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 carrying out a conversation system and / or chatbot system. As a result of the mirrored answer, proposed here for the first time, in the form of a mirroring and / or calibrating question by a chatbot, the user is motivated to remain in the conversation and offer up more information. This additional conversation with the information is then used again, by means of the detection via a processor and AI, to automatically store a classification of the customer, feedback about a product categorisation or a personality categorisation of a negotiation partner in connection with the user in a memory suitable therefor.

Description

[0001] Description

[0002] Method and system for interactive electronic communication

[0003] The invention relates to a method and a system for interactive electronic communication, in particular a method for implementing a conversation system and / or chatbot system.

[0004] A conversational system—hereafter also referred to as a "chatbot"—is a computer-based system capable of conducting human-like conversations with users. Chatbots are often used in social media, on websites, or in instant messaging applications to provide users with information, answer questions, or offer services.

[0005] Chatbots can be programmed in a variety of ways to conduct human-like conversations. Some chatbots use rule-based systems, which use predefined rules and patterns to respond to user input. Other chatbots use machine learning to conduct human-like conversations and improve based on user input.

[0006] Chatbots can also be used in a variety of industries, such as customer service, e-commerce, and the entertainment industry. They can help improve the efficiency of business processes and provide users with a faster and easier way to respond to their inquiries and needs.

[0007] There are different ways a chatbot can be trained, depending on how it was programmed.

[0008] A rule-based chatbot is typically programmed by a developer using predefined rules and patterns to respond to user input. These rules and patterns are built into the chatbot from the start.

[0009] This interactive electronic communication, which takes place through "chatbots", has existed for some time, but it is only through the rapid development of natural language processing that they have gained economic importance.

[0010] The growing market for customized technical solutions requires ever greater knowledge of customers, their needs, and, above all, their reactions to mass-produced and / or partially or fully customized products. Customized solutions offer a competitive advantage in all areas of life, regardless of the technology used, be it health, mobility, nutrition, information, housing, heating, clothing, education, leisure, sports, or anything else.

[0011] In order to achieve customer specificity, it is therefore sensible to collect as much information as possible about individual customers, for example customer reactions to new products, to classify it and to make it available to product developers as feedback for their work.

[0012] However, it is practically impossible to query customer data in a personalized manner because, firstly, it cannot be guaranteed that a human contact person will be available at the time a customer reacts to the new product and, secondly, because data collection and classification by a human user are less reliable than can be carried out by artificial intelligence.

[0013] Known chatbots are passive and unable to respond to users in an individualized manner. Therefore, the object of the present invention is to provide a method and system for interactive electronic communication that not only retrieves and forwards answers and information, but also continuously evolves.

[0014] This object is achieved by the subject matter of the present invention as disclosed in the description, the figures and the claims.

[0015] Accordingly, the subject of the present invention is a computer-implemented method comprising the following method steps:

[0016] - A) Receiving and performing continuous recognition of an input made by a user to generate one or more intermediate results in the form of user statements, which are stored by a system for interactive electronic communication and usable for processing by an artificial intelligence AI,

[0017] - B) computer-implemented and AI-implemented processing of the intermediate result, each based on the user's statement(s) stored as an intermediate result, and generating at least one result in the form of a reflecting and / or calibrating question in the context,

[0018] - C) Generation of one or more outputs of the system for interactive electronic communication to the user, each comprising the reflecting and / or calibrating question from step B), followed by

[0019] - D) Output of the result to the user by the interactive electronic communication system,

[0020] - E ) Capturing and storing the user’s response

[0021] - F) Classify and assign the user’s response and

[0022] - G) If necessary, repeat steps A) to F). Furthermore, the present invention relates to a system for interactive communication, which comprises the following modules: at least one recording unit which is connected to

[0023] - a processor unit which - using a microcontroller for performing data transformation and data processing,

[0024] - a storage unit containing computer program instructions executed by the processor unit,

[0025] - wherein the computer program instructions have a program code for generating one or more intermediate results in the form of statements based on the data transformation and

[0026] - wherein the computer program instructions have access to a processor with a program code for transforming the intermediate results into results in the form of questions, with which the intermediate results in the form of statements can be transformed into results in the form of questions by sentence rearrangement, interpretation, repetition of the last 2 to 7 words, generation of calibrating questions and / or adaptation of the pronouns.

[0027] Up to now, chatbots that work on a rule basis and have stored patterns and rules have given the same answers to the same questions from any user.

[0028] The "A. L. I. C. E. " ("Artificial Linguistic Internet Computer Entity" system) operates on the basis of a broad base of human users. If the chatbot, i.e. the system for interactive electronic communication, cannot find an answer to a user's question, it asks another user the question and forwards the answer to the first user. However, the system is not suitable for obtaining user-specific information or training the chatbot, because the chatbot does not learn anything from the forwarding.

[0029] A. L . I . C . E . is a natural language chatbot that

[0030] "XML-DTD" or "extensible markup language" - a "document type definition" called AIML (Artificial Intelligence Markup Language) uses to easily generate likely answers. ALICE was inspired by Joseph Weizenbaum's psychotherapy program ELIZA.

[0031] However, ALICE is not suitable for obtaining user-specific information and training the chatbot because the chatbot does not learn anything from the forwarding.

[0032] The general finding of the invention is that a chatbot with appropriately designed AI can firstly extract a statement from a user's input and secondly, using a simple algorithm, generate a reflecting and / or calibrating question from this statement, which automatically transmits a large number of so-called user-specific "soft skills" that otherwise could not be formulated automatically even with the greatest effort in modules for emotional recognition of the user. This is particularly because, by simply rearranging sentences and / or repetition and, if necessary, slight linguistic revision of the last words of a sentence, grammatically or at least colloquially correct reflecting questions can be formulated in most languages, but at least in German and English.This can be taught to a chatbot's AI, which then continuously generates and outputs these "under the skin" questions automatically, and based on the answers receives new training data, classifies it, and makes it usable as training data, thus continuously evolving in operation.

[0033] Further features and advantages of the invention will become apparent from the following explanations of an embodiment using schematic drawings.

[0034] BRIEF DESCRIPTION OF THE DRAWINGS

[0035] There show: FIG. 1 a scheme of the disclosed system, and

[0036] FIG. 2 is a flow diagram of an exemplary embodiment with speech recognition.

[0037] FIG. 1 shows a block diagram of an exemplary arrangement of the invention relating to an embodiment of the system for interactive electronic communication, comprising a chatbot 1, a processor with a memory unit 2, an artificial intelligence 3 and a further processor 4 connected to the IoT 6 and / or a cloud 5, as well as the user 7.

[0038] The respective arrows indicate a signal-transmitting connection between the respective modules; in the case of the connection between user 7 and chatbot 1, this is an acoustic connection, according to the embodiment shown here as an example, in which the input of the interactive electronic communication takes place at least partly via speech on the part of the user 7 and speech recognition on the part of the chatbot 1, and the output of the interactive electronic communication on the part of the chatbot 1 via a loudspeaker (not shown) and ears of the user 7 (not shown).

[0039] The signal transmission connection between the other modules 2 to 6 can be realized wirelessly and / or via cable.

[0040] The chatbot 1 is conventionally constructed, which means that it comprises at least a monitor, a microphone and / or a loudspeaker, a central circuit board on which most of the components of a computer, 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 the person skilled in the art in order to function.

[0041] The system for interactive electronic communication shown in Figure 1 shows how the user 7 communicates with the chatbot 1 and the chatbot 1 with the processor with storage unit 2, the processor 4 with access to the IoT 6 on the one hand, to an AI 3 on the other hand and finally to a cloud 5.

[0042] User 7 is represented here as a person, but it could also be an avatar and / or another chatbot.

[0043] FIG. 2 shows a flow diagram of an exemplary embodiment of the disclosed invention of a computer-implemented method for carrying out an interactive electronic conversation. In a first step 101, a user 100 makes an input, in the case shown a voice input, but within the meaning of the invention it can just as well be a sound, text, touch and / or image input. In the case of an image input, it can be provided, for example, that the system comprises a camera which is set up in such a way that it carries out face and / or gesture recognition and makes the corresponding data available to the system as input, so that an input, for example in the form of a voice input from a user 100, is supplemented by the face and / or gesture recognition of the user 100 for the system to make it easier to interpret.

[0044] In the following step 102, the voice input is received and, using a processor of the chatbot 1 from Figure 1 that is set up and suitable for this purpose, speech recognition is first carried out, particularly preferably based on NLP, and an intermediate result is stored therefrom, e.g. in the processor with memory unit 2 from Figure 1. In method step 103, this intermediate result is forwarded to an AI - for example the AI ​​3 from Figure 1 - which uses the intermediate result to identify a statement that is an essential component of the input. In method step 104, this statement is forwarded to a further AI to generate either a reflecting and / or a calibrating question. In method step 104, the AI ​​generates the question and loads it into an output device that is a loudspeaker, a monitor, a printer or the like.The AI ​​can also generate the question as speech but also as an image, text, writing, sound, etc. and forward it to the user via a suitable output device.

[0045] The output device then outputs the question to the user 100. The user answers in method step 105, with the answer of the user 100 being evaluated, received, and processed as input, which, as in method step 101 in the case shown here, is a voice input.

[0046] Alternatively, the response from user 100 can be received in method step 105, analyzed, and stored as an intermediate result in method step 106. It can be fed back into method step 102 to supplement the intermediate result achieved in method step 101 and / or stored as information in the system in method step 107. The storage as information can be categorized according to various categories, such as person-related, product-related, location-related, date-related, etc.

[0047] Regardless of the grammatical gender of a particular term, persons with male, female or other gender identity are included.

[0048] According to an advantageous embodiment of the method, the input is recognized by speech recognition. This is particularly true of natural language recognition, which, with the involvement of AI (artificial intelligence), analyzes and processes human language so that machines can understand human language. This is referred to as "natural language processing" (NLP). NLP combines computational linguistics, i.e., the rule-based modeling of human language, with statistical, machine, and deep learning models.

[0049] It has been recognized that these technologies together enable processors to process human language in the form of text and / or speech data and understand its full meaning, including the intent and / or emotion of the speaker or writer.

[0050] A key role is played by the generation of the intermediate result. In the intermediate result, for example, an intermediate result in text form is generated from the user's input, regardless of whether it is voice input, text input, sound input, touchscreen input and / or other input. The AI ​​identifies at least one statement in the text that was contained in the input and saves it as an intermediate result. The data of the intermediate result is then transferred to an AI, which generates a result in the form of a reflecting and / or calibrating question from the intermediate result.

[0051] According to an advantageous embodiment, the recognition is character recognition, for example by touchscreen and / or keyboard input, by scanning and / or handwriting.

[0052] The output from the interactive electronic communication system can also be speech and / or writing.

[0053] The method and system for interactive electronic communication proposed according to the invention enables chatbots to learn because a user's reactions are recorded and processed. This is achieved in particular by the fact that the recorded inputs are first summarized and saved as interim results before being transformed into questions and returned to the user. The resulting reactions and inputs are in turn recognized and processed; for example, they are classified, linked to the user or similar users and / or - again for example - passed on to the product developers as feedback. The chatbot then has different types of users who, for example, correspond to specific customer groups.In this way, the chatbot has a reservoir of questions and answers for this customer class, and when a new associated user communicates with the chatbot, it further develops the class of users or customers and itself.

[0054] The source of the data for training the chatbot is the user himself, because he answers the chatbot individually and thus provides information about himself and does not just receive stereotypical "best match" answers from the chatbot.

[0055] Chatbots that use machine learning to conduct human-like communication use machine learning. This is done by providing large amounts of data, known as "training." This data includes typical user inputs and the chatbot's corresponding responses. The machine learning model is then used to identify patterns in this data and learn how to respond to user input.

[0056] There are different 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 better suited to certain use cases than others.

[0057] It was recognized that for optimal benefit from the chatbot, its training as a process is preferably continuous and, in particular, that the chatbot is preferably continuously trained to improve its performance and adapt to new user inputs and / or needs.

[0058] Currently, chatbot administrators offer two main functionalities. One is the input of at least 20 versions of the same question to trigger what is known as NLU (Natural Language Understanding) training from intents, e.g., "Can I have some ice cream?", "Do you have any ice cream?", "I really want ice cream," etc., to train the intent "The user wants ice cream." The other is hard-coded answers that the chatbot can provide when such an intent is recognized, e.g., "Here you can find the nearest ice cream parlor."

[0059] Typically, this content is handed over to the data scientist as unstructured prose. It's also common for the content to be collected from multiple sources in various formats before this one-stop handover.

[0060] A chatbot administrator is someone responsible for managing and maintaining a chatbot. This may include programming, training, and updating the chatbot, depending on how the chatbot is configured and what type of functions it performs.

[0061] A chatbot administrator might also be responsible for monitoring the chatbot's performance and analyzing user inputs and responses to ensure the chatbot provides accurate and helpful answers. A chatbot administrator might also be responsible for integrating the chatbot with other systems and / or applications, ensuring it functions properly and meets the company's needs.

[0062] In some cases, a chatbot administrator might also be responsible for communicating with users and customers when the chatbot is unable to answer their queries or assist them. In this case, the chatbot administrator might interact directly with users to answer their queries or help them resolve their issues.

[0063] Overall, the chatbot administrator is responsible for managing and maintaining the chatbot, ensuring it functions properly and provides users with helpful information and services. A data scientist is a person 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 in data. They often work with large amounts of structured and unstructured data, using specialized tools and techniques to process and analyze the data.

[0064] The work of a data scientist often involves analyzing data to support business decisions or improve processes. They can also make predictions or forecasts by identifying patterns in the data and presenting the results in a visual format. Data scientists often work closely with other professionals, such as software developers and business analysts, using their knowledge of data analysis to solve complex problems and make decisions.

[0065] Data collection for chatbot systems is laborious and typically requires manual input by a programmer. Questions and corresponding answers are entered into a chatbot CMS (e.g., Botpress), and at least 10 alternatives are found for NLU training.

[0066] Overall, none of these methods and / or systems of known chatbot applications provide any indication of how a system for interactive electronic communication would have to be designed so that a chatbot can respond in a user-specific manner. This is changed by the present invention because a system and a method are disclosed in which a processor unit is suitable and configured for input recognition, in particular for speech recognition, and which is capable of transforming a user's statements into questions and is configured to ask these questions to a user via a suitable output device. For example, the following transformations take place in the processor and are then accessible to the user via an output device:

[0067] "System" here stands for "system for interactive electronic communication"

[0068] Identification of the user, i.e. input into the system:

[0069] - So, my friends and I tried out an escape room.

[0070] System output:

[0071] - Have you and your friends tried an escape room?

[0072] Identification of the user, input into the system:

[0073] - After breakfast we went to the sauna.

[0074] System output:

[0075] - Did you and your friends go to the sauna after breakfast?

[0076] Input into the system:

[0077] - The tool is difficult to grip with the left hand.

[0078] System output:

[0079] - Is it difficult to grip the tool with your left hand?

[0080] Input into the system:

[0081] - I have difficulty controlling the machine when it is near the window.

[0082] System output:

[0083] - Does the control cause problems when the machine is placed near a window?

[0084] Input into the system: We cannot make any changes.

[0085] System output:

[0086] - Can't you make any changes?

[0087] Through this type of simple transformation of a statement into a query, the chatbot can retrieve and process information from the user.

[0088] The processing goes through several steps, on the one hand it can be saved and asked for by a random generator to other users of the same machine and / or it can be stored as information about user X, i.e. saved.

[0089] This makes it possible to train a chatbot specifically for the user, not just the machine, and above all, it makes it possible to continuously develop the chatbot through operation with concrete findings that are tailored to the user.

[0090] In particular, the system is enabled to ask questions based on user input and generate a user profile from the answers. Then, a program code, a pattern, a set of rules, and / or machine learning can handle and answer the user's future questions in a user-specific manner.

[0091] According to an advantageous embodiment, the system is capable of asking further questions that are in the context of the statement but do not just repeat the exact words of the statement, so that further information about the user becomes available.

[0092] Input into the system:

[0093] - The tool is difficult to grip with the left hand. System output:

[0094] - How should the tool be gripped, better with the right hand or better with the left hand?

[0095] Input into the system:

[0096] - I have difficulty controlling the machine when it is near the window.

[0097] System output:

[0098] - Do the lighting conditions at the window affect the control of the machine?

[0099] This approach has several advantages: a more human-like conversation with the chatbot, more understanding responses, and the gathering of more information and insights about the user. A negotiation expert will want to quickly assess the personality type of the negotiating partner in order to determine how best to deal with them.

[0100] With the present invention, a chatbot can classify the personality type of a user and also of a negotiating partner and provide optimized communication for them. At the same time, the system can collect valuable KPIs (key performance indicators) of the user.

[0101] According to the invention, the chatbot is enabled to mirror the user's statements and thereby give the user a good feeling of being "understood". The system thus opens up a new dimension of communication between user and system, because the user feels comfortable during communication and, in turn, understands the question because it reflects and reflects his statement. Here, whether the user is aware of it or not, a harmony, an adaptation and a correspondence is created in the interactive electronic communication, which extends to the content of the communication and also to the form of communication, right down to the choice of words.

[0102] The means to achieve the goal are incredibly easy to achieve with the help of an AI; often it is enough to rearrange the last three words of an input to generate the question.

[0103] Example input:

[0104] "I have three apartments of my own"

[0105] Edition Variation 1 :

[0106] - simple mirroring question from the chatbot: "Do you have three apartments?"

[0107] Edition Variation 2 :

[0108] - Restriction to the last three words to formulate the mirroring question:

[0109] "Do you have apartments?" or - colloquially - "have three apartments of your own?"

[0110] Edition Variation 3 :

[0111] - Calibrating question: "Are you interested in real estate?"

[0112] If necessary, the chatbot can simply rearrange the last three, four, five, or seven words in virtually any way it sees fit, turning them into a question that invites the user to tell more about themselves in a very individual and personality-related way. Grammatical correctness or colloquial language can be selected and adjusted, depending on the chatbot's context.

[0113] The chatbot remains an automated system that classifies things based on keywords, which in turn enable new connections to other users, other products, other technologies, other teams, etc. For example, if a user says, "That's not good," the chatbot replies, "You sound pretty disappointed?", to which the user will reply. Or the user says, "I'll pick up the kids." The chatbot then says, "You sound like you're a very involved parent?", and the user will willingly elaborate.

[0114] Another possibility provided by the system for interactive electronic communication according to the present invention according to a preferred embodiment are calibrated and / or calibrating questions, such as what to do, for example

[0115] User's question: "Show me the nearest restaurant!" Chatbot answers "How can I do that?" then the user will probably answer: "Show me on the map" or "send me an email with a list"

[0116] In this way, the user has effectively solved the problem and specified the preferred output format. This allows the chatbot to provide an improved, personalized user experience.

[0117] By combining the present invention in 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. This chatbot continuously trains itself, generating data for further training and / or data for automated classification from the user's input, in particular from the user's statements, in particular through mirroring, i.e., by converting the input into a question exactly or analogously.

[0118] The mirroring of the user's input can be done, for example, by using the same words, in particular by rearranging the last three or five words according to very simple rules.

[0119] On the other hand, the chatbot can also just ask a question in a meaningful way or in context and then improve itself further through the reactions and answers.

[0120] Finally, the chatbot can also ask only calibrated or calibrating questions and generate output from them. A calibrated or calibrating question in the context of NLP refers to the calibration of perception to the conversation partner, or rather, the user. This occurs by recognizing their inner state by reading nonverbal signals, for example, in combination with gesture and / or facial recognition, etc.

[0121] The mirrored response, proposed here for the first time in the form of a mirroring and / or calibrating question from a chatbot, motivates the user to continue the conversation and provide more information about themselves. This further conversation with the information is then used again through recognition via processor and AI to automatically classify the customer, provide feedback on a product or personality classification of a negotiating partner in connection with the user, and store it in a suitable memory.

Claims

Patent claims 1. Computer-implemented method for interactive electronic communication, comprising the following process steps: - A) receiving and performing (101) continuous recognition (102) of an input made by a user (7,100) to generate one or more intermediate results in the form of statements of the user (7,100) that are stored by a system for interactive electronic communication and usable for processing by an artificial intelligence AI (3), - B) computer-implemented and AI-implemented processing (103) of the intermediate result, each on the basis of the user's (7, 100) findings stored as an intermediate result, and generating at least one result in the form of a reflecting and / or calibrating question in the context, - C) generating (105) one or more outputs of the system for interactive electronic communication to the user (7,100), each comprising the reflecting and / or calibrating question from step B), subsequently - D) output (104) of the result by the interactive electronic communication system to the user, - E) Capturing and storing (106, 102) the user's response - F) Classifying and assigning (107) the user’s response and - G) If necessary, repeat steps A) to F).

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

3. Method according to one of claims 1 or 2, in which an avatar is used as the user. 4 . Method according to one of claims 1 to 3, in which a chatbot is used as the user.

5. Method according to one of the preceding claims, in which the continuous recognition of the input takes place in the form of speech recognition on the basis of natural language processing (NLP).

6. Method according to one of the preceding claims, in which the intermediate result is stored in text form. 7 . Method according to one of the preceding claims, in which the generation of the mirroring question takes place in the form of a sentence rearrangement of the input converted into text. 8 . Method according to one of the preceding claims, in which the generation of the mirroring question takes place in the form of a repetition of the last 2 to 7 words of the input.

9. System for interactive communication, comprising the following modules: at least one recording unit ( 1 ) connected to - a processor unit ( 1 ) which - using an AI - is configured to carry out data transformation and data processing, - a memory unit with computer program instructions (2) which are executed by the processor unit, - wherein the computer program instructions have a 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 have access to a processor ( 4 ) with a program code for transforming the intermediate results into results in the form of questions, with which the intermediate results in the form of statements can be generated by sentence rearrangement, interpretation, repetition of the last 2 to 7 words, generation calibrating questions and / or adapting the pronouns can be transformed into results in the form of questions.

10. System according to claim 9, wherein the data transformation is carried out by means of a processor that is configured and arranged to carry out natural language processing NLP.

11. System according to one of claims 9 or 10, comprising a processor which is suitable, arranged and configured so that the system generates new training data for the AI ​​of the system from the user's answers.

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

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

14. A computer program product comprising program code means which cause a processor and / or a neural network for executing an artificial intelligence (AI) to carry out a method according to any one of claims 1 to 8 when the program code means are processed by the processor and / or the neural network.

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

Images