Method and system for interactive electronic communication

Abstract

The invention relates to a method and a system for interactive electronic communication, in particular a method for executing a session system and / or a chat robot system. Through the mirror-image-type answer which is proposed for the first time and is given by the chatting robot in a mirror-image-type and / or calibration-type question form, the user is stimulated to continue dialogue and reveal more information about the user. This further conversation with information is then reused by means of the processor and the identification of the AI to automatically store feedback of classification of the customer, classification of the product or classification of the character of the negotiation partner in the context related to the user in a memory suitable for this.

Description

Technical Field

[0001] The present invention relates to a method and system for interactive electronic communication, in particular to a method for implementing a conversational system and / or a chatbot system. Summary of the Invention

[0002] A conversational system—hereinafter also referred to as a "chatbot"—is a computer-assisted system capable of engaging in human-like conversations with users. Chatbots are often used on social media, websites, or instant messaging applications to provide users with information, answer questions, or offer services.

[0003] Chatbots can be programmed in different ways to enable human-like conversations. Some chatbots use rule-based systems, where predefined rules and patterns are used to respond to user input. Other chatbots use machine learning to conduct human-like conversations and improve themselves based on user input.

[0004] Chatbots can also be applied in numerous industries, such as customer service, e-commerce, and entertainment. Chatbots can help improve the efficiency of business processes and provide users with a faster and simpler possibility to respond to their requests and needs.

[0005] There are various ways to train a chatbot, depending on how it is programmed.

[0006] Rule-based chatbots are typically programmed by developers to respond to user input using predefined rules and patterns that are built into the chatbot from the outset.

[0007] Although this type of interactive electronic communication through "chatbots" has existed for quite some time, it only gained economic importance due to the rapid development of natural language processing.

[0008] The growing market for customer-specific technical solutions requires an increasing understanding of customers, their needs, and, in particular, their reactions to mass-produced and / or partially or completely individually manufactured products. Customer-specific solutions offer competitive advantages in all areas of life, regardless of the specific technology they are used for, whether in health, mobility, nutrition, information, housing, heating, clothing, education, leisure, sports, or other areas.

[0009] Therefore, in order to establish customer specificity, it makes sense to collect as much information as possible about individual customers (e.g., how customers react to new products), categorize this information, and provide it as feedback to product developers for their work.

[0010] However, personalized querying of customer data is practically impossible because, firstly, there is no guarantee that a human conversation partner will be available at the moment a customer reacts to a new product, and secondly, the reliability of data collection and classification by human users is lower than when performed by artificial intelligence.

[0011] Known chatbots are passive and unable to respond to users in a personalized manner. Summary of the Invention

[0012] SUMMARY OF THE INVENTION It is therefore an object of the present invention to provide a method and a system for interactive electronic communication which not only enables the retrieval and forwarding of answers and information, but which also allows for continuous development.

[0013] This object is achieved by the subject matter of the invention, as disclosed in the description, the drawings and the claims.

[0014] The subject of the present invention is therefore a computer-implemented method comprising the following steps: - A) receiving and performing continuous recognition of inputs made by a user to generate one or more intermediate results in the form of user statements, said statements being stored by the system for interactive electronic communication and available for processing by artificial intelligence (AI), - B) subjecting the intermediate results to computer-implemented and AI-implemented processing, respectively, based on one or more user statements stored as intermediate results, and generating in this context at least one result in the form of a mirroring and / or calibration question, - C) generating one or more outputs from the system for interactive electronic communication to the user, each output comprising the mirrored and / or calibrated questions of step B), followed by - D) outputting said result to the user by the system for interactive electronic communication, - E) collect and store the user's answers, - F) categorize and attribute user responses, and - G) Repeat steps A) to F) if necessary.

[0015] Furthermore, the subject matter of the invention is a system for interactive communication, comprising the following modules: - at least one receiving unit connected to - a processor unit configured to perform data conversion and data processing, where AI is used, - a memory unit having computer program instructions executed by the processor unit, - wherein the computer program instructions have program code for generating one or more intermediate results in a statement form based on the data transformation, and - wherein the computer program instructions can access a processor having a program code for converting an intermediate result into a result in the form of a question, wherein the program code can convert an intermediate result in the form of a statement into a result in the form of a question by sentence conversion, paraphrase, repeating the last 2 to 7 words, generating a calibrated question and / or adapting pronouns.

[0016] Until now, chatbots that were rule-based and stored patterns and rules would give the same answer to the same question from any arbitrary user.

[0017] The "ALICE" (Artificial Linguistic Internet Computer Entity) system operates based on a broad user base. If a chatbot (a system for interactive electronic communication) cannot find an answer to a user's question, the chatbot will ask another user the same question and forward the other user's answer to the first user. However, this system is not suitable for acquiring user-specific information and training the chatbot, as the chatbot learns nothing from this forwarding.

[0018] ALICE is a natural language chatbot that uses an XML-DTD or Extensible Markup Language-Document Type Definition called AIML (Artificial Intelligence Markup Language) to easily generate likely responses. ALICE was inspired by Joseph Weizenbaum's psychotherapy program ELIZA.

[0019] However, ALICE is not suitable for obtaining user-specific information and training chatbots because the chatbot cannot learn anything from such forwarding.

[0020] The general insight of the present invention is that a chatbot with a correspondingly implemented AI can first extract statements from user input and, secondly, generate mirror and / or calibration questions from these statements using simple algorithms. These questions automatically convey a large number of so-called user-specific "soft skills" that cannot be automatically developed even with the most sophisticated modules for user emotion recognition. This is particularly true because grammatically or at least colloquially correct mirror questions can be formed in most languages ​​(at least in German and English) through simple sentence transformations and / or repetitions, and possibly slight linguistic modifications to the last few words of the sentence. This can be taught to the chatbot's AI, which can then continuously and automatically generate these "touching" questions, output them, receive new training data based on the responses, classify these questions, and use them as training data, thereby continuously developing them during operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Further features and advantages of the present invention will become apparent from the following explanation of exemplary embodiments based on the schematic drawings.

[0022] Figure 1 A schematic diagram showing the disclosed system, and Figure 2 A flow chart illustrating an exemplary implementation with speech recognition is shown. DETAILED DESCRIPTION

[0023] Figure 1 A block diagram showing an exemplary arrangement of an embodiment of the present invention regarding a system for interactive electronic communication, comprising a chatbot 1, a processor with a storage unit 2, an artificial intelligence 3 and a further processor 4 connected to an IoT 6 and / or a cloud 5, and a user 7.

[0024] Each arrow represents a signal transmission connection between corresponding modules. In the case of a connection between user 7 and chatbot 1, according to the exemplary embodiment shown, the connection is an acoustic connection. In this embodiment, the input of interactive electronic communications is at least partially carried out via the voice of user 7 and speech recognition on the chatbot 1 side, while the output of interactive electronic communications is carried out on the chatbot 1 side via a speaker (not shown) and the ear (not shown) of user 7.

[0025] The signal transmission connections between the other modules 2 to 6 may be realized wirelessly and / or via cables.

[0026] The chatbot 1 is conventionally constructed, i.e. it comprises at least a display, a microphone and / or a loudspeaker, a central motherboard on which most of the components of a calculator, computer or server are connected to one another, a CPU (“Central Processing Unit”) and a working memory (e.g. RAM, “Random Access Memory”), as well as other modules known to a person skilled in the art that enable it to function properly.

[0027] Figure 1 The system for interactive electronic communication shown in shows how a user 7 communicates with a chatbot 1 which in turn communicates with a processor 4 with a storage unit 2, access to an IoT 6 on the one hand, an AI 3 on the other hand and finally a cloud 5.

[0028] The user 7 is shown here as a person, but may also be an avatar and / or another chatbot.

[0029] Figure 2 A flow chart illustrating an exemplary embodiment of a computer-implemented method for conducting an interactive electronic conversation according to the disclosed invention is shown. In a first step 101, user 100 performs an input, which in the illustrated case is voice input, but within the meaning of the present invention, could equally well be sound input, text input, touch input, and / or image input. In the case of image input, it can be provided, for example, that the system includes a camera configured to perform facial recognition and / or gesture recognition and provide corresponding data as input to the system, so that input, for example, in the form of user 100's voice input, can be supplemented by facial recognition and / or gesture recognition of user 100 for easier interpretation by the system.

[0030] In the next step 102, the voice input is received and Figure 1 A suitable processor of the chatbot 1 in the embodiment of the invention first performs speech recognition (particularly preferably based on NLP) and stores intermediate results, for example in Figure 1 In a processor with a memory unit 2. In method step 103, the intermediate result is forwarded to the AI ​​- for example Figure 1 AI 3 in the example identifies statements that form the basic components of the input from the intermediate results. In method step 104, these statements are forwarded to another AI for generating mirrored and / or calibrated questions. In method step 104, the AI ​​generates the questions and loads them into an output device, such as a speaker, monitor, printer, or the like. The AI ​​can also generate the questions as speech, but can also generate them as images, text, words, sounds, or the like, and forward them to the user via a suitable output device.

[0031] The output device then outputs the question to user 100. The user answers in method step 105, wherein user answer 100 is evaluated, received, and processed as input (which, in the case shown here, is a voice input as in method step 101).

[0032] Alternatively, the answers of user 100 can be received and analyzed in method step 105, stored as intermediate results in method step 106, and fed back to method step 102 to supplement the intermediate results obtained in method step 101, and / or stored as information in the system in method step 107. The information can be stored according to different categories, such as by person, product, location, date, etc.

[0033] Individuals with male, female, or other gender identities are included regardless of the grammatical gender of a particular term.

[0034] According to one advantageous embodiment of the method, the input recognition is speech recognition. This is particularly natural language recognition, which uses AI (artificial intelligence) to analyze and process human language, enabling machines to understand it. This is known as "natural language processing" (NLP). NLP combines computational linguistics (i.e., rule-based modeling of human language) with statistical, machine learning, and deep learning models.

[0035] It is recognized that these technologies collectively enable processors to process human language in the form of text and / or voice data and understand its full meaning, including the intent and / or emotion of the speaker or author.

[0036] The generation of intermediate results plays a key role here. These intermediate results are generated, for example, from user input—whether voice, text, sound, touchscreen, or other input. The AI ​​identifies at least one statement contained in the input in the text and stores it as an intermediate result. The data from the intermediate results is then transmitted to the AI, which generates results from the intermediate results in the form of mirror and / or calibration questions.

[0037] According to an advantageous embodiment, the recognition is text recognition, for example via a touch screen and / or keyboard input, by scanning and / or handwriting.

[0038] Output via the system for interactive electronic communication can also be via voice and / or text.

[0039] The method and system proposed in accordance with the present invention enable chatbots to learn by collecting and processing user responses. This is achieved, in particular, by first aggregating and storing the collected input as an intermediate result before converting it into a question and returning it to the user. The responses and input generated are then recognized and processed again, for example, categorized, associated with the user or similar users, and / or (again, exemplarily) passed on as feedback to product developers.

[0040] The chatbot then has different types of users, which correspond, for example, to specific customer groups. The chatbot thus has a library of questions and answers for these customers, and when new, corresponding users communicate with the chatbot, it further develops these users or customers and itself.

[0041] In operation, the data source used to train the chatbot is the users themselves, as users answer the chatbot in a completely personalized way, providing information about themselves rather than just receiving a stereotyped "best match" answer from the bot.

[0042] Chatbots that use machine learning to conduct human-like communication use machine learning. To do this, they are fed a large amount of data, known as "training data." This data contains typical user input and the chatbot's associated responses. A machine learning model is then used to identify patterns in this data and learn how to respond to user input.

[0043] There are various types of machine learning models that can be used in chatbots, such as neural networks and decision trees. Each model has its own strengths and weaknesses and may be more suitable for specific application scenarios than others.

[0044] It has been recognized that in order to obtain optimal utilization from the chatbot, the training of the chatbot is preferably an ongoing process, and it is particularly preferred that the chatbot is continuously further trained to improve its performance and adapt to new user input and / or user needs.

[0045] Currently, chatbot administrators offer two main features. One is to input at least 20 versions of the same question to trigger so-called NLU (Natural Language Understanding) training of intents, such as "Can I have ice cream?", "Do you have ice cream?", "I really want ice cream," and so on, to train the intent "The user wants ice cream." The other is to hard-code answers that the chatbot can provide when such an intent is recognized, such as "Here is the nearest ice cream shop."

[0046] Often, this content is handed over to the data scientist as unstructured prose. Equally common is that the content has already been collected from different sources in different formats before this one-stop handover moment.

[0047] A chatbot administrator is someone who manages and maintains a chatbot. This can include programming, training, and updating the chatbot, depending on how the chatbot is configured and what types of functions it performs.

[0048] A chatbot administrator may also be responsible for monitoring the chatbot's performance and analyzing user input and user responses to ensure the chatbot provides correct and helpful responses. A chatbot administrator may also be responsible for integrating the chatbot into other systems and / or applications and ensuring it works as specified and meets the organization's requirements.

[0049] In some cases, chatbot administrators may also be responsible for communicating with users and customers if the chatbot is unable to answer users' questions or help them. In such cases, the chatbot administrator may interact directly with users to answer their questions or help them resolve their issues.

[0050] In summary, a chatbot administrator is responsible for managing and maintaining the chatbot and is responsible for ensuring that it works as specified and provides useful information and services to users.

[0051] A data scientist is someone who analyzes and understands large amounts of data. Data scientists use mathematics, statistics, machine learning, and other tools and techniques to identify and understand patterns and relationships within data. They often work with large amounts of structured and unstructured data and use specialized tools and techniques to process and analyze it.

[0052] Data scientists often analyze data to support business decisions or improve processes. They may also create predictions or forecasts by identifying patterns in data and presenting the results in an intuitive form. Data scientists often work closely with other professionals, such as software developers and business analysts, and leverage their knowledge of data analysis to solve complex problems and make decisions.

[0053] Collecting data for a chatbot system is tedious and usually requires manual input by programmers. Questions and corresponding answers are entered into a chatbot CMS (such as Botpress) and at least 10 alternatives are found for NLU training.

[0054] In summary, none of these known chatbot-based methods and / or systems clearly demonstrate how a system for interactive electronic communication must be designed to enable the chatbot to provide user-specific responses. This is addressed by the present invention, which discloses a system and method in which a processor unit, adapted and configured for input recognition (particularly speech recognition), is capable of converting user statements into questions and is configured to present these questions to the user via a suitable output device. For example, the following conversion occurs in the processor and is then provided to the user via an output device: Here, "system" means "a system for interactive electronic communications" User statements, i.e. input to the system: - Well, my friends and I tried an escape room.

[0055] The output of the system: - Have you and your friends tried an escape room? User statements, i.e. input to the system: - After breakfast we went to the sauna.

[0056] The output of the system: - Did you and your friends go to the sauna after breakfast? Input to the system: - The tool is difficult to hold with the left hand.

[0057] The output of the system: - Is the tool difficult to hold with your left hand? Input to the system: - When the machine is next to the window, the controls are difficult for me.

[0058] The output of the system: - Is it difficult to control the machine when it is close to the window? Input to the system: - We can't make any changes.

[0059] The output of the system: - Can't you make any changes? By simply converting a statement into a rhetorical question, the chatbot can call out information to the user and process said information.

[0060] The process here goes through several steps. On the one hand, the process can be stored and asked to other users using the same machine based on the random generator; and / or the process can be stored as information about user X.

[0061] This makes it possible to train the chatbot in a user-specific, rather than just a machine-specific, manner and, in particular, to use specific, user-specific statements to allow the chatbot to continuously evolve through operation.

[0062] In particular, the system is enabled to ask questions based on user input and generate a user profile based on the responses. Program code, patterns, rule sets, and / or machine learning can then provide user-specific processing and responses to future user questions.

[0063] According to an advantageous embodiment, the system is adapted to ask further questions that are relevant to the statement context, but not just the exact repetition of the statement words, so that further information about the user can be obtained.

[0064] Input to the system: - The tool is difficult to hold with the left hand.

[0065] The output of the system: - How to hold the tool, right or left hand? Input to the system: - When the machine is next to the window, the controls give me difficulty.

[0066] The output of the system: - Does the light condition near the window affect the control of the machine? This approach has a number of advantages: conversations with chatbots feel more human and responses are easier to understand, but it also allows for more information and insights about the user. Negotiation experts would like to quickly categorize their partner’s personality type to figure out how to best approach them.

[0067] With the help of the present invention, the chatbot can classify the personality type of the user and the personality type of the negotiation partner and provide them with optimized communication methods. At the same time, the system can also collect valuable KPIs (key performance indicators) of the user in the process.

[0068] According to the present invention, the chatbot is enabled to mirror the user's statements, thereby conveying a positive feeling of "understanding" to the user. This system thus opens up a new dimension of understanding between the user and the system, as the user feels comfortable during communication and understands the question, which is a paraphrase and mirror of their own statement. Here, whether the user is aware of it or not, harmony, adaptation, and consistency are created in the interactive electronic communication, extending to the content of the communication, as well as the form and even the wording.

[0069] The means to achieve this goal can be incredibly simple with AI; often simply transforming the last three words of the input is enough to create a problem.

[0070] Input example: - "I have three apartments of my own" Output variant 1: - Simple mirroring question for chatbot: “Do you have three apartments? ” Output variant 2: - Construct mirror questions by limiting yourself to the last three words: “Do you have an apartment? ” or - colloquially - “have three apartments of their own ” Output variant 3: - Calibrated question: "Are you interested in real estate? ” If necessary, the chatbot can simply rearrange the last three, four, five, or seven ("a few") words almost arbitrarily and create a question from it, through which the user is encouraged to tell more about themselves in a very personalized and personality-specific way. Depending on the chatbot's environment, grammatical correctness or colloquialisms can be selected and set.

[0071] Here, the chatbot is still an automated system that classifies based on keywords, which in turn can create new associations with other users, other products, other technologies, other teams, etc.

[0072] For example, when a user says, “That’s not good,” the chatbot responds, “You sound quite disappointed. ”, the user replies again. Or the user says, “I’m going to pick up the kids.” The chatbot says, “You sound like a very dedicated parent. ”, which the user will happily elaborate on further.

[0073] Another possibility of providing a system for interactive electronic communication according to a preferred embodiment of the present invention is calibrated and / or calibrated questions, such as questions about how to: User's question: "Show me the nearest restaurants ” The chatbot responded: “How can I ” The user might then respond: “Show me on a map” or “Email me with the list.”

[0074] In this way, the user actually solves the problem and pre-defines their preferred output format, allowing the chatbot to provide an improved, personalized user experience.

[0075] By integrating the present invention into a system for interactive electronic communication—at least one input device, a processor for recognition, a storage module, an AI for processing, and an output device—a completely new chatbot can be created. The chatbot continuously trains itself, wherein the chatbot generates data for further training and / or data for automatic classification from user input, in particular from user statements, in particular by mirroring (i.e., by converting them exactly or roughly into a question).

[0076] Mirroring of the user input can be done, for example, using the same words, in particular by reordering the last three or five words according to very simple rules.

[0077] On the other hand, a chatbot can also ask questions based solely on the general idea or context and further improve itself again through reactions (i.e. answers).

[0078] Finally, chatbots can also ask only calibrated or calibrated questions and generate output from them. In the NLP framework, calibrated or calibrated questions are understood as calibrating perception to the conversation partner (i.e., the user). This calibration is achieved by reading non-verbal signals, such as by combining gesture recognition and / or facial recognition, to identify the user's internal state.

[0079] The mirrored responses presented here for the first time by the chatbot in the form of mirrored and / or calibrated questions encourage the user to continue the conversation and reveal more about themselves. This further conversation, which contains information, is then used again with the help of a processor and AI recognition to automatically store feedback on customer classification, product classification, or personality classification of the negotiation partner in the context of the user in a suitable memory.

Claims

1. A computer-implemented method for interactive electronic communication, comprising the following method steps: - A) receiving and performing (101) continuous recognition (102) of inputs made by a user (7, 100) to generate one or more intermediate results in the form of statements of the user (7, 100), said statements being stored by the system for interactive electronic communication and being available for processing by artificial intelligence AI (3), - B) subjecting the intermediate results to computer-implemented and AI-implemented processing (103) based on one or more statements of the user (7, 100) stored as intermediate results, and generating in this context at least one result in the form of a mirroring and / or calibration question, - C) generating (105) one or more outputs of the system for interactive electronic communication to the user (7, 100), each output comprising the mirrored and / or calibrated questions of step B), followed by - D) outputting (104) said result to said user by said system for interactive electronic communication, - E) collecting and storing (106, 102) the user's answer, - F) categorizing and attributing user responses (107), and - G) Repeat steps A) to F) if necessary.

2. The method according to claim 1, wherein the input by the user (7, 100) is performed in the form of a voice input.

3. The method according to any one of claims 1 or 2, wherein a virtual avatar is used as the user.

4. The method according to any one of claims 1 to 3, wherein a chatbot is used as the user. The method according to claim 1 , wherein the continuous recognition of the input takes place in the form of speech recognition based on natural language processing (NLP). The method according to claim 1 , wherein the intermediate results are stored in text form.

7. The method according to any one of the preceding claims, wherein the generation of the mirror question is performed in the form of sentence transformation of the input converted into text.

8. The method according to any one of the preceding claims, wherein the generation of the mirror question is performed in the form of repeating the last 2 to 7 words of the input.

9. A system for interactive communication, comprising the following modules: - at least one receiving unit (1) connected to - a processor unit (1) configured to perform data conversion and data processing when using AI, - a storage unit (2) with computer program instructions, said stored instructions being executed by said processor unit, - wherein the computer program instructions have program code for generating one or more intermediate results in the form of statements based on the data transformation, and - wherein the computer program instructions are able to access a processor (4), which has a program code for converting an intermediate result into a result in the form of a question, using which the intermediate result in the form of a statement can be converted into a result in the form of a question by sentence conversion, paraphrase, repeating the last 2 to 7 words, generating a calibrated question and / or adapting pronouns.

10. The system of claim 9, wherein the data conversion is performed by a processor configured and set up to perform natural language processing (NLP).

11. A system according to any one of claims 9 or 10, having a processor adapted, set up and configured to enable the system to generate new training data for the system's AI from the user's answers.

12. The system according to any one of claims 9 to 11, wherein an avatar and / or a second chatbot is provided, which acts as a user in the system.

13. A system according to any one of claims 9 to 12, comprising a camera for facial recognition and / or gesture recognition.

14. A computer program product comprising program code means which, when processed by a processor and / or a neural network for executing artificial intelligence AI, causes the processor and / or the neural network to execute the method according to any one of claims 1 to 8.

15. A computer-readable storage medium having a computer program product according to claim 14.

Images