG. 3 is a flowchart of an implementation of step S30 in FIG. 1;
  • FIG. 5 is a flowchart of another implementation of the smart robot training method provided in Embodiment 1 of the present application.
  • FIG. 7 is a schematic diagram of an intelligent robot training device provided in Embodiment 2 of the present application.
  • FIG. 8 is a schematic diagram of a computer device provided in Embodiment 4 of the present application.
  • FIG. 1 shows an implementation flow of the intelligent robot training method provided by this embodiment.
  • the intelligent robot training method is applied to the intelligent robot for human-computer interaction with the personnel to be trained, so as to personally train the trainer based on the relevant request input by the user.
  • the customer service training applied in the insurance financial industry is explained. Details are as follows:
  • An intelligent robot training method includes the following steps performed by an intelligent robot:
  • S10 Acquire a mode selection request, and enter a training presentation interface if the mode selection request is a training speech mode request.
  • the mode selection request refers to a request for controlling the intelligent robot to enter the corresponding mode.
  • different training modes such as a training speech mode, a practical exercise mode, and a classroom practice mode, are set in advance according to the business nature of different industries and the needs of business training.
  • the training speech mode mainly displays the relevant courseware by the intelligent robot and treats the training personnel through the lecture.
  • the training mode is theoretical, the knowledge points are concentrated, and the pertinence is strong, so that the trainees can quickly receive the need to understand. knowledge.
  • the practice drill mode is mainly based on the scenario where the intelligent robot gives simulation related problems, and the training personnel are simulated.
  • the training mode has a strong sense of situation and deepens the memory of the trainees through interaction.
  • the classroom practice mode is to display relevant exercises according to the progress of the course by the intelligent robot, to evaluate or train the trainers, and to consolidate the knowledge received by the trainees.
  • the mode configuration of the intelligent robot training in the embodiment is exemplarily a training speech mode and a practical exercise mode. .
  • the training presentation interface is an interface for the user to operate on the intelligent robot, and is also an initial program interface for the training speech mode.
  • the intelligent robot acquires the mode selection request of the training speech mode selected by the user, it enters the training presentation interface, and the system starts the initial procedure of the training speech mode.
  • the intelligent robot sets different training modes according to the business nature, business training or other requirements of different industries in advance, so that each trainer can independently input the corresponding mode selection request according to his own needs and the training progress. Bring intelligent robots into different training modes to meet the individual needs of the trainees.
  • the user needs to register in advance through the client on the corresponding training webpage, and after the registration is successful, log in the intelligent robot based on the corresponding user account and password, and the intelligent robot is based on the pre-recorded each time the user successfully logs in.
  • User information interacts with the user. For example, the use of the user's name and the gender extension in the title makes the human-computer interaction more substitutable.
  • the intelligent robot can call the learning history record of each user account from its corresponding database for storing user information, and can recommend information to the user based on the latest learning history record, and specifically recommend the user to select different training. Mode, you can also recommend users to choose different courseware or other training-related information, so that users can learn effectively with their own learning progress during training and learning.
  • S20 Acquire a search request, the search request including at least one search keyword.
  • the search request is a request that the user sends to the intelligent robot to acquire the knowledge point that wants to be trained.
  • the search request may be a voice search request input by the user using a recording device (such as a microphone) or a text search request input by the user using an input device (such as a keyboard), and the input method of the search request may be selected according to the user's habit.
  • a voice search request is a voice-based search request
  • a text search request is a text-based search request.
  • the search request includes at least one search keyword related to the training content.
  • the search keyword is a keyword extracted by the user according to the training knowledge point that he wants to acquire.
  • the keywords extracted may be car insurance sales, car insurance business, and the like.
  • the search keyword can be extracted by the user and input into the intelligent robot to form a search request, or the user can input a search sentence input by using a natural language, and the keyword extraction algorithm built by the intelligent robot can perform keyword extraction. To obtain corresponding at least one search keyword.
  • S30 Search a training database based on at least one search keyword to obtain a correlated target presentation and target handout voice data.
  • the training database is a database for storing training materials, and specifically stores pre-recorded collected interrelated lecture voice data and presentations.
  • the database can be a MySQL database, an Oracle database, or another database.
  • a search formula may be formed based on at least one search keyword and a corresponding search logical connector (such as and, not, and or), and the training database is searched based on the search formula to obtain corresponding to at least one search keyword. Interrelated target presentations and target handouts for voice data.
  • the presentation includes documents such as PPTs, videos, and pictures that can be presented through the display screen of the smart robot.
  • the handout voice data may include not only an audio file explaining the PPT, video, and pictures, but also an audio file explaining the documents such as test papers, textbooks, and exercises.
  • the interrelated target presentation and the target speech voice data specifically mean that each presentation page of the target presentation is correspondingly configured with target speech data for explaining the content of the presentation page.
  • the interrelated speech data and the presentation are pre-stored, and the corresponding interrelated association can be obtained by using a simple search query operation according to at least one of the search keywords input by the user.
  • Target presentation and target handout voice data the operation process is simple and convenient, and the interrelated handout voice data and presentation can be repeatedly searched by different trainees, which can achieve the purpose of multiple recordings once, so that the training process can be Significantly reduce the reliance on training instructors and reduce training costs.
  • S40 Display the target presentation and play the target handout voice data associated with the target presentation.
  • the intelligent robot After acquiring the target target presentation and the target handout voice data, the intelligent robot performs voice playback through the built-in player or the external playing device of the intelligent robot; meanwhile, the target presentation is displayed through the screen of the intelligent robot. Or project through the projector to the large screen for display.
  • This display of the target presentation while playing the associated target speech data, enables the intelligent robot to have a sense of substitution for the training of the trainers, making it easier for the trainees to receive the knowledge to be trained.
  • the presentation and the speech data are correlated with each other through an associated program built in the intelligent robot, and the presentation of the page identifier of a certain page is associated with the corresponding speech data according to the presentation, and the presentation is played.
  • the lecture speech data associated with the presentation page is played. For example, if the intelligent robot is currently displayed as the Nth page of a target presentation, the intelligent robot simultaneously plays the target handout voice data that matches the Nth page presentation page according to the associated program.
  • the desired target presentation and the target handwritten voice data are selected according to the needs of the personnel to be trained, so as to conform to the personality of the trainee to be trained. Demand.
  • the intelligent robot displays the target presentation and simultaneously plays the target speech data to realize the training of the training personnel, which makes the training process more flexible, reduces the training pressure of the training instructors, and solves the problem of lack of resources for the training instructors.
  • the intelligent robot training method can realize one-to-one training for the trainees with different needs, which can be more convenient for training and more suitable for the development of the enterprise.
  • step S20 a search request is obtained, and the search request includes at least one search keyword, which specifically includes the following steps:
  • S201 Acquire a voice search request, where the voice search request includes target voice search data.
  • the voice search request is a voice-based search request, that is, in the present embodiment, the intelligent robot acquires the user's target voice search data by collecting and recognizing the user's voice.
  • Step S201 specifically adopts the following steps to obtain target voice search data:
  • the intelligent robot uses the microphone array to collect the original voice search data.
  • the original voice search data refers to voice data collected by the intelligent robot through the microphone array but not processed, and the voice data contains related information of the voice search request.
  • the microphone array adopts a distributed array, and the sub-array or sub-array can be arranged in a larger range, so that the sub-array or the sub-array exchanges and shares data with each other by wire or wirelessly, and Based on this, the techniques of sound source localization and beamforming in a broad sense are implemented to achieve signal processing.
  • the intelligent robot further processes the original voice search data by using a de-reverberation algorithm to obtain the target voice search data.
  • the dereverberation algorithm adopts a Beamforming based approach to form a pickup beam in the signal direction of the original voice search data by weighting the signals of the original voice search data collected by the microphone array. While attenuating the reflected sound from other directions.
  • a method based on beamforming that is, by separately forming a sound beam to sound sources in different directions and suppressing sounds in other directions, the original voice search data is extracted or separated to obtain target voice search data.
  • the target voice search data refers to the voice data collected by the intelligent robot through the microphone array and extracted from the original voice search data by using the de-reverberation algorithm, and the target voice search data contains the information of the voice search request. It can be understood that through the distributed array of microphone arrays, the original voice search data in the space can be acquired from multiple directions in a wider range, so that the intelligent robot acquires a larger range of the user voice search request, and is more conducive to receiving the voice search request. By using the de-reverberation algorithm to extract the target voice search data from the original voice search data, the target voice search data can be extracted from the original voice search data more accurately, and the more accurate voice search request information can be obtained.
  • S202 Perform conversion processing on the target voice search data by using a preset voice recognition system, and obtain text search data.
  • the speech recognition system is a system that converts speech into words using a built-in speech recognition algorithm.
  • the intelligent robot after receiving the target voice search data, the intelligent robot calls a preset voice recognition system to process the target voice search data to obtain text search data.
  • S203 Analyze the text search data by using the NLP library to obtain at least one search keyword.
  • the NLP (Natural Language Processing) library is a natural language processing database of artificial intelligence, and is a database for matching natural language analysis to a machine language that can be recognized by a computer.
  • the NLP library can be developed using the Stanford CoreNLP developed by the Stanford Natural Speech Processing team to quickly parse text search data to obtain at least one search keyword.
  • the intelligent robot collects the voice search request input by the user through the built-in microphone array, and uses the voice recognition system to convert the target voice search data in the voice search request into text search data, and uploads the text search data. Semantic parsing to the NLP library to facilitate access to at least one search keyword quickly and easily. The corresponding search keyword is obtained based on the voice search request, so that the user can select the voice input mode to input the search request, and the operation process is simple and convenient, and it is easier to meet the user's needs.
  • the search request may include text search data entered by the user via an input device such as a keyboard and/or mouse.
  • the intelligent robot acquires text search data input by the user through an input device such as a keyboard and/or a mouse, and then parses the text search data by using the NLP library to obtain at least one search keyword, and the processing procedure is the same as step S203. Avoid narration and no longer describe them one by one.
  • the search request can be accurately input to improve the search accuracy.
  • step S30 the training database is searched based on at least one search keyword, and the associated target presentation and target handout voice data are obtained, which specifically includes the following steps:
  • S301 Using each search keyword as a search field, processing at least one search field by using a logical relationship to form a target search formula.
  • the corresponding logical relationship is combined for the at least one search keyword as the target search formula.
  • the plurality of search keywords are searched by logical relationship combination, the determined search range is smaller, the search accuracy is higher, and the required information can be searched more quickly.
  • S302 Obtain and display a search result list based on the target search-type search training database, where the search result list includes at least one basic courseware information, and each basic courseware information includes a basic courseware ID, an associated basic presentation, and basic lecture voice data.
  • the basic courseware information is courseware information pre-stored in the training database.
  • the basic courseware ID is an identifier for uniquely identifying basic courseware information, and the identifier may be a storage location of the basic courseware information in the training database.
  • Each basic courseware ID is associated with the corresponding basic courseware information to quickly find the corresponding basic courseware information based on the basic courseware ID.
  • the basic presentation and the basic lecture voice data are correlated with each other, so that the basic lecture data is played on the basis of displaying the basic presentation, thereby improving the sense of substitution of the training process.
  • the fuzzy matching method may be adopted, and the training database is searched based on the target search formula to obtain and display the search result list, and the higher the matching degree of the basic courseware information, the greater the priority, the search result is The higher the ranking in the list, the trainers can quickly find the courseware information they want to train according to their needs.
  • S303 Acquire a search result selection request, where the search result selection request includes a target courseware ID.
  • the search result selection request refers to a selection request that selects one of the search result lists.
  • the target courseware ID refers to the basic courseware ID corresponding to the basic courseware information selected by the user, wherein the basic courseware information selected by the user is the target courseware information, and the target courseware ID is the identifier of the target courseware information.
  • the search result selection request may be a selection request entered through a recording device (such as a microphone) provided by the intelligent robot or a selection request input through a keyboard, or may be input by using a mouse dragging or clicking selection. Select the request to select the required target courseware ID.
  • the above input mode is simple, easy to implement and high in operation accuracy, which can meet the operating habits of different trainees, and is more conducive to improving the satisfaction of the trainees.
  • S304 Find the corresponding target courseware information based on the target courseware ID, and acquire the associated target presentation and target handout voice data.
  • the target courseware ID may point to the storage location of the target courseware information, and is associated with the information of the target courseware. Therefore, the target courseware ID may be linked to the storage location of the target courseware information to invoke the associated target presentation corresponding to the target courseware information.
  • Handouts and target speech data Specifically, the target handwritten voice data associated with the target presentation file may be automatically invoked when the target presentation is acquired; or the target presentation document associated with the target handout voice data may be automatically invoked when the target handout voice data is acquired.
  • the smart robot may generate the reminder information, so that the reminder information is sent to the mailbox corresponding to the preset management personnel to remind the management
  • the personnel supplement the basic courseware information such as the corresponding presentation and handout voice data, so as to supplement the corresponding basic courseware information, so that the training content in the training database is more perfect.
  • the training database in the intelligent robot can be updated in real time in the case of networking, so as to expand the knowledge surface to better meet the needs of the personnel to be trained.
  • the target search formula may be acquired based on the at least one search keyword, and the target search search search database is facilitated to obtain and display a search result list including at least one basic courseware information, so that the trainee can pass Check out this list of search results to learn more about the knowledge points you want to train and to personalize your choices.
  • the search result selection request can be input by using various input methods to obtain the target courseware information, thereby determining the personalized courseware information, the operation process is simple and convenient, easy to implement and high in accuracy.
  • the intelligent robot training method before the step of searching for the training database based on the at least one search keyword mentioned in step S30, the intelligent robot training method further comprises the following steps:
  • each basic courseware information includes a basic courseware ID, an associated basic presentation, and basic lecture voice data.
  • the training database is created in advance, so that the trainee can quickly select the corresponding interrelated target presentation and target handout voice data through the search request, so as to implement the target presentation and play the target handout at the same time.
  • the voice data makes the operation process simple and convenient, and it is easier to meet the individual needs of the personnel to be trained.
  • creating a training database includes the following steps:
  • S311 Acquire at least one basic presentation and at least one basic lecture voice data, and each basic presentation includes at least one presentation page.
  • the basic presentation refers to a pre-recorded training-related presentation collected and stored in the training database.
  • the basic lecture voice data refers to the training-related voice data pre-recorded and stored in the training database.
  • a demo page is a content page that makes up a presentation.
  • the basic presentation can include at least one presentation page for display via the onboard screen of the smart robot or by an external projector.
  • S312 Associate each presentation page with a basic lecture voice data, so that each of the presentation pages displays the associated basic lecture voice data when displayed.
  • each presentation page corresponds to basic lecture voice data for explaining the content of the presentation page.
  • the presentation page is associated with the basic lecture voice data explaining the content of the presentation page, and during the training process, the basic lecture voice data is played while being displayed to a certain presentation page.
  • explain the contents of the demo page which will help the trainees to better understand and understand the knowledge points of the demo page.
  • each presentation page is associated with a basic lecture voice data, so that the trainee can synchronously play the basic lecture instructions for explaining the presentation page when displaying a certain presentation page of the target presentation.
  • Voice data is configured to be associated with a basic lecture voice data, so as to be visually and audibly matched in real time, so as to be visually and audibly matched in real time, the more vivid display of the knowledge to be trained is more beneficial to the trainer to learn and understand.
  • step S312 specifically includes the following steps:
  • S3121 Enter the association configuration interface to obtain a document selection request, and the document selection request includes a document identifier.
  • the association configuration interface is an interaction interface displayed on the intelligent robot for the user to operate when setting the association relationship between the presentation page and the basic lecture voice data.
  • a document selection request is a request by the user to select a presentation that requires an associated configuration.
  • the document ID is the identifier used to determine the presentation you want to select, either the name of the presentation, the storage address of the presentation in the training database, or other identifiers that can be associated with the presentation.
  • S3122 Select a corresponding basic presentation according to the document identification, and display at least one presentation page corresponding to the basic presentation.
  • At least one presentation page of the basic presentation is displayed on the associated configuration interface, so that corresponding basic lecture voice data is configured for each presentation page.
  • S3123 Acquire an association selection request, where the association selection request includes a page identifier and a voice identifier.
  • the association selection request is a request for associating any presentation page in the basic presentation with its corresponding basic lecture voice data.
  • Each presentation page of the basic presentation is provided with a page identifier (which may be composed of a document identification and a page number) to determine a corresponding presentation page based on the page identification.
  • the voice identification is an identifier used to uniquely identify the basic handout voice data in the training database.
  • the intelligent robot may obtain the page identifier and the voice identifier of the associated selection request in association with each other, so that the corresponding presentation page corresponding to each page identifier is associated with the corresponding basic lecture voice data.
  • association selection request is input by using a drag and drop selection method or a check selection method, and the input mode is flexible, and it is more convenient to configure the relationship between the basic presentation and the basic lecture voice data.
  • S3124 Associate each presentation page with a basic lecture voice data according to the page identifier and the voice identifier, so that each basic presentation page displays the associated basic lecture voice data when displayed.
  • the basic lecture voice data is played according to the associated voice identifier.
  • each presentation page is associated with the basic lecture voice data for explaining the content thereof, so that when each presentation interface is displayed, the basic lecture voice data of the explanation explanation is synchronously played, and the transition presentation can be realized.
  • the corresponding basic lecture voice data is automatically switched in synchronization, so that the content of the demonstration during the training process is synchronized with the content of the explanation.
  • the intelligent robot training method further includes the following steps:
  • S411 Acquire a question request, the question request includes at least one question keyword.
  • the question request refers to a request sent by the user (ie, the person to be trained) to the intelligent robot to ask a question for a specific question during the training process.
  • the question request may be a voice question request input by the user using a recording device (such as a microphone), or may be a text question request input by the user through an input device such as a keyboard and/or a mouse, and the processing procedure is basically as follows in step S20. Consistent, I will not repeat them here.
  • S412 Search for a training database based on at least one question keyword, and obtain reply data corresponding to at least one question keyword, and the reply data includes reply voice data.
  • the reply data refers to the answer data that responds to the user's question for a specific question.
  • the reply voice data is voice data in the answer data that answers the user's question.
  • the process of searching for the training database based on the at least one question keyword is substantially the same as step S30, and details are not described herein.
  • the intelligent robot can raise a question about the knowledge point that the user does not understand, and the intelligent robot receives at least the question request from the user, according to at least the question request.
  • a question keyword search obtains the corresponding reply voice data
  • step S413 the target handout voice data needs to be resumed from the target handout voice data pause to ensure the continuity of the training process.
  • the intelligent robot extracts the question keyword by acquiring the user's question request, searches the training database to obtain the corresponding reply voice data, pauses the play of the target handwritten voice data, and plays the reply voice data to answer the user's question in real time.
  • the knowledge points that the user encounters during the training process so that the user can better understand the knowledge to be absorbed and enhance the interaction during the training process.
  • step S413 the playback of the target speech data is suspended based on the request, specifically including the following steps:
  • the program built in the intelligent robot determines whether the currently played statement is played according to the energy of the current play statement of the target speech data.
  • the target speech voice data will have a corresponding pause after each statement is played, and the silence is paused when the pause is performed, and the energy is lower than the energy played by the statement when the pause is not paused.
  • reply voice data After the reply voice data is obtained, it is detected that the currently played statement has been played, and the playback of the target voice data is suspended.
  • S4133 If it is detected that the currently played statement has not been played, continue playing until the statement is finished playing, and then perform the operation of pausing the playback of the target speech data.
  • the statement After obtaining the reply voice data, detecting that the currently played statement has not been played, the statement continues to be played until the energy of the current play statement is determined to be a pause, and then the target speech data is suspended for subsequent execution. Play the steps to reply to the voice data.
  • replying to the data may also include replying to the presentation.
  • the reply presentation is the presentation data in the answer data that answers the user's question.
  • step S413 the playing of the target speech data is suspended based on the question request in step S413, and the reply voice data is played, and the following steps are further included:
  • the reply presentation will overwrite the currently displayed target presentation, and the reply voice data will be played to answer the user's question in a more detailed and vivid manner.
  • the intelligent robot returns to the currently displayed target presentation, and continues to play the target speech data to continue the current training, making the training process very flexible and personalized to the user's knowledge of the training. Understand the degree and help users better grasp the knowledge of the training.
  • the real-time reply to the question request input by the trainer in the training speech mode can obtain the difficult question of the trainee in real time and provide the reply data in real time, and can timely solve the training personnel in the training speech mode.
  • Doubt explain the problems and knowledge points that are not understood in detail, enhance the interaction between intelligent robots and users, and make the training process more flexible.
  • the intelligent robot training method further includes the following steps:
  • S111 Acquire a mode selection request, and if the mode selection request is a practical drill mode request, enter a practical exercise interface.
  • the practice drill mode is a mode in which the intelligent robot interacts with the personnel to be trained, and the training personnel are subjected to the actual operation assessment by the intelligent robot.
  • the training mode has a strong sense of reality and is closely integrated with the actual scene. It is beneficial to improve the learning interest of the trainees, and the practical application effect is good, which is suitable for testing the knowledge mastery of the trainees.
  • S112 Acquire a drill mode selection request, and select a request search training database based on the drill mode, and obtain a test questionnaire, where the test questionnaire includes at least one test question.
  • test questionnaire refers specifically to the questionnaire that tests the knowledge mastery of the trainees according to the training-related content.
  • the test question specifically refers to the problem set by the training-related content for testing the knowledge points of the person to be trained.
  • the walkthrough mode may include a scenario test mode, a random test mode, and an exam test mode.
  • the scenario test mode is to practice the training of the person to be trained to select a certain preset scene, which can be adjusted according to the learning progress of the personnel to be trained, so as to enhance the memory.
  • the random test method is that the system randomly selects the scene for the trainees to test, which is beneficial to train the trainees' ability to adapt to the randomness.
  • the test test method is to collect a variety of test exercises according to the training content of the trainees to be trained, and the trainees are required to conduct test tests. This method is beneficial for the trainees to test their mastery of knowledge.
  • test reply information includes a test answer corresponding to the at least one question identifier.
  • the problem identifier is an identifier for uniquely identifying a corresponding test question, and the problem identifier can be associated with the test question and the corresponding standard answer.
  • the test reply information may be voice reply information input by the user using a recording device (such as a microphone), or text reply information input by the user through an input device such as a keyboard and/or a mouse, and the processing procedure is basically the same as step S20. Do not repeat them one by one.
  • the intelligent robot saves the voice reply information in the user's actual exercise practice process in real time, and converts the content of the voice reply information into test response information.
  • S114 Call the corresponding standard answer based on the question identifier, and obtain the test score of each test question according to the test answer and the standard answer.
  • test answers corresponding to the at least one question identifier in the test response information of the user are respectively matched with the corresponding standard answers, and the test reply information is scored according to the matching relevance, and each word is not required to be matched. You only need to fuzzy match the successful keyword to score. The higher the matching relevance, the higher the test score.
  • the user selects the actual exercise mode, and the intelligent robot broadcasts the test question according to the test questionnaire in the drill mode selected by the user, and each test question corresponds to a problem identifier. Then, after obtaining the test reply information input by the user based on the test question, the corresponding at least one standard answer is searched based on the question identifier, and each standard answer corresponds to a different level of the test score. The test answers corresponding to the at least one question identifier in the test reply information are respectively matched with the corresponding standard answers. Finally, according to the matching result, the test score corresponding to the test answer is judged, the test score is given to the test questionnaire, and the relevant learning suggestions are given based on the test score result.
  • the training database pre-configures the standard answers of all test questions.
  • One test question can be configured with multiple standard answers, which is not unique. As long as the test response information answered by the user during the actual exercise can match the keywords, the score can be scored.
  • the drill mode is selected as the scene test mode, and the scenario selected to promote the car insurance is set, and the test question is related to the simulation of the smart robot rejecting the sales call of the person to be trained. Scenes. The trainee will dial out the sales call according to the test question, and the intelligent robot simulates the rejection type reply: "I am busy now, no time.” At this time, if the respondent gives a reply of “OK, then talk again next time”, the intelligent robot first converts the voice reply information into test response information, and then includes the test reply information. The test answer matches the standard answer.
  • the standard answer preset in the training database is to ask the trainee to make an appointment for the customer and ask for the next free time, and the test answer of the trainee does not involve the key point of the standard answer, then the trainee's The test score is relatively low.
  • the reply to be sent by the trainer at this time is: "Okay, sorry, when will you be available, will I call you next time?"
  • the test reply message mentions the next time.
  • the trainee can obtain a relatively high test score.
  • there are many ways to ask for the next time. need to be configured in advance in the answers of the training database. As long as the relevant keywords can be fuzzyly matched, the corresponding test scores can be obtained.
  • different drill modes are set through the practice drill mode, which is convenient for the trainees to select a suitable training mode, fully meet the individual needs of the trainees, and independently search the test questionnaire according to the progress and content of the training, so that
  • the interactive mode of man-machine answering is conducive to improving the interest of the trainees, the practical application effect is good, and the test answers of the trainers are judged by the test scores, which is convenient for testing the trainees to be trained.
  • the degree of knowledge after the mastery is convenient for the trainees to select a suitable training mode, fully meet the individual needs of the trainees, and independently search the test questionnaire according to the progress and content of the training, so that
  • the interactive mode of man-machine answering is conducive to improving the interest of the trainees, the practical application effect is good
  • the test answers of the trainers are judged by the test scores, which is convenient for testing the trainees to be trained.
  • the degree of knowledge after the mastery is convenient for the trainees to select a suitable training mode, fully meet the individual needs of the trainee
  • FIG. 7 shows an intelligent robot training device corresponding to the intelligent robot training method provided in Embodiment 1.
  • FIG. 7 shows an intelligent robot training device corresponding to the intelligent robot training method provided in Embodiment 1.
  • the parts related to the embodiments of the present application are shown.
  • the intelligent robot training device includes a mode selection module 10, a search request acquisition module 20, a target search module 30, and a display broadcast module 40.
  • the mode selection module 10 is configured to obtain a mode selection request, and enter a training presentation interface if the mode selection request is a training speech mode request.
  • the search request obtaining module 20 is configured to acquire a search request, and acquire at least one search keyword based on the search request.
  • the target search module 30 is configured to search the training database based on the at least one search keyword to obtain the associated target presentation and the target handout voice data.
  • the display announcement module 40 is configured to display the target presentation and simultaneously play the target handout voice data associated with the target presentation.
  • the search request acquisition module 20 includes a voice search request acquisition unit 201, a voice data conversion unit 202, and a text data parsing unit 203.
  • the voice search request obtaining unit 201 is configured to acquire a voice search request, where the voice search request includes target voice search data.
  • the voice data conversion unit 202 is configured to perform conversion processing on the target voice search data by using a preset voice recognition system, and acquire text search data.
  • the character data parsing unit 203 is configured to perform parsing processing on the text search data by using the NLP library to acquire at least one search keyword.
  • the target search module 30 includes a logical processing unit 301, a courseware search unit 302, a search result selection unit 303, and a target courseware acquisition unit 304.
  • the logic processing unit 301 is configured to process each search keyword as a search field, and use a logical relationship to process at least one search field to form a target search formula.
  • the courseware search unit 302 is configured to search and display a training result database based on the target search formula, and the search result list includes at least one basic courseware information, where each basic courseware information includes a basic courseware ID, an associated basic presentation, and a foundation. Handout voice data.
  • the search result selection unit 303 is configured to obtain a search result selection request, and the search result selection request includes a target courseware ID.
  • the target courseware obtaining unit 304 is configured to find the corresponding target courseware information based on the target courseware ID, and acquire the associated target presentation and target handout voice data.
  • the intelligent robot training device further comprises: a training database creation module 31 for creating a training database.
  • the training database creation module 31 includes a data acquisition unit 311 and an association configuration unit 312.
  • the data obtaining unit 311 is configured to acquire at least one basic presentation and at least one basic lecture voice data, and each basic presentation includes at least one presentation page.
  • the association configuration unit 312 is configured to associate each presentation page with a basic lecture voice data to play the associated basic lecture voice data when each presentation page is displayed.
  • the intelligent robot training device further includes a question answering module 41.
  • the question answering module 41 includes a question request obtaining unit 411, a question search unit 412, and a reply data answering unit 413.
  • the question request obtaining unit 411 is configured to obtain a question request, and the question request includes at least one question keyword.
  • the question searching unit 412 is configured to search the training database based on the at least one question keyword to obtain reply data corresponding to the at least one question keyword, and the reply data includes replying the voice data.
  • the reply data response unit 413 is configured to suspend playback of the target speech data based on the question request and play back the reply voice data.
  • the intelligent robot training device further comprises a practice drill module 11.
  • the practice drill module 11 includes a mode selection unit 111, a drill mode selection unit 112, a test reply information acquisition unit 113, and a test score acquisition unit 114.
  • the mode selection unit 111 is configured to acquire a mode selection request, and enter a practical exercise interface if the mode selection request is a real operation drill mode request.
  • the exercise mode selection unit 112 is configured to acquire a drill mode selection request, and select a search training database based on the drill mode selection method to obtain a test questionnaire, and the test questionnaire includes at least one test question.
  • the test reply information obtaining unit 113 is configured to obtain test response information input by the user, and the test reply information includes a test answer corresponding to the at least one question identifier.
  • the test score obtaining unit 114 is configured to call a corresponding standard answer based on the question identifier, and obtain a test score for each test question according to the test answer and the standard answer.
  • This embodiment provides one or more non-volatile readable storage media having computer readable instructions stored thereon.
  • the one or more non-transitory readable storage mediums storing computer readable instructions, when executed by one or more processors, causing one or more processors to perform the intelligent robot training of Embodiment 1
  • the method in order to avoid repetition, will not be described here.
  • the functions of the units/modules in the intelligent robot training device in Embodiment 2 are implemented when the computer readable instructions are executed by one or more processors. To avoid repetition, details are not described herein again.
  • non-volatile readable storage media storing computer readable instructions may comprise: any entity or device capable of carrying the computer readable instructions, a recording medium, a USB flash drive, a mobile hard drive, a magnetic Disc, optical disc, computer memory, Read-Only Memory (ROM), Random Access Memory (RAM), electrical carrier signals, and telecommunications signals.
  • FIG. 8 is a schematic diagram of a computer device according to an embodiment of the present application.
  • computer device 80 of this embodiment includes a processor 81, a memory 82, and computer readable instructions 83 stored in memory 82 and executable on processor 81.
  • the processor 81 executes the steps of the intelligent robot training method in the first embodiment, such as the steps S10 to S40 shown in FIG. 1, when the computer readable instructions 83 are executed.
  • the processor 81 executes the computer readable instructions 83
  • the functions of the modules/units of the intelligent robot training device in the second embodiment are implemented, for example, the mode selection module 10, the search request acquisition module 20, and the target search module 30 shown in FIG. And displaying the function of the broadcast module 40.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Computational Linguistics (AREA)
  • Health & Medical Sciences (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • General Engineering & Computer Science (AREA)
  • General Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • General Physics & Mathematics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Human Computer Interaction (AREA)
  • Acoustics & Sound (AREA)
  • Multimedia (AREA)
  • Electrically Operated Instructional Devices (AREA)
  • Manipulator (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

本申请公开了一种智能机器人培训方法、装置、计算机设备及存储介质，方法包括：获取模式选择请求，若模式选择请求为培训演讲模式请求时，进入培训演讲界面，获取搜索请求，基于搜索请求获取至少一个搜索关键词，基于至少一个搜索关键词搜索培训数据库，获取相互关联的目标演示文稿和目标讲义语音数据，显示目标演示文稿，同时播放与目标演示文稿相互关联的目标讲义语音数据。本申请的技术方案通过采用智能机器人培训方法对待培训人员进行培训，达到了按照不同需求对待培训人员进行一一培训的效果，从而有效的满足了待培训人员的个性化需求。

Description

智能机器人培训方法、装置、计算机设备及存储介质
本申请以2018年3月12日提交的申请号为201810200486.1，名称为“智能机器人培训方法、装置、计算机设备及存储介质”的中国发明专利申请为基础，并要求其优先权。
技术领域
本申请涉及人工智能技术领域，尤其涉及一种智能机器人培训方法、装置、计算机设备及存储介质。
背景技术
随着社会经济的不断发展，人工智能在我们生活当中起着越来越重要的作用。人工智能(Artificial Intelligence，简称AI)，是指由人工制造出来的系统所表现出来的智能，人工智能是研究如何制造出智能设备或者智能系统，来模拟人类智能活动的一门技术。智能机器人是应用人工智能的设备，伴随着人工智能这门技术的快速发展，智能机器人也在快速发展中。
随着知识的不断更新，为了更好地适应于企业的发展，通常需要对职工进行专业培训，这种专业培训一般采用由培训讲师用演示文稿向多人讲授的方式来进行，而现在知识的更新速度快，必然决定了进行培训的频率越来越快。例如，传统专业培训(如客服培训)往往由培训讲师对待培训人员进行面对面培训，但这种培训方式存在着任务重、时间和地点不灵活、培训讲师资源匮乏和辅导培训的人员有限等问题。而且，传统专业培训主要由培训讲师依据其培训计划对待培训人员进行常规培训，无法满足每一待培训人员对其所需要了解的知识进行一一培训的要求，即满足不了待培训人员的个性化需求。
发明内容
本申请实施例提供一种智能机器人培训方法、装置、计算机设备及存储介质，以解决传统专业培训无法满足待培训人员的个性化需求的问题。
第一方面，本申请实施例提供一种智能机器人培训方法，包括：
获取模式选择请求，若所述模式选择请求为培训演讲模式请求时，进入培训演讲界面；
获取搜索请求，所述搜索请求包括至少一个搜索关键词；
基于至少一个所述搜索关键词搜索培训数据库，获取相互关联的目标演示文稿和目标讲义语音数据；
显示所述目标演示文稿，同时播放与所述目标演示文稿相互关联的所述目标讲义语音数据。
第二方面，本申请实施例提供一种智能机器人培训装置，包括：
模式选择模块，用于获取模式选择请求，若所述模式选择请求为培训演讲模式请求时，进入培训演讲界面；
搜索请求获取模块，用于获取搜索请求，基于所述搜索请求获取至少一个搜索关键词；
目标搜索模块，用于基于至少一个所述搜索关键词搜索培训数据库，获取相互关联的目标演示文稿和目标讲义语音数据；
显示播报模块，用于显示所述目标演示文稿，同时播放与所述目标演示文稿相互关联的所述目标讲义语音数据。
第三方面，本申请实施例提供一种计算机设备，包括存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机可读指令，所述处理器执行所述计算机可读指令时实现如下步骤：
获取模式选择请求，若所述模式选择请求为培训演讲模式请求时，进入培训演讲界面；
获取搜索请求，基于所述搜索请求获取至少一个搜索关键词；
基于至少一个所述搜索关键词搜索培训数据库，获取相互关联的目标演示文稿和目标讲义语音数据；
显示所述目标演示文稿，同时播放与所述目标演示文稿相互关联的所述目标讲义语音数据。
第四方面，本申请实施例提供一个或多个存储有计算机可读指令的非易失性可读存储介质，所述计算机可读指令被一个或多个处理器执行时，使得所述一个或多个处理器执行如下步骤：
获取模式选择请求，若所述模式选择请求为培训演讲模式请求时，进入培训演讲界面；
获取搜索请求，基于所述搜索请求获取至少一个搜索关键词；
基于至少一个所述搜索关键词搜索培训数据库，获取相互关联的目标演示文稿和目标讲义语音数据；
显示所述目标演示文稿，同时播放与所述目标演示文稿相互关联的所述目标讲义语音数据。
本申请的一个或多个实施例的细节在下面的附图及描述中提出。本申请的其他特征和优点将从说明书、附图以及权利要求书变得明显。
附图说明
为了更清楚地说明本申请实施例的技术方案，下面将对本申请实施例的描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本申请的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动性的前提下，还可以根据这些附图获得其他的附图。
图1是本申请实施例1提供的智能机器人培训方法的一实现流程图；
图2是图1中步骤S20的一实现流程图；
图3是图1中步骤S30的一实现流程图；
图4是本申请实施例1中提供的智能机器人培训方法的一实现流程图；
图5是本申请实施例1中提供的智能机器人培训方法的另一实现流程图；
图6是本申请实施例1中提供的智能机器人培训方法的又一实现流程图；
图7是本申请实施例2中提供的智能机器人培训装置的一示意图；
图8是本申请实施例4中提供的计算机设备的一示意图。
具体实施方式
下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例是本申请一部分实施例，而不是全部的实施例。基于本申请中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。
实施例1
请参阅图1，图1示出了本实施例提供的智能机器人培训方法的实现流程。该智能机器人培训方法应用在智能机器人中，用于与待培训人员进行人机交互，以便基于用户输入的相关请求对待培训人员进行个性化培训。本实施例中，以应用在保险金融行业的客服培训进行说明。详述如下：
一种智能机器人培训方法，包括由智能机器人执行的如下步骤：
S10：获取模式选择请求，若模式选择请求为培训演讲模式请求时，进入培训演讲界面。
其中，模式选择请求是指用于控制智能机器人进入相应模式的请求。
本实施例中，预先根据不同行业的业务性质及业务培训的需求设定不同的培训模式，如培训演讲模式、实操演练模式和课堂练习模式等。
具体地，培训演讲模式主要由智能机器人显示相关课件并通过演讲的方式对待培训人员进行培训讲解，此培训模式理论性强，知识点集中，针对性强，便于待培训人员快速接收到其需要了解的知识。
实操演练模式主要是由智能机器人给出模拟相关问题的场景，对待培训人员进行模拟练习，此培训模式情景感强，通过互动方式加深待培训人员的记忆。
课堂练习模式是由智能机器人根据课程进度显示相关习题，以对待培训人员进行考核或训练，巩固待培训人员接收到的知识。
需要说明的是，由于保险金融行业的客服培训主要涉及其业务的基本信息培训和座席沟通技巧培训，故在本实施例中智能机器人培训的模式配置示例性地为培训演讲模式和实操演练模式。
其中，培训演讲界面为智能机器人上设置的用于供用户操作的界面，也是培训演讲模式的初始程序界面。当智能机器人获取用户选择的培训演讲模式的模式选择请求时，进入到培训演讲界面，系统启动培训演讲模式的初始程序。
可以理解地，智能机器人通过预先根据不同行业的业务性质、业务培训或其他需求设置不同的培训模式，以使每一待培训人员可以根据自己的需求结合培训进度自主输入对应的模式选择请求，以使智能机器人进入不同的培训模式，以满足待培训人员个性化的需求。
进一步地，用户需预先通过客户端在对应的培训网页上进行注册，并在注册成功后，基于其对应的用户账号和密码登录智能机器人，智能机器人在用户每次登录成功后，基于预先录入的用户信息与用户交互。例如，在称谓上采用用户的名字及性别延伸的称呼，使人机交互更有代入感。并且，智能机器人可以从其对应的用于存储用户信息的数据库中调用每个用户账号的学习历史记录，并可基于最近一次的学习历史记录给用户进行信息推荐，具体可以推荐用户选择不同的培训模式，也可以推荐用户选择不同课件或其他与培训相关的信息，以使用户在培训学习时可以结合自身的学习进度进行高效学习。
S20：获取搜索请求，该搜索请求包括至少一个搜索关键词。
在本申请实施例中，搜索请求是用户对智能机器人发送的获取想要培训的知识点的请求。搜索请求可以为用户采用录音设备(如麦克风)输入的语音搜索请求或用户采用输入设备(如键盘)输入的文字搜索请求，可以根据用户的习惯自主选择搜素请求的输入方式。顾名思义，语音搜索请求是基于语音的搜索请求，文字搜索请求是基于文字的搜索请求。
具体地，搜素请求包括至少一个与培训内容相关的搜索关键词。其中，搜索关键词是用户根据想要获取的培训知识点提取的关键词。例如，用户想要获取推销车险业务的电话培训，则其提取的关键词可以为车险销售、车险业务等。可以理解地，该搜索关键词可以由用户自行提炼并输入到智能机器人以形成搜索请求，也可以由用户输入采用自然语言输入的搜索语句，通过智能机器人内置的关键词提取算法进行关键词提取，以获取对应的至少一个搜索关键词。
S30：基于至少一个搜索关键词搜索培训数据库，获取相互关联的目标演示文稿和目标讲义语音数据。
其中，培训数据库是用于存储培训资料的数据库，具体存储有预先录制收集的相互关联的讲义语音数据和演示文稿。该数据库可以是MySQL数据库、Oracle数据库或者其他数据库。本实施例中，可以基于至少一个搜索关键词和对应的搜索逻辑连接符(如and、not和or)形成搜索式，基于该搜索式对培训数据库进行搜索，以获取与至少一个搜索关键词对应的相互关联的目标演示文稿和目标讲义语音数据。
演示文稿包括PPT、视频和图片等可以通过智能机器人的显示屏幕进行演示的文稿。讲义语音数据不仅可以包括对PPT、视频和图片等进行讲解说明的音频文件，还可以包括对试卷、课本和习题等文档进行讲解的音频文件。
相互关联的目标演示文稿和目标讲义语音数据具体是指该目标演示文稿的每一演示页面均对应配置有对该演示页面的内容进行讲解的目标讲义语音数据。
可以理解地，培训数据库中预先存储相互关联的讲义语音数据和演示文稿，只需根据用户输入的搜索请求中的至少一个搜索关键词，采用简单的搜索查询操作，即可获取对应的相互关联的目标演示文稿和目标讲义语音数据，操作过程简单方便，而且，相互关联的讲义语音数据和演示文稿可以被不同的待培训人员重复搜索到，可实现一次录制多次使用的目的，使得培训过程可大幅减少对培训讲师的依赖，降低培训成本。
S40：显示目标演示文稿，同时播放与目标演示文稿相互关联的目标讲义语音数据。
具体地，智能机器人在获取相互关联的目标演示文稿和目标讲义语音数据之后，通过智能机器人内置的播放器或外置播放设备进行语音播放；同时，目标演示文稿通过智能机器人自带的屏幕进行显示或者通过投影仪投影到大屏幕上进行显示。这种在显示目标演示文稿的同时，播放相互关联的目标讲义语音数据，可使智能机器人对待培训人员进行培训的过程更具有代入感，使得待培训人员更容易接收所要培训的知识。
在一个具体实施方式中，通过智能机器人内置的关联程序，将演示文稿与讲义语音数据相互关联，根据该演示文稿演示某一页面的页面标识关联到与之相匹配的讲义语音数据，在播放演示文稿某一演示页面的同时，播放该演示页面相互关联的讲义语音数据。例如，若智能机器人当前显示为某一目标演示文稿的第N页，则根据关联程序，智能机器人同时播放与第N页演示页面相匹配的目标讲义语音数据。
本申请实施例所提供的智能机器人培训方法中，通过控制智能机器人自主的搜索培训数据库，根据待培训人员的需求选择所需的目标演示文稿和目标讲义语音数据，以贴合待培训人员的个性化需求。且由智能机器人显示目标演示文稿并同时播放目标讲义语音数据，以实现对待培训人员进行培训，使得培训过程更灵活，且减少培训讲师的培训压力，解决了培训讲师资源匮乏的问题。该智能机器人培训方法，可实现对不同需求的待培训人员进行一对一培训，可以更方便的开展培训，更适应于企业的发展。
在一具体实施方式中，如图2所示，步骤S20中，即获取搜索请求，搜索请求包括至少一个搜索关键词，具体包括如下步骤：
S201：获取语音搜索请求，语音搜索请求包括目标语音搜索数据。
顾名思义，语音搜索请求是基于语音的搜索请求，即在本实施方式中，智能机器人通过采集并识别用户的语音以获取用户的目标语音搜索数据。步骤S201具体采用如下步骤获取目标语音搜索数据：
首先，智能机器人采用麦克风阵列采集原始语音搜索数据。原始语音搜索数据是指智能机器人通过麦克风阵列采集到的但未经过处理的语音数据，该语音数据包含语音搜索请求的相关信息。其中，麦克风阵列采用分布式阵列，可以将子阵元或子阵列布局到更大的范围内，使得子阵元或子阵列相互之间通过有线或者无线的方式进行数据的交换和共享，并在此基础上进行广义上的声源定位和波束形成等技术实现信号处理。
然后，在麦克风阵列采集到原始语音搜索数据之后，智能机器人采用去混响算法对原始语音搜索数据进行进一步处理，以获取目标语音搜素数据。
进一步地，去混响算法采用的是基于波束形成的方法(Beamforming based approach)，通过对麦克风阵列收集的原始语音搜索数据的信号进行加权处理，在原始语音搜索数据的信号方向形成一个拾音波束，同时衰减来自其他方向的反射声。基于波束形成的方法，即通过向不同方向的声源分别形成拾音波束，并且抑制其他方向的声音，来对原始语音搜索数据进行提取或分离，以获取目标语音搜索数据。
目标语音搜索数据是指智能机器人通过麦克风阵列采集，采用去混响算法从原始语音搜索数据中提取到的语音数据，目标语音搜索数据包含了语音搜素请求的信息。可以理解地，通过麦克风阵列的分布式阵列，可以在更大范围内从多方位获取空间中的原始语音搜索数据，使得智能机器人获取用户语音搜索请求的范围更大，更利于接收语音搜索请求。采用去混响算法从原始语音搜索数据中提取目标语音搜索数据，可以更精准地从原始语音搜索数据中提取出目标语音搜索数据，获取更准确的语音搜素请求的信息。
S202：采用预设的语音识别系统对目标语音搜索数据进行转换处理，获取文字搜索数据。
其中，语音识别系统是采用内置的语音识别算法把语音转变成文字的系统。在本实施例中，智能机器人在接收到目标语音搜索数据之后，调用预设的语音识别系统对该目标语音搜索数据进行处理，以获取文字搜索数据。
S203：采用NLP库对文字搜索数据进行解析处理，获取至少一个搜索关键词。
其中，NLP(Natural Language Processing，自然语言处理)库是人工智能的一个自然语言处理数据库，是用于将自然语言解析匹配为计算机能识别的机器语言的数据库。该NLP库可以采用斯坦福自然语音处理团队开发的Stanford CoreNLP，可以快速对文字搜索数据进行解析处理，以获取至少一个搜索关键词。
基于步骤S201-S203，智能机器人通过内置的麦克风阵列将采集用户输入的语音搜索请求，并采用语音识别系统把语音搜索请求中的目标语音搜索数据转变成文字搜索数据，并将该文字搜索数据上传到NLP库进行语义解析，以便于方便快捷地获取至少一个搜索关键词。基于语音搜索请求获取对应的搜索关键词，可以使得用户选择语音输入方式输入搜索请求，操作过程简单方便，更容易满足用户的需求。
在另一具体实施方式中，搜索请求可以包括用户通过键盘和/或鼠标等输入设备输入的文字搜索数据。具体地，智能机器人获取用户通过键盘和/或鼠标等输入设备输入的文字搜索数据，再采用NLP库对该文字搜索数据进行解析，获取至少一个搜索关键词，其处 理过程与步骤S203相同，为避免赘述，不再一一描述。采用该键盘和/或鼠标等输入设备输入搜索请求的方式，可以精准的输入搜索请求，提高搜索准确度。
在一具体实施方式中，如图3所示，步骤S30中，即基于至少一个搜索关键词搜索培训数据库，获取相互关联的目标演示文稿和目标讲义语音数据，具体包括如下步骤：
S301：将每一搜索关键词作为搜索字段，采用逻辑关系对至少一个搜索字段进行处理，形成目标搜索式。
具体地，基于搜索请求提取出至少一个搜索关键词后，针对至少一个搜索关键词，采用对应的逻辑关系(与/或/非等)进行组合，作为目标搜索式。
其中，将多个搜索关键词采用逻辑关系组合进行搜索，其确定的搜索范围更小，搜索的精确度更高，可以更快捷地搜索到需要的信息。
S302：基于目标搜索式搜索培训数据库，获取并显示搜索结果列表，搜索结果列表包括至少一个基础课件信息，每一基础课件信息包括基础课件ID、相互关联的基础演示文稿和基础讲义语音数据。
其中，基础课件信息是预先存储在培训数据库中的课件信息。基础课件ID是用于唯一识别基础课件信息的标识，该标识可以是基础课件信息在培训数据库中的存储位置。每一基础课件ID与对应的基础课件信息关联，以便基于该基础课件ID快速查找到对应的基础课件信息。本实施例中，基础演示文稿和基础讲义语音数据相互关联，以便后续在显示基础演示文稿的基础上，播放基础讲义数据，提高其培训过程的代入感。
在本申请实施例中，可以采用模糊匹配的方法，基于目标搜索式搜索培训数据库，以获取并显示搜索结果列表，并且，匹配度越高的基础课件信息，优先级越大，其在搜索结果列表中的排序越靠前，使得待培训人员可以根据自己的需求快速找到自己最想培训的课件信息。
S303：获取搜索结果选择请求，搜索结果选择请求包括目标课件ID。
搜索结果选择请求是指从搜索结果列表中选择其中一个的选择请求。目标课件ID是指用户选择的基础课件信息对应的基础课件ID，其中，用户选择的基础课件信息为目标课件信息，目标课件ID即为该目标课件信息的标识。本实施例中，搜索结果选择请求可以是通过智能机器人自带的录音设备(如麦克风)录入的选择请求或者是通过键盘输入的选择请求，也可以是采用鼠标拖拽或点击选择的方式输入的选择请求，以选取所需要的目标 课件ID，以上输入方式操作简单、易于实现且操作准确度高，可以满足不同待培训人员的操作习惯，更有利于提高待培训人员的满意度。
S304：基于目标课件ID查找到对应的目标课件信息，获取相互关联的目标演示文稿和目标讲义语音数据。
其中，目标课件ID可以指向目标课件信息的存储位置，与目标课件的信息相关联，因此，可以根据目标课件ID链接至目标课件信息的存储位置，以调用目标课件信息对应的相互关联的目标演示文稿和目标讲义语音数据。具体地，可以在获取目标演示文稿时，自动调用与目标演示文稿相互关联的目标讲义语音数据；也可以在获取目标讲义语音数据时，自动调用与目标讲义语音数据相互关联的目标演示文稿。
进一步地，若搜索结果列表中没有用户(即待培训人员)所需的目标课件信息，该智能机器人可生成提醒信息，使得将该提醒信息发送给预先设置的管理人员对应的邮箱，以提醒管理人员补充相应的演示文稿和讲义语音数据等基础课件信息，以便后续进行补充相应的基础课件信息，使得培训数据库中的培训内容更完善。
进一步地，智能机器人中的培训数据库在联网的情况能够实时更新，以扩展知识面，便于更好地满足待培训人员的需求。
步骤S301-S304中，可基于至少一个搜索关键词获取目标搜索式，并利于该目标搜索式搜索培训数据库，以获取并显示包含至少一个基础课件信息的搜索结果列表，以使待培训人员可通过查阅该搜索结果列表，了解更多与自己想要培训的知识点相关的信息，并进行个性化选择。并且，在显示搜索结果列表之后，可以采用多种输入方式输入搜索结果选择请求，以获取目标课件信息，从而确定个性化选择的课件信息，操作过程简单方便，易于实现且准确度高。
在一具体实施方式中，在步骤S30提及的基于至少一个搜索关键词搜索培训数据库的步骤之前，智能机器人培训方法还包括如下步骤：
S31：创建培训数据库，培训数据库中存储至少一个基础课件信息，每一基础课件信息包括基础课件ID、相互关联的基础演示文稿和基础讲义语音数据。
本实施例中，预先创建培训数据库，以便于在待培训人员可以通过搜索请求快速选择对应的相互关联的目标演示文稿和目标讲义语音数据，以实现后续在演示该目标演示文稿并同时播放目标讲义语音数据，使得其操作过程简单方便，更容易满足待培训人员的个性化需求。
进一步地，如图4所示，创建培训数据库(即步骤S31)包括以下步骤：
S311：获取至少一个基础演示文稿和至少一个基础讲义语音数据，每一基础演示文稿包括至少一个演示页面。
其中，基础演示文稿是指预先录制收集并存储在培训数据库中的与培训相关的演示文稿，基础讲义语音数据是指预先录制收集并存储在培训数据库中的与培训相关的语音数据。演示页面是指构成演示文稿的内容页面。基础演示文稿可以包括至少一个演示页面，用于通过智能机器人自带的屏幕进行显示或者通过外置的投影仪进行显示。
S312：将每一演示页面与一基础讲义语音数据关联配置，以使每一演示页面显示时播放相互关联的基础讲义语音数据。
具体地，每一演示页面均对应有对其演示页面的内容进行讲解说明的基础讲义语音数据。通过智能机器人内置的关联程序，将演示页面与对其演示页面的内容进行讲解说明的基础讲义语音数据相关联，在培训过程中，显示到某一演示页面的同时，播放该基础讲义语音数据以对演示页面的内容进行讲解说明，利于待培训人员更好地学习和理解该演示页面的知识点。
在本实施例中，将每一演示页面与一基础讲义语音数据关联配置，以便于待培训人员在显示目标演示文稿的某一演示页面时，能同步播放对该演示页面进行讲解说明的基础讲义语音数据。步骤S311-S312中，通过使每一演示页面与一基础讲义语音数据关联配置，以便后续在视觉和听觉上实时配合，更生动形象的展示所要培训的知识，更利于待培训人员学习和理解。
进一步地，步骤S312中将每一演示页面与一基础讲义语音数据关联配置具体包括如下步骤：
S3121：进入关联配置界面，获取文稿选择请求，文稿选择请求包括文稿标识。
其中，关联配置界面为在设置演示页面与基础讲义语音数据的关联关系时，智能机器人上显示的供用户进行操作的交互界面。文稿选择请求是指用户选择所需要进行关联配置的演示文稿的请求。文稿标识是用于确定所要选择的演示文稿的标识，可以是演示文稿的名称，也可以是演示文稿在培训数据库中的存储地址或其他可以关联演示文稿的标识等。
S3122：根据文稿标识选择对应的基础演示文稿，显示基础演示文稿对应的至少一个演示页面。
具体地，根据文稿选择请求中的文稿标识选择出对应的基础演示文稿后，在关联配置界面上显示该基础演示文稿的至少一个演示页面，以便为每一演示页面配置对应的基础讲义语音数据。
S3123：获取关联选择请求，关联选择请求包括页面标识和语音标识。
其中，关联选择请求是用于将基础演示文稿中的任一演示页面与其对应的基础讲义语音数据关联起来的请求。基础演示文稿的每一演示页面均设置有页面标识(该页面标识可以由文稿标识和页码组成)，以便基于该页面标识确定对应的演示页面。语音标识是用于唯一识别培训数据库中的基础讲义语音数据的标识。智能机器人在接收到关联选择请求后，可获取关联选择请求中具有关联关系的页面标识和语音标识，以便后续其每一页面标识对应的演示页面和其对应的基础讲义语音数据关联过来。
进一步地，采用拖拽选择方式或勾选选择方式输入该关联选择请求，输入方式灵活，更方便配置相互关联的基础演示文稿和基础讲义语音数据的关关系。
S3124：根据页面标识和语音标识，将每一演示页面与一基础讲义语音数据关联配置，以使每一基础演示页面显示时播放相互关联的基础讲义语音数据。
具体地，通过将每一演示页面的页面标识与基础讲义语音数据的语音标识关联在一起，显示某一演示页面时，根据其关联的语音标识播放基础讲义语音数据。
基于步骤S3121-S3124，将每一演示页面与对其内容进行讲解说明的基础讲义语音数据相互关联，使得在显示每一演示界面时，同步播放讲解说明的基础讲义语音数据，可实现在切换演示界面时，同步自动切换相应的基础讲义语音数据，使得培训过程中演示的内容与讲解的内容同步。
在一具体实施方式中，在培训过程中，待培训人员可能对智能机器人培训过程讲解的知识点不理解，为了使得培训效果更好，需使得智能机器人的培训过程配置有实时解答用户(即待培训人员)提出问题的功能。因此，在步骤S40之后，即显示目标演示文稿，同时播放与目标演示文稿相互关联的目标讲义语音数据的步骤之后，如图5所示，该智能机器人培训方法还包括如下步骤：
S411：获取提问请求，该提问请求包括至少一个提问关键词。
其中，提问请求是指用户(即待培训人员)在培训过程中对智能机器人发送的想要针对特定的问题进行提问的请求。在本实施方式中，提问请求可以是用户采用录音设备(如 麦克风)输入的语音提问请求，也可以是用户通过键盘和/或鼠标等输入设备输入的文字提问请求，其处理过程如步骤S20基本一致，在此不一一赘述。
S412：基于至少一个提问关键词搜索培训数据库，获取与至少一个提问关键词相对应的回复数据，回复数据包括回复语音数据。
其中，回复数据是指对用户针对特定问题的提问进行答复的答案数据。回复语音数据是对用户的提问进行答复的答案数据中的语音数据。在本申请实施例中，基于至少一个提问关键词搜索培训数据库的过程与步骤S30基本一致，在此不一一赘述。
S413：基于提问请求暂停目标讲义语音数据的播放，并播放回复语音数据。
具体地，智能机器人在显示目标演示文稿和播放目标讲义语音数据的过程中，用户可以就自己不理解的知识点提出问题，智能机器人在收到用户提出的提问请求后，根据提问请求中的至少一个提问关键词搜索获取对应的回复语音数据后，需停止当前的目标讲义语音数据的讲解，并播放获取到的回复语音数据，以实现实时回复用户(即待培训人员)提出的问题的目的，提高培训过程的代入感。
可以理解地，在步骤S413中播放完回复语音数据之后，需从目标讲义语音数据暂停处继续播放目标讲义语音数据，以保证培训过程的连续性。
基于步骤S411-S413，智能机器人通过获取用户的提问请求提取提问关键词，以搜索培训数据库获取对应的回复语音数据，暂停目标讲义语音数据的播放并播放回复语音数据，以实时解答用户的疑问，对用户在培训过程中遇到不理解的知识点进行详细讲解，便于用户更好的理解吸收所要培训的知识，且增强了培训过程中的互动性。
进一步地，步骤S413中基于提问请求暂停目标讲义语音数据的播放，具体包括如下步骤：
S4131：检测目标讲义语音数据当前播放的语句是否播放完毕。
具体地，智能机器人内置的程序根据目标讲义语音数据当前播放语句的能量来判断当前播放的语句是否播放完毕。其中，目标讲义语音数据每一语句播放完毕后会有相应的停顿，停顿时为静音，其能量低于非停顿时语句播放的能量。
S4132：若检测到当前播放的语句已播放完毕，则暂停该目标讲义语音数据的播放。
在获取到回复语音数据之后，检测当前播放的语句已播放完毕，则暂停该目标语音数据的播放。
S4133：若检测到当前播放的语句未播放完毕，则继续播放至该语句播放完毕后，再执行暂停该目标讲义语音数据的播放的操作。
在获取到回复语音数据之后，检测当前播放的语句还未播放完毕，则继续播放完毕该语句，直至检测到当前播放语句的能量确定其为停顿时，再暂停该目标讲义语音数据，以便后续执行播放回复语音数据的步骤。
基于步骤S4131-S4133，通过智能机器人在播放回复语音数据之前，检测目标讲义语音数据当前播放的语句是否播放完毕，以保证在当前播放的语句播放完毕之后再进行回复语音数据的播放，从而保证了当前播放的语句的完整性，使得当前播放的语句能播放完毕，使得用户能够听取完整该语句，便于理解该语句表达的意思。
具体地，回复数据还可以包括回复演示文稿。其中，回复演示文稿是对用户的提问进行答复的答案数据中的演示文稿数据。
进一步地，步骤S413中的基于提问请求暂停目标讲义语音数据的播放，并播放回复语音数据，还包括如下步骤：
S4134：将回复演示文稿覆盖当前显示的目标演示文稿，并在回复语音数据播放完成后，返回当前显示的目标演示文稿，继续目标讲义语音数据的播放。
具体地，回复数据包括回复语音数据和回复演示文稿时，将回复演示文稿覆盖当前显示的目标演示文稿，并播放回复语音数据，更详细且生动直观地解答用户提出的问题。当用户提出的问题得到解答后，智能机器人返回当前显示的目标演示文稿，继续目标讲义语音数据的播放，以继续进行当前的培训，使得培训过程非常灵活，个性化地适应用户对培训知识点的理解程度，帮助用户更好地掌握所培训的知识。
在本实施方式中，通过在培训演讲模式下对待培训人员输入的提问请求进行实时回复，可实时获取待培训人员的疑难提问并实时给出回复数据，能及时解决培训演讲模式中待培训人员的疑惑，对不理解的问题和知识点进行详细重点讲解，增强了智能机器人与用户的交互性，使得培训过程更灵活。
在一具体实施方式中，如图6所示，在步骤S10提及的获取模式选择请求的步骤之后，该智能机器人培训方法还包括以下步骤：
S111：获取模式选择请求，若模式选择请求为实操演练模式请求时，进入实操演练界面。
在本申请实施例中，实操演练模式为智能机器人与待培训人员互动练习的模式，通过智能机器人出题对待培训人员进行实际操作考核，此培训模式真实感强，与实际现场结合紧密，有利于提高待培训人员的学习兴趣，实际应用效果好，适用于检验待培训人员的知识掌握程度。
S112：获取演练方式选择请求，并基于演练方式选择请求搜索培训数据库，获取测试问卷，测试问卷包括至少一个测试问题。
顾名思义，测试问卷具体是指根据培训相关的内容，测试待培训人员的知识掌握程度的问卷。测试问题具体是指根据培训相关的内容设置的用于测试待培训人员的知识点的问题。
在该具体实施方式中，演练方式可以包括场景测试方式、随机测试方式和考试测试方式。其中，场景测试方式为待培训人员自主选择某种预先设置的场景进行练习，此方式可根据待培训人员的学习进度进行调整，以利于加强记忆。随机测试方式为系统随机挑选场景，以供待培训人员进行测试，此方式利于训练待培训人员的随机应变能力。考试测试方式为按照待培训人员的培训内容记录集合多种测试习题，待培训人员以进行考试测试，此方式利于待培训人员检测自己对知识的掌握程度。
S113：获取用户输入的测试答复信息，测试答复信息包括与至少一个问题标识相对应的测试答案。
其中，问题标识是用于唯一识别对应的测试问题的标识，该问题标识可以关联测试问题和对应的标准答案。测试答复信息可以是是用户采用录音设备(如麦克风)输入的语音答复信息，也可以是用户通过键盘和/或鼠标等输入设备输入的文字答复信息，其处理过程如步骤S20基本一致，在此不一一赘述。
具体地，在人机问答对话方式中，智能机器人实时保存用户的实操演练练习过程中的语音答复信息，并将语音答复信息的内容转换为测试答复信息。
S114：基于问题标识调用对应的标准答案，根据测试答案和标准答案获取每一测试问题的测试分值。
进一步地，将用户的测试答复信息中与至少一个问题标识相对应的测试答案分别与其对应的标准答案进行匹配，根据匹配相关度对测试答复信息进行打分，匹配时不需要每个字都一样，只需要模糊匹配成功关键字就可以得分，匹配相关度越高其测试分值越高。
在本实施方式中，用户选择实操演练模式，智能机器人根据用户选择的演练方式中的测试问卷播报测试问题，每一测试问题对应一个问题标识。然后，在获取用户基于该测试问题输入的测试答复信息后，基于问题标识查找对应的至少一个标准答案，每一标准答案分别对应不同级别的测试分值。将该测试答复信息中与至少一个问题标识相对应的测试答案分别与其对应的标准答案进行匹配。最后，再根据匹配结果，判断该测试答案对应的测试分值，对本次测试问卷给出测试分值，基于测试分值的结果给出相关的学习建议。
其中，培训数据库预先配置好所有测试问题的标准答案，一个测试问题可以配置多种标准答案，不是唯一的，只要用户实操演练时回答的测试答复信息能够匹配其中的关键词，即可得分。
在一具体实施方式中，例如，在实操演练模式中，选择演练方式为场景测试方式，设定选择为推销车险的场景，测试问题为智能机器人针对待培训人员的推销电话予以拒绝的相关模拟场景。待培训人员根据测试问题拨出推销电话，智能机器人模拟拒绝型回复：“我现在比较忙，没空”。此时，若待培训人员若给出的回复为：“好的，那下次再聊”，则智能机器人先将该语音答复信息进行语言转换成测试答复信息，然后，将该测试答复信息包含的测试答案与标准答案进行匹配。具体地，培训数据库中预设的标准答案是要求待培训人员预约客户，询问下次空闲时间，而该待培训人员的测试答案并未涉及到该标准答案的关键点，则该待培训人员的测试分值比较低。又例如，若此时待培训人员给出的回复为：“好的，不好意思，请问您什么时候有空，我下次再给您来电？”，则该测试答复信息提到了询问下次时间的，根据模糊匹配关键字，则该待培训人员可以获得比较高的测试分值。本实施例中，询问下次时间方式很多，这些需要预先在培训数据库的答案中都配置好，只要相关的关键词能够模糊匹配就可以获得相应的测试分值。
在本实施方式中，通过实操演练模式设置不同演练方式，便于供待培训人员选择适合的培训模式，充分满足待培训人员的个性化需求，根据培训的进度和内容，自主搜索测试问卷，以便于巩固培训的知识，人机对答的互动方式，有利于提高待培训人员的学习兴趣，实际应用效果好，并且对待培训人员的测试答案进行测试分值的判定，便于检验待培训人员经过培训演讲后的知识掌握程度。
应理解，上述实施例中各步骤的序号的大小并不意味着执行顺序的先后，各过程的执行顺序应以其功能和内在逻辑确定，而不应对本申请实施例的实施过程构成任何限定。
实施例2
对应于实施例1中的智能机器人培训方法，图7示出了与实施例1提供的智能机器人培训方法一一对应的智能机器人培训装置。为了便于说明，仅示出了与本申请实施例相关的部分。
请参阅图7，该智能机器人培训装置包括：模式选择模块10、搜索请求获取模块20、目标搜索模块30和显示播报模块40。
模式选择模块10，用于获取模式选择请求，若模式选择请求为培训演讲模式请求时，进入培训演讲界面。
搜索请求获取模块20，用于获取搜索请求，基于搜索请求获取至少一个搜索关键词。
目标搜索模块30，用于基于至少一个搜索关键词搜索培训数据库，获取相互关联的目标演示文稿和目标讲义语音数据。
显示播报模块40，用于显示目标演示文稿，同时播放与目标演示文稿相互关联的目标讲义语音数据。
优选地，搜索请求获取模块20包括：语音搜素请求获取单元201、语音数据转换单元202和文字数据解析单元203。
语音搜素请求获取单元201，用于获取语音搜索请求，语音搜索请求包括目标语音搜索数据。
语音数据转换单元202，用于采用预设的语音识别系统对目标语音搜索数据进行转换处理，获取文字搜索数据。
文字数据解析单元203，用于采用NLP库对文字搜索数据进行解析处理，获取至少一个搜索关键词。
优选地，目标搜索模块30包括：逻辑处理单元301、课件搜索单元302、搜索结果选择单元303和目标课件获取单元304。
逻辑处理单元301，用于将每一搜索关键词作为搜索字段，采用逻辑关系对至少一个搜索字段进行处理，形成目标搜索式。
课件搜索单元302，用于基于目标搜索式搜索培训数据库，获取并显示搜索结果列表，搜索结果列表包括至少一个基础课件信息，每一基础课件信息包括基础课件ID、相互关联的基础演示文稿和基础讲义语音数据。
搜索结果选择单元303，用于获取搜索结果选择请求，搜索结果选择请求包括目标课件ID。
目标课件获取单元304，用于基于目标课件ID查找到对应的目标课件信息，获取相互关联的目标演示文稿和目标讲义语音数据。
优选地，该智能机器人培训装置还包括：培训数据库创建模块31，用于创建培训数据库。
培训数据库创建模块31包括：数据获取单元311和关联配置单元312。
数据获取单元311，用于获取至少一个基础演示文稿和至少一个基础讲义语音数据，每一基础演示文稿包括至少一个演示页面。
关联配置单元312，用于将每一演示页面与一基础讲义语音数据关联配置，以使每一演示页面显示时播放相互关联的基础讲义语音数据。
优选地，该智能机器人培训装置还包括提问解答模块41。
提问解答模块41包括提问请求获取单元411、提问搜索单元412和回复数据应答单元413。
提问请求获取单元411，用于获取提问请求，该提问请求包括至少一个提问关键词。
提问搜索单元412，用于基于至少一个提问关键词搜索培训数据库，获取与至少一个提问关键词相对应的回复数据，回复数据包括回复语音数据。
回复数据应答单元413，用于基于提问请求暂停目标讲义语音数据的播放，并播放回复语音数据。
优选地，该智能机器人培训装置还包括实操演练模块11。
实操演练模块11包括模式选择单元111、演练方式选择单元112、测试答复信息获取单元113和测试分值获取单元114。
模式选择单元111，用于获取模式选择请求，若模式选择请求为实操演练模式请求时，进入实操演练界面。
演练方式选择单元112，用于获取演练方式选择请求，并基于演练方式选择请求搜索培训数据库，获取测试问卷，测试问卷包括至少一个测试问题。
测试答复信息获取单元113，用于获取用户输入的测试答复信息，测试答复信息包括与至少一个问题标识相对应的测试答案。
测试分值获取单元114，用于基于问题标识调用对应的标准答案，根据测试答案和标准答案获取每一测试问题的测试分值。
本实施例提供的智能机器人培训装置中各模块实现各自功能的过程，具体可参考前述实施例的描述，此处不再赘述。
实施例3
本实施例提供一个或多个存储有计算机可读指令的非易失性可读存储介质。该一个或多个存储有计算机可读指令的非易失性可读存储介质，计算机可读指令被一个或多个处理器执行时，使得一个或多个处理器执行实施例1中智能机器人培训方法，为避免重复，这里不再赘述。或者，该计算机可读指令被一个或多个处理器执行时实现实施例2中智能机器人培训装置中各单元/模块的功能，为避免重复，这里不再赘述。
可以理解地，一个或多个存储有计算机可读指令的非易失性可读存储介质可以包括：能够携带所述计算机可读指令的任何实体或装置、记录介质、U盘、移动硬盘、磁碟、光盘、计算机存储器、只读存储器(Read-Only Memory，ROM)、随机存取存储器(Random Access Memory，RAM)、电载波信号和电信信号等。
实施例4
图8是本申请一实施例提供的计算机设备的示意图。如图8所示，该实施例的计算机设备80包括：处理器81、存储器82以及存储在存储器82中并可在处理器81上运行的计算机可读指令83。处理器81执行计算机可读指令83时实现上述实施例1中智能机器人培训方法的步骤，例如图1所示的步骤S10至步骤S40。或者，处理器81执行计算机可读指令83时实现上述实施例2中智能机器人培训装置的各模块/单元的功能，例如图7所示模式选择模块10、搜索请求获取模块20、目标搜索模块30和显示播报模块40的功能。
所属领域的技术人员可以清楚地了解到，为了描述的方便和简洁，仅以上述各功能单元、模块的划分进行举例说明，实际应用中，可以根据需要而将上述功能分配由不同的功能单元、模块完成，即将所述装置的内部结构划分成不同的功能单元或模块，以完成以上描述的全部或者部分功能。
以上所述实施例仅用以说明本申请的技术方案，而非对其限制；尽管参照前述实施例对本申请进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本申请各实施例技术方案的精神和范围，均应包含在本申请的保护范围之内。

Claims (20)

  1. 一种智能机器人培训方法，其特征在于，包括：
    获取模式选择请求，若所述模式选择请求为培训演讲模式请求时，进入培训演讲界面；
    获取搜索请求，基于所述搜索请求获取至少一个搜索关键词；
    基于至少一个所述搜索关键词搜索培训数据库，获取相互关联的目标演示文稿和目标讲义语音数据；
    显示所述目标演示文稿，同时播放与所述目标演示文稿相互关联的所述目标讲义语音数据。
  2. 如权利要求1所述的智能机器人培训方法，其特征在于，所述搜索请求包括语音搜索请求；
    所述获取搜索请求，基于所述搜索请求获取至少一个搜索关键词，包括：
    获取语音搜索请求，所述语音搜索请求包括目标语音搜索数据；
    采用预设的语音识别系统对所述目标语音搜索数据进行转换处理，获取文字搜索数据；
    采用NLP库对所述文字搜索数据进行解析处理，获取至少一个搜索关键词。
  3. 如权利要求1所述的智能机器人培训方法，其特征在于，所述基于至少一个所述搜索关键词搜索培训数据库，获取相互关联的目标演示文稿和目标讲义语音数据，包括：
    将每一所述搜索关键词作为搜索字段，采用逻辑关系对至少一个所述搜索字段进行处理，形成目标搜索式；
    基于所述目标搜索式搜索所述培训数据库，获取并显示搜索结果列表，所述搜索结果列表包括至少一个基础课件信息，每一所述基础课件信息包括基础课件ID、相互关联的基础演示文稿和基础讲义语音数据；
    获取搜索结果选择请求，所述搜索结果选择请求包括目标课件ID；
    基于所述目标课件ID查找到对应的目标课件信息，获取相互关联的所述目标演示文稿和所述目标讲义语音数据。
  4. 如权利要求1所述的智能机器人培训方法，其特征在于，在所述基于至少一个所述搜索关键词搜索培训数据库的步骤之前，智能机器人培训方法还包括：
    创建培训数据库，所述培训数据库中存储至少一个基础课件信息，每一所述基础课件信息包括基础课件ID、相互关联的基础演示文稿和基础讲义语音数据；
    所述创建培训数据库，包括：
    获取至少一个所述基础演示文稿和至少一个所述基础讲义语音数据，每一所述基础演示文稿包括至少一个演示页面；
    将每一所述演示页面与一所述基础讲义语音数据关联配置，以使每一所述演示页面显示时播放相互关联的所述基础讲义语音数据。
  5. 如权利要求1所述的智能机器人培训方法，其特征在于，在所述显示所述目标演示文稿，同时播放与所述目标演示文稿相互关联的所述目标讲义语音数据的步骤之后，所述智能机器人培训方法还包括：
    获取提问请求，所述提问请求包括至少一个提问关键词；
    基于至少一个所述提问关键词搜索所述培训数据库，获取与至少一个所述提问关键词相对应的回复数据，所述回复数据包括回复语音数据；
    基于所述提问请求暂停所述目标讲义语音数据的播放，并播放所述回复语音数据。
  6. 如权利要求5所述的智能机器人培训方法，其特征在于，所述回复数据还包括回复演示文稿；
    所述基于所述提问请求暂停所述目标讲义语音数据的播放，并播放所述回复语音数据，还包括：
    将所述回复演示文稿覆盖当前显示的所述目标演示文稿，并在所述回复语音数据播放完成后，返回当前显示的所述目标演示文稿，继续所述目标讲义语音数据的播放。
  7. 如权利要求1所述的智能机器人培训方法，其特征在于，在所述获取模式选择请求的步骤之后，所述智能机器人培训方法还包括：
    获取模式选择请求，若模式选择请求为实操演练模式请求时，进入实操演练界面；
    获取演练方式选择请求，并基于所述演练方式选择请求搜索所述培训数据库，获取测试问卷，所述测试问卷包括至少一个测试问题；
    获取用户输入的测试答复信息，所述测试答复信息包括与至少一个问题标识相对应的测试答案；
    基于所述问题标识调用对应的标准答案，根据所述测试答案和所述标准答案获取每一所述测试问题的测试分值。
  8. 一种智能机器人培训装置，其特征在于，包括：
    模式选择模块，用于获取模式选择请求，若所述模式选择请求为培训演讲模式请求时，进入培训演讲界面；
    搜索请求获取模块，用于获取搜索请求，基于所述搜索请求获取至少一个搜索关键词；
    目标搜索模块，用于基于至少一个所述搜索关键词搜索培训数据库，获取相互关联的目标演示文稿和目标讲义语音数据；
    显示播报模块，用于显示所述目标演示文稿，同时播放与所述目标演示文稿相互关联的所述目标讲义语音数据。
  9. 一种计算机设备，包括存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机可读指令，其特征在于，所述处理器执行所述计算机可读指令时实现如下步骤：
    获取模式选择请求，若所述模式选择请求为培训演讲模式请求时，进入培训演讲界面；
    获取搜索请求，基于所述搜索请求获取至少一个搜索关键词；
    基于至少一个所述搜索关键词搜索培训数据库，获取相互关联的目标演示文稿和目标讲义语音数据；
    显示所述目标演示文稿，同时播放与所述目标演示文稿相互关联的所述目标讲义语音数据。
  10. 如权利要求9所述的计算机设备，其特征在于，所述搜索请求包括语音搜索请求；
    所述获取搜索请求，基于所述搜索请求获取至少一个搜索关键词，包括：
    获取语音搜索请求，所述语音搜索请求包括目标语音搜索数据；
    采用预设的语音识别系统对所述目标语音搜索数据进行转换处理，获取文字搜索数据；
    采用NLP库对所述文字搜索数据进行解析处理，获取至少一个搜索关键词。
  11. 如权利要求9所述的计算机设备，其特征在于，所述基于至少一个所述搜索关键词搜索培训数据库，获取相互关联的目标演示文稿和目标讲义语音数据，包括：
    将每一所述搜索关键词作为搜索字段，采用逻辑关系对至少一个所述搜索字段进行处理，形成目标搜索式；
    基于所述目标搜索式搜索所述培训数据库，获取并显示搜索结果列表，所述搜索结果列表包括至少一个基础课件信息，每一所述基础课件信息包括基础课件ID、相互关联的基础演示文稿和基础讲义语音数据；
    获取搜索结果选择请求，所述搜索结果选择请求包括目标课件ID；
    基于所述目标课件ID查找到对应的目标课件信息，获取相互关联的所述目标演示文稿和所述目标讲义语音数据。
  12. 如权利要求9所述的计算机设备，其特征在于，在所述基于至少一个所述搜索关键词搜索培训数据库的步骤之前，所述处理器执行所述计算机可读指令时还实现如下步骤：
    创建培训数据库，所述培训数据库中存储至少一个基础课件信息，每一所述基础课件信息包括基础课件ID、相互关联的基础演示文稿和基础讲义语音数据；
    所述创建培训数据库，包括：
    获取至少一个所述基础演示文稿和至少一个所述基础讲义语音数据，每一所述基础演示文稿包括至少一个演示页面；
    将每一所述演示页面与一所述基础讲义语音数据关联配置，以使每一所述演示页面显示时播放相互关联的所述基础讲义语音数据。
  13. 如权利要求9所述的计算机设备，其特征在于，在所述显示所述目标演示文稿，同时播放与所述目标演示文稿相互关联的所述目标讲义语音数据的步骤之后，所述处理器执行所述计算机可读指令时还实现如下步骤：
    获取提问请求，所述提问请求包括至少一个提问关键词；
    基于至少一个所述提问关键词搜索所述培训数据库，获取与至少一个所述提问关键词相对应的回复数据，所述回复数据包括回复语音数据；
    基于所述提问请求暂停所述目标讲义语音数据的播放，并播放所述回复语音数据。
  14. 如权利要求9所述的计算机设备，其特征在于，在所述获取模式选择请求的步骤之后，所述处理器执行所述计算机可读指令时还实现如下步骤：
    获取模式选择请求，若模式选择请求为实操演练模式请求时，进入实操演练界面；
    获取演练方式选择请求，并基于所述演练方式选择请求搜索所述培训数据库，获取测试问卷，所述测试问卷包括至少一个测试问题；
    获取用户输入的测试答复信息，所述测试答复信息包括与至少一个问题标识相对应的测试答案；
    基于所述问题标识调用对应的标准答案，根据所述测试答案和所述标准答案获取每一所述测试问题的测试分值。
  15. 一个或多个存储有计算机可读指令的非易失性可读存储介质，其特征在于，所述计算机可读指令被一个或多个处理器执行时，使得所述一个或多个处理器执行如下步骤：
    获取模式选择请求，若所述模式选择请求为培训演讲模式请求时，进入培训演讲界面；
    获取搜索请求，基于所述搜索请求获取至少一个搜索关键词；
    基于至少一个所述搜索关键词搜索培训数据库，获取相互关联的目标演示文稿和目标讲义语音数据；
    显示所述目标演示文稿，同时播放与所述目标演示文稿相互关联的所述目标讲义语音数据。
  16. 如权利要求15所述的非易失性可读存储介质，其特征在于，所述搜索请求包括语音搜索请求；
    所述获取搜索请求，基于所述搜索请求获取至少一个搜索关键词，包括：
    获取语音搜索请求，所述语音搜索请求包括目标语音搜索数据；
    采用预设的语音识别系统对所述目标语音搜索数据进行转换处理，获取文字搜索数据；
    采用NLP库对所述文字搜索数据进行解析处理，获取至少一个搜索关键词。
  17. 如权利要求15所述的非易失性可读存储介质，其特征在于，所述基于至少一个所述搜索关键词搜索培训数据库，获取相互关联的目标演示文稿和目标讲义语音数据，包括：
    将每一所述搜索关键词作为搜索字段，采用逻辑关系对至少一个所述搜索字段进行处理，形成目标搜索式；
    基于所述目标搜索式搜索所述培训数据库，获取并显示搜索结果列表，所述搜索结果列表包括至少一个基础课件信息，每一所述基础课件信息包括基础课件ID、相互关联的基础演示文稿和基础讲义语音数据；
    获取搜索结果选择请求，所述搜索结果选择请求包括目标课件ID；
    基于所述目标课件ID查找到对应的目标课件信息，获取相互关联的所述目标演示文稿和所述目标讲义语音数据。
  18. 如权利要求15所述的非易失性可读存储介质，其特征在于，在所述基于至少一个所述搜索关键词搜索培训数据库的步骤之前，所述计算机可读指令被一个或多个处理器执行时，使得所述一个或多个处理器还执行如下步骤：
    创建培训数据库，所述培训数据库中存储至少一个基础课件信息，每一所述基础课件信息包括基础课件ID、相互关联的基础演示文稿和基础讲义语音数据；
    所述创建培训数据库，包括：
    获取至少一个所述基础演示文稿和至少一个所述基础讲义语音数据，每一所述基础演示文稿包括至少一个演示页面；
    将每一所述演示页面与一所述基础讲义语音数据关联配置，以使每一所述演示页面显示时播放相互关联的所述基础讲义语音数据。
  19. 如权利要求15所述的非易失性可读存储介质，其特征在于，在所述显示所述目标演示文稿，同时播放与所述目标演示文稿相互关联的所述目标讲义语音数据的步骤之后，所述计算机可读指令被一个或多个处理器执行时，使得所述一个或多个处理器还执行如下步骤：
    获取提问请求，所述提问请求包括至少一个提问关键词；
    基于至少一个所述提问关键词搜索所述培训数据库，获取与至少一个所述提问关键词相对应的回复数据，所述回复数据包括回复语音数据；
    基于所述提问请求暂停所述目标讲义语音数据的播放，并播放所述回复语音数据。
  20. 如权利要求15所述的非易失性可读存储介质，其特征在于，在所述获取模式选择请求的步骤之后，所述计算机可读指令被一个或多个处理器执行时，使得所述一个或多个处理器还执行如下步骤：
    获取模式选择请求，若模式选择请求为实操演练模式请求时，进入实操演练界面；
    获取演练方式选择请求，并基于所述演练方式选择请求搜索所述培训数据库，获取测试问卷，所述测试问卷包括至少一个测试问题；
    获取用户输入的测试答复信息，所述测试答复信息包括与至少一个问题标识相对应的测试答案；
    基于所述问题标识调用对应的标准答案，根据所述测试答案和所述标准答案获取每一所述测试问题的测试分值。
PCT/CN2018/081504 2018-03-12 2018-04-02 智能机器人培训方法、装置、计算机设备及存储介质 WO2019174072A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810200486.1A CN108536672A (zh) 2018-03-12 2018-03-12 智能机器人培训方法、装置、计算机设备及存储介质
CN201810200486.1 2018-03-12

Publications (1)

Publication Number Publication Date
WO2019174072A1 true WO2019174072A1 (zh) 2019-09-19

Family

ID=63483607

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/081504 WO2019174072A1 (zh) 2018-03-12 2018-04-02 智能机器人培训方法、装置、计算机设备及存储介质

Country Status (2)

Country Link
CN (1) CN108536672A (zh)
WO (1) WO2019174072A1 (zh)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112131743A (zh) * 2020-09-23 2020-12-25 中广核工程有限公司 核电厂设备的立体培训方法和系统
CN112767199A (zh) * 2020-12-25 2021-05-07 科讯嘉联信息技术有限公司 一种企业员工培训系统及培训方法
CN113630309A (zh) * 2021-08-12 2021-11-09 深圳追一科技有限公司 机器人会话系统、方法、装置、计算机设备和存储介质
CN114120740A (zh) * 2021-11-24 2022-03-01 科大国创云网科技有限公司 一种基于人机交互的客服行业业务人员培训方法及系统
CN115860995A (zh) * 2022-12-09 2023-03-28 广州沐思信息科技有限公司 基于云计算的智能培训监督方法及系统
CN116189681A (zh) * 2023-05-04 2023-05-30 北京水晶石数字科技股份有限公司 一种智能语音交互系统及方法

Families Citing this family (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109146432A (zh) * 2018-09-26 2019-01-04 北京城市网邻信息技术有限公司 针对于程序开发人员的面试方法、装置、设备及存储介质
CN109584880A (zh) * 2018-11-26 2019-04-05 山东弘享机器人有限公司 一种机器人演讲ppt文件的实现方法
CN109262617A (zh) * 2018-11-29 2019-01-25 北京猎户星空科技有限公司 机器人控制方法、装置、设备及存储介质
CN110209957A (zh) * 2019-06-06 2019-09-06 北京猎户星空科技有限公司 基于智能机器人的讲解方法、装置、设备及存储介质
CN110517688A (zh) * 2019-08-20 2019-11-29 合肥凌极西雅电子科技有限公司 一种语音关联提示系统
CN110556028A (zh) * 2019-08-23 2019-12-10 王英伟 基于人机交互的麻醉医师培训方法、设备及存储介质
CN110717020B (zh) * 2019-09-06 2022-09-27 平安国际智慧城市科技股份有限公司 语音问答方法、装置、计算机设备和存储介质
CN111553555A (zh) * 2020-03-27 2020-08-18 深圳追一科技有限公司 培训方法、装置、计算机设备和存储介质
CN111696010A (zh) * 2020-05-28 2020-09-22 深圳市元征科技股份有限公司 基于场景的培训方法、服务器、终端设备及存储介质
CN113177126A (zh) * 2021-03-24 2021-07-27 珠海金山办公软件有限公司 一种处理演示文稿的方法、装置、计算机存储介质及终端
CN113539267B (zh) * 2021-07-27 2023-11-07 广东电网有限责任公司 一种计量采集运维装置和方法
CN115249359B (zh) * 2021-09-16 2023-03-31 达闼科技（北京）有限公司 讲解方法、机器人、电子设备和存储介质
CN113920789A (zh) * 2021-10-21 2022-01-11 中石化宁波工程有限公司 工程类知识智能培训方法及系统
CN117033700A (zh) * 2023-10-09 2023-11-10 南京维赛客网络科技有限公司 基于ai助手辅助课件展示的方法、系统及存储介质
CN117493531B (zh) * 2023-12-29 2024-04-09 深圳星网信通科技股份有限公司 培训材料的生成方法、设备及存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103761261A (zh) * 2013-12-31 2014-04-30 北京紫冬锐意语音科技有限公司 一种基于语音识别的媒体搜索方法及装置
US20140122513A1 (en) * 2005-01-03 2014-05-01 Luc Julia System and method for enabling search and retrieval operations to be performed for data items and records using data obtained from associated voice files
CN104952009A (zh) * 2015-04-23 2015-09-30 阔地教育科技有限公司 资源管理方法、系统、资源管理服务器及互动教学终端
CN105895085A (zh) * 2016-03-30 2016-08-24 科大讯飞股份有限公司 一种多媒体转写方法和系统

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080134038A1 (en) * 2006-12-05 2008-06-05 Electronics And Telecommunications Research Interactive information providing service method and apparatus
CN202534180U (zh) * 2011-12-30 2012-11-14 安凯(广州)微电子技术有限公司 一种手持式自主学习机
CN106023693B (zh) * 2016-05-25 2018-09-04 北京九天翱翔科技有限公司 一种基于虚拟现实技术和模式识别技术的教育系统及方法
CN105868419A (zh) * 2016-06-02 2016-08-17 泉港区奇妙工业设计服务中心 一种童伴机器人搜索学习装置
CN106548773B (zh) * 2016-11-04 2020-06-23 百度在线网络技术（北京）有限公司 基于人工智能的儿童用户搜索方法及装置

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140122513A1 (en) * 2005-01-03 2014-05-01 Luc Julia System and method for enabling search and retrieval operations to be performed for data items and records using data obtained from associated voice files
CN103761261A (zh) * 2013-12-31 2014-04-30 北京紫冬锐意语音科技有限公司 一种基于语音识别的媒体搜索方法及装置
CN104952009A (zh) * 2015-04-23 2015-09-30 阔地教育科技有限公司 资源管理方法、系统、资源管理服务器及互动教学终端
CN105895085A (zh) * 2016-03-30 2016-08-24 科大讯飞股份有限公司 一种多媒体转写方法和系统

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112131743A (zh) * 2020-09-23 2020-12-25 中广核工程有限公司 核电厂设备的立体培训方法和系统
CN112131743B (zh) * 2020-09-23 2024-03-29 中广核工程有限公司 核电厂设备的立体培训方法和系统
CN112767199A (zh) * 2020-12-25 2021-05-07 科讯嘉联信息技术有限公司 一种企业员工培训系统及培训方法
CN113630309A (zh) * 2021-08-12 2021-11-09 深圳追一科技有限公司 机器人会话系统、方法、装置、计算机设备和存储介质
CN114120740A (zh) * 2021-11-24 2022-03-01 科大国创云网科技有限公司 一种基于人机交互的客服行业业务人员培训方法及系统
CN115860995A (zh) * 2022-12-09 2023-03-28 广州沐思信息科技有限公司 基于云计算的智能培训监督方法及系统
CN115860995B (zh) * 2022-12-09 2024-02-06 广州兴趣岛信息科技有限公司 基于云计算的智能培训监督方法及系统
CN116189681A (zh) * 2023-05-04 2023-05-30 北京水晶石数字科技股份有限公司 一种智能语音交互系统及方法
CN116189681B (zh) * 2023-05-04 2023-09-26 北京水晶石数字科技股份有限公司 一种智能语音交互系统及方法

Also Published As

Publication number Publication date
CN108536672A (zh) 2018-09-14

Similar Documents

Publication Publication Date Title
WO2019174072A1 (zh) 智能机器人培训方法、装置、计算机设备及存储介质
Lewis Serious use of a serious game for language learning
CN110489756B (zh) 会话式人机交互口语测评系统
US20160171387A1 (en) Digital companions for human users
CN106897950A (zh) 一种基于单词认知状态模型适应性学习系统及方法
EP1986175A2 (en) Method, interface and system for obtaining user input
KR20160077200A (ko) 언어-관련 장애들의 진단 및 치료를 위한 컴퓨팅 기술들
US20120329013A1 (en) Computer Language Translation and Learning Software
US20150037765A1 (en) System and method for interactive electronic learning and assessment
US20160358489A1 (en) Dynamic learning supplementation with intelligent delivery of appropriate content
CN110111611B (zh) 信息处理方法、装置、介质和电子设备
CN112053595B (zh) 由计算机实现的训练系统
CN110910694A (zh) 一种智能客服培训系统
CN109377802A (zh) 一种自动交互式智能教育系统及方法
CN110046290B (zh) 一种个性化自主教学课程系统
KR20070006742A (ko) 언어교육방법
Ockey et al. Evaluating technology-mediated second language oral communication assessment delivery models
CN101739852A (zh) 基于语音识别的实现自动口译训练的方法和装置
KR102272567B1 (ko) 음성 인식 수정 시스템
AU2018102139A4 (en) Advanced Learning System
KR101944628B1 (ko) 화상 학습 기반의 일대일 외국어 학습 시스템
US10453354B2 (en) Automatically generated flash cards
Duan et al. [Retracted] English Listening Prediction Strategy and Training Method with Data Mining
KR20160086152A (ko) 인터넷 상에서의 소리 기반 영어 훈련 방법 및 시스템
CN110428674A (zh) 一种听力学习装置的使用方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18909605

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 11.12.2020)

122 Ep: pct application non-entry in european phase

Ref document number: 18909605

Country of ref document: EP

Kind code of ref document: A1