Method for automated and context-sensitive data processing

Abstract

The present invention relates to a method for capturing, processing and presenting data using artificial intelligence (AI), comprising the following steps: a) determining context information relevant to a user, the context information being generated on the basis of user data, user inputs, stored information and / or environmental conditions; b) analyzing the context information by means of an algorithm with the use of AI, in order to identify relevant topics, ideas and / or prompts which are suitable for prompting the user to input further data; c) generating one or more personalized suggestions on the basis of the analysis of the contextual information, the suggestions being specifically matched to the determined topics or interests of the user; d) automatically providing the generated suggestions to the user in the form of a spoken, textual, haptic and / or pictorial output; e) capturing inputs of the user in response to the generated output.

Description

RELEFANT Status: 3.1E25 RLE200001EPWO Methods for automated and context-sensitive data processing

[0001] The present invention relates to a method for automated and context-sensitive data processing, in particular for user metrics-based adaptation of a suggestion pipeline for efficient data processing.

[0002] Traditional data processing systems have the disadvantage that large amounts of information are not processed efficiently and no or only slightly relevant results are presented to the user without the user having to interact extensively and independently with and guide the system.

[0003] These classic systems are often based on fixed, rule-based algorithms that operate independently of the individual user context. This lack of flexibility means they lack the ability to respond to specific user interests or situations, often resulting in an overload of irrelevant data or missed, potentially important information. Insufficient consideration of the user's context can negatively impact the user experience and drastically reduce the practical value of such systems.

[0004] In order to take user-relevant context into account, newer methods and systems are therefore increasingly using artificial intelligence (AI) to improve the relevance and efficiency of data processing.

[0005] AI-supported algorithms can recognize patterns and relationships within large datasets, thus better adapting the results to the user's needs.

[0006] Although AI systems offer a more advanced form of data preparation, they are often complex to implement and require large amounts of training data to function reliably. Furthermore, their decision-making can be opaque, which poses challenges in data-intensive applications where it is unclear what actions are being taken and why.

[0007] One particular challenge is addressing people in a structured way and interacting with them using either very little or a very large amount of data stored in a system. This is especially true when... RELEFANT Status: 3.1E25 RLE200001EPWO The structured collection of information from a user and their encouragement to provide further information is important.

[0008] There are existing methods and devices that, for example, support a user in automatically capturing and documenting spoken text and then pre-structuring it with the help of artificial intelligence and processing it in a textual or multimedia format, for example by integrating photos or videos.

[0009] However, this is a passive use of artificial intelligence, as the user is not actively encouraged by the system to interact with it and enter information. If a user doesn't know what to enter, the systems known from the state of the art fail right from the start, as no relevant information can be provided to the system. These systems therefore act more like a classic assistant that writes down the entered information in a structured way, but nothing more.

[0010] Zulfikar W. Deen et al., “Memoro,” describe a context-sensitive AI dialogue system that continuously captures conversation data and suggests context-appropriate memory aids to the user. Furthermore, US 2023 / 245651 A1 discloses a framework for context-relevant interaction by evaluating environmental data and recognized objects. Both state-of-the-art documents demonstrate context-based suggestion generation; however, they do not explicitly teach user-metric-sensitive control of the suggestion pipeline aimed at internal efficiency (computing time / energy).

[0011] It would therefore be desirable to have a procedure that actively encourages users to contribute content to a specific topic based on minimal, standard, or extensive information. The barrier to interaction with the system should be as low as possible and require no prior technical experience. Ideally, the system should also be fully accessible.

[0012] The object of the present invention is therefore to provide a method for processing and displaying data using artificial intelligence, with which a user can be guided through an interactive process. RELEFANT Status: 3.1E25 RLE200001EPWO It is automatically stimulated to provide information, and efficiently controls internal processing with regard to computing time, processor / memory load and / or energy consumption.

[0013] This task is solved by a method for capturing, processing and presenting data using artificial intelligence (AI), comprising the following steps: a) Determining contextual information relevant to a user, generated based on user data, user input, stored information and / or environmental conditions; b) Analyzing the contextual information using an algorithm with the aid of AI to identify relevant topics, ideas and / or suggestions that are suitable to encourage the user to enter further data; c) Generating one or more personalized suggestions based on the analysis of contextual information, with the suggestions specifically tailored to the identified topics or interests of the user; d) Automatically providing the generated suggestions to the user in the form of a linguistic, textual, haptic and / or visual output, e) Capturing user input in response to the generated output, wherein at least one user-related metric is automatically captured and / or derived from interaction data during or between steps a)-e), and that the generation and / or output of the suggestions is adaptively controlled depending on the at least one user-related metric in such a way as to reduce computing time, processor / memory load and / or energy consumption of the data processing device while maintaining a predetermined minimum output quality.

[0014] The surprising insight of the present invention lies in the fact that a user can be specifically encouraged to systematically record memories through user-metrics-based, context-adaptive suggestions, while simultaneously controlling internal data processing in a resource-efficient manner. The method according to the invention solves the technical problem of how to efficiently process large amounts of contextual information and generate user-relevant suggestions. RELEFANT Status: 3.1E25 RLE200001EPWO Suggestions can be automatically generated to encourage the entry of further data.

[0015] The technical solution according to the invention consists of processing data in such a way that it specifically encourages a user to create structured documentation of their memories, going beyond a simple data display. The combination of AI-supported analysis, the generation of personalized suggestions, and the immediate output constitutes the technical solution according to the present invention.

[0016] The use of AI to analyze contextual information and automatically generate personalized suggestions requires systematic and structured processing that goes beyond a purely algorithmic solution and efficiently utilizes system resources.

[0017] The method according to the invention therefore provides optimized data processing and display logic, which is not available in classic software solutions without AI. The method according to the invention interacts systematically with sensor, storage, and computing resources.

[0018] The AI ​​takes over tasks of analyzing and recognizing contextual information that would far exceed purely manual processing, and controls the depth of processing (context window), model variants and output rate based on user metrics.

[0019] Human interaction supports the technical processes and continuously improves the accuracy of AI analyses. This aspect contributes to the technical solution because the interaction process is controlled systematically and automatically.

[0020] The interaction between user data and the AI ​​algorithm is not limited to an abstract process, but relies on technical means such as memory management, sensor data processing, and context-dependent content display. The invention produces a technical effect because the context-dependent data display and analysis are automated and technically performed. RELEFANT Status: 3.1E25 RLE200001EPWO

[0021] The method according to the invention thus provides a solution for the task of automated and context-sensitive data processing, which is optimized through human interaction.

[0022] According to the invention, it has proven advantageous to first generate relevant contextual information. User data includes all information stored specifically for the user, such as personal preferences, interests, previous interactions, and behavioral patterns.

[0023] In addition to this information, user data can also include demographic information such as age, gender, language preferences, and previous topics that the user found interesting. User data can be entered by the user themselves, but this is not mandatory. According to the inventive method, it is also possible for third parties to store or enter this information for a user.

[0024] User input, in contrast to the previously described user data, encompasses all interactions and information that the user has actively entered into the system, such as search queries, answers to questions, preferences, or feedback on suggestions. This also includes spontaneous entries or keywords that the user has used in previous interactions.

[0025] The user-metrics-based control of communication according to the invention produces a demonstrable technical effect, as the internal operation of the data processing device becomes more efficient (e.g., shortened inference paths, reduced context depth, limited number of parallel hypotheses, adaptive output rate). This leads to a measurable reduction in computing time, load, and / or energy consumption while maintaining a minimum output quality. The technical benefit thus arises from the control logic, which is aligned with the internal functioning of the computer, and not merely from content-based personalization.

[0026] In this context, "user metric" refers to a A quantitative measure derived from user interactions, device parameters, and / or sensor data that controls the adjustment of generation and output steps. Examples of user metrics include, but are not limited to, frequency. RELEFANT Status: 3.1E25 RLE200001EPWO These include repeat requests, the severity of device jitter, a device-set zoom / font size level, typing error rate and / or autocorrect statistics, color perception parameters, and a memory performance indicator. The examples mentioned are not exhaustive and can be combined.

[0027] Information stored according to the present invention includes all data collected and archived during use, such as previous interactions / use of the method, archived notes, completed tasks, or specific topics created by the user in the past. System-stored events or milestones may also be included.

[0028] By optionally accessing archived data, the inventive method can find links to previous topics or memories that might still be relevant to the user. For example, it could make suggestions that build on a previously created topic or remind the user of topics that were important to them. This creates continuity and helps to organize information and topics that the user might want to develop further.

[0029] Environmental conditions according to the present invention relate to factors such as the user's current location, the time of day, the weather, the device used (smartphone, laptop, etc.), and other external factors that can influence the user's context. For example, it may be useful to contextualize data seasonally, such as Christmas in December and / or Easter in spring. Certain events, such as the Olympics or World Championships, may also provide points of reference.

[0030] According to the invention, there are therefore various ways to obtain the relevant contextual information. One possibility is that user data is directly recorded. Another possibility is that the user enters data themselves. A further possibility is that stored information and / or contextual data is accessed. This thus makes it possible to RELEFANT Status: 3.1E25 RLE200001EPWO This means that an interaction is triggered based on a very small amount of data, for example, that it is Easter, or based on a very large amount of data. This will be explained in more detail below using an example.

[0031] After the relevant contextual information has been determined as a starting point, it is analyzed using an algorithm employing AI to identify relevant topics, ideas and / or suggestions that are suitable to encourage the user to enter further data.

[0032] Such suggestions can be generated by identifying topics or keywords in the data already entered and tailored to the user's interests and preferences. This ensures that the content suggested by the system is highly useful to the user and encourages them to enter further information.

[0033] According to the inventive method, artificial intelligence (AI) is used as a central component of the analysis and personalization mechanisms. Its main objective is to continuously process the user's contextual information and generate individual suggestions that motivate the user to further interaction and data input.

[0034] The AI ​​dynamically adapts the content to the current user context, enabling a deep understanding of the user's preferences and needs.

[0035] The AI ​​captures and integrates diverse data sources, such as user input, past interactions, location information, and time data, to create a comprehensive picture of the user's context. Through machine learning (ML), the AI ​​is trained to recognize relevant patterns and behaviors from this data and accurately assess the user's current state. For example, the AI ​​can compare past interests with current circumstances to predict the most likely topics that might appeal to the user.

[0036] Within the AI ​​architecture, various techniques are used according to the invention, including NLP for analyzing text input and ML algorithms for pattern recognition and personalization. NLP is used to RELEFANT Status: 3.1E25 RLE200001EPWO To identify key themes and emotions in user input, enabling the AI ​​to formulate suggestions in a way that appeals to the user.

[0037] Machine learning algorithms learn from previous interactions which suggestions were most relevant to the user and adapt these suggestions in real time based on current and previous contextual information. This continuous learning capability ensures that the AI ​​improves its accuracy and relevance over time.

[0038] The AI ​​algorithms according to the invention then subsequently generate one or more personalized suggestions based on the analysis of the context information, wherein the suggestions are specifically tailored to the identified topics or interests of the user.

[0039] These are provided to the user in the form of a linguistic, textual, haptic and / or visual output, and their reaction to the generated output is recorded and further processed.

[0040] According to the invention, it may be provided that the determination of contextual information from the user is carried out by means of questions from a digital avatar that addresses the user by means of audio and / or video.

[0041] A digital avatar according to the invention can be advantageous in communication with people because it makes the interaction more personal and engaging. Avatars can simulate human characteristics through facial expressions, gestures, and speech, which promotes understanding and emotional connection with users.

[0042] According to the invention, it may be provided that an avatar creates a friendly and often trusting atmosphere that encourages the user to exchange information.

[0043] According to one embodiment of the invention, it can also be advantageous for avatars to dynamically present information and adapt it to the respective user, making the communication experience more interactive and natural.

[0044] According to one embodiment of the present invention, user input can be provided by means of text, haptic input, speech and / or visual recognition, wherein in particular a keyboard, a RELEFANT Status: 3.1E25 RLE200001EPWO A mouse, a trackpad, a microphone, a camera and / or a sensor device for capturing movements is used.

[0045] The use of different input devices such as keyboard, mouse, trackpad, microphone, camera and sensor devices enables versatile and user-friendly interaction according to the present inventive procedure.

[0046] This wide variety of input options gives the user the flexibility to choose the method they consider best suited to their particular environment and preferences.

[0047] The keyboard offers precise, text-based input, while the mouse and trackpad allow for intuitive navigation and selection. These haptic input methods are particularly helpful when users require a high degree of control and accuracy over their input, such as when editing text or selecting options on a screen.

[0048] However, there are also many situations where using a keyboard and / or mouse is not practical, for example when walking, driving, etc.

[0049] There are also physically impaired users who find it difficult to use the keyboard and mouse, or who simply have poorer visual perception skills.

[0050] In these cases and / or for users with disabilities, voice input via a microphone can represent a fast and natural form of interaction, offering significant advantages, especially in situations where haptic input is difficult or time-consuming.

[0051] It allows users to easily submit complex requests or instructions via voice command, increasing ease of use and improving interaction efficiency. At the same time, voice input promotes barrier-free use for people with limited mobility or typing difficulties.

[0052] Finally, cameras and sensor devices, which are also optionally provided according to the present invention, expand the possibilities of RELEFANT Status: 3.1E25 RLE200001EPWO Input is via visual recognition and motion detection. The camera can be used, for example, for facial recognition or gesture capture, while sensor devices can detect and interpret user movements.

[0053] These input methods enable intuitive, touchless control, which is particularly useful in environments where the user cannot make direct physical inputs.

[0054] The combination of these different input methods enables the inventive method to offer an individual, seamless interaction.

[0055] According to one embodiment of the present invention, it may be provided that information about one or more users or a group of users is entered or determined, in particular including whether it is an individual, a group of persons or an organization, and / or characteristics of the user(s), in particular including age, gender, state of health.

[0056] This investigation preferably takes place before step a), but can take place continuously.

[0057] In this context, it can be particularly advantageous for an algorithm to use AI, with the provision of a photo, an audio file, a video recording and / or audio- or audiovisual communication with the user, to determine the information and / or characteristics of the user(s).

[0058] The described method, which collects and utilizes information about individual users or user groups, including their characteristics, offers significant advantages for precise and context-aware adaptation of system interactions. The ability to inform the system whether the user is an individual, a group, or an organization allows for better adaptation of suggestions and outputs to the specific situation.

[0059] This allows the system to take into account different interests and preferences within a group of users, or to facilitate common interactions. RELEFANT Status: 3.1E25 RLE200001EPWO promote, while in the case of an individual it specifically addresses individual needs.

[0060] A group of people within the meaning of the present invention could, for example, be a family or a circle of friends who wish to jointly document and record certain experiences, events or journeys.

[0061] Organizations could have their specific organizational needs taken into account in order to offer relevant content.

[0062] An organization within the meaning of the present invention may be a company, an association, or the like, which, for example, wants to create an association chronicle.

[0063] In addition to recording the number of users and the user type, according to one embodiment of the present invention it is advantageous to record specific characteristics of the users such as age, gender and health status, since this information increases the personalization and relevance of the content.

[0064] For example, the system can provide age-appropriate content or take health aspects into account if the information is aimed at the well-being of the users.

[0065] According to the inventive method, this information can be entered by the user(s) themselves or by third parties. However, it may be advantageous for the determination of this user information and characteristics to be carried out using an AI-supported algorithm, as well as for the possibility of providing photo, audio, or video files for the identification of these characteristics.

[0066] For example, the AI ​​can use a photo to determine the user's estimated age, gender, or mood, and thus encourage targeted, needs-based interactions.

[0067] Audio files or videos allow for an analysis of voice and body language, which could enable the system to identify emotional states or preferred communication styles.

[0068] This AI-supported analysis ensures that information about the user is updated and used in real time, resulting in a dynamic and personalized experience. RELEFANT Status: 3.1E25 RLE200001EPWO This allows for coordinated adaptation of content and communication methods. This increases the efficiency and relevance of the responses from the algorithm according to the invention and enables user-tailored interaction.

[0069] According to an embodiment of the invention, it may be provided that contextual information for the user or contextual information about the user is entered by another person.

[0070] It can be advantageous if another user inputs information into the algorithm according to the invention, e.g., asks questions or creates another user profile. For example, the additional user can specifically ask questions about the past or about other contacts.

[0071] Furthermore, according to the invention, it may be preferred that contextual information is obtained by the algorithm asking the user targeted questions about places, interests and / or persons.

[0072] It may be provided that when a person is mentioned, the algorithm according to the invention generates questions about that person, e.g. where the user knows them from, how old they are, where they come from, etc.

[0073] Similarly, such context-based follow-up questions about art / artists, sports / athletes, etc. can be helpful.

[0074] Furthermore, according to one embodiment of the present invention, it may be preferred that an output rate control is implemented which automatically varies the speech or text output rate and / or the length of individual output segments depending on the user-related metric.

[0075] To implement this, the system can have one or more resource controllers that, based on user metrics, (i) select model variants (e.g., lightweight vs. full model), (ii) define context retrieval and depth (maximum number of candidates / embedding length), and / or (iii) dynamically control the output rate (tokens / second or speech rate). The parameters can be adjusted via thresholds and hysteresis to prevent unstable switching. RELEFANT Status: 3.1E25 RLE200001EPWO

[0076] It may be provided that the number and / or depth of the generated suggestions are dynamically limited depending on the user-related metric, whereby a minimum quality is ensured by means of a confidence-Z quality measure.

[0077] In a preferred embodiment, frequently reusable content is cached locally and, when user metrics are unfavorable, is preferentially retrieved from the cache to reduce network and inference load. The cache can feature metric-weighted prioritization (e.g., LRU with extended retention times) and a quality check (minimum quality).

[0078] According to one embodiment, it may be preferred that frequently reusable content is cached locally and preferably output from the cache depending on the user-related metric.

[0079] Furthermore, according to the invention, the system can automatically switch between an active dialog mode (system initiates suggestions) and a passive participation mode (system reacts to input) depending on user metrics, in order to reduce avoidable system activity and save resources. Thresholds and hysteresis prevent frequent mode changes.

[0080] The system may be designed to collect user metrics, including the frequency of repeat requests and the length of pauses in speech. If the repeat frequency exceeds a threshold, the system automatically reduces the number of parallel suggestion candidates, increases the output resolution (shorter segments), and lowers the inference budget limits. Input processing is debounced if device jitter is detected; at high zoom levels, compact, high-contrast outputs are selected.

[0081] The invention also provides a system comprising a data processing device as well as an input and an output device, designed and configured for carrying out a method according to the invention.

[0082] Finally, the invention provides the use of a method or system according to the invention for interacting with a user to systematically capture the user's memories. RELEFANT Status: 3.1E25 RLE200001EPWO It may be possible for the recorded memories to be displayed in textual or multimedia format, with the processing being carried out using algorithms and AI.

[0083] The present method according to the invention will now be explained in more detail with reference to an exemplary embodiment.

[0084] The user is, for example, an elderly gentleman whose children wish him to systematically record memories concerning his life. The method according to the invention is intended to be used for this purpose, enabling a structured approach.

[0085] In the first step according to the inventive method, the contextual information relevant to the user is recorded.

[0086] For this purpose, the method according to the invention can be generated as described on user data, user inputs, stored information and / or environmental conditions.

[0087] It may therefore be envisaged, for example, that the user's children create a profile of the user and provide comprehensive information about their age, place of birth, places of residence, as well as family information about life partners, parents, siblings and profession.

[0088] Alternatively, the method according to the invention can begin with minimal relevant contextual information and then acquire or iteratively supplement the necessary further contextual information during the course of the method itself. For example, if only a minimal photograph of the user is provided or taken by the system itself, a starting point for the relevant contextual information, such as approximate age, gender, etc., can already be determined from this using the algorithms according to the invention.

[0089] In addition, user-related metrics are automatically collected and / or derived from interaction data; these may include, in particular: (i) the frequency of retry requests, (ii) the intensity of device jitter, (iii) a device-set zoom / font size level, (iv) typing error rate and / or autocorrect statistics, (v) color perception parameters, and (vi) a memory performance indicator (not exhaustive). The collection may RELEFANT Status: 3.1E25 RLE200001EPWO This can be done on the device side (e.g., OS settings) and / or based on sensor signals (microphone, camera, motion sensors).

[0090] Based on this very small, medium or very extensive contextual information and the user-related metrics, the algorithm according to the invention uses AI to identify relevant topics, ideas and / or suggestions that are suitable to encourage the user to enter further data and addresses the user accordingly.

[0091] A resource controller adaptively manages the suggestion pipeline based on user-related metrics to reduce computation time, processor / memory load, and / or energy consumption while maintaining a minimum quality standard. This includes, in particular, output rate control (tokens / s or TTS speed), limiting the number / depth of suggestions (retrieval candidate cap, context window), an early exit strategy at high confidence levels, and / or model variant selection (lighter / quantized model).

[0092] A dialogue is created by repeating the steps of interactive user input and queries.

[0093] In this example, one of the children enters the specific question, "What was your wedding day like?" The system links the input with stored contextual data (e.g., date / location, available photos) and simultaneously considers current user metrics: If there is a high repetition frequency and longer pauses in speech are detected, the system automatically reduces the number of parallel suggestion candidates, increases the output resolution (shorter segments), and lowers the Inference budget limits; if device jitter is detected, input is debounced; at high zoom levels and font sizes, compact, high-contrast outputs are selected. Thus, in the first output, only two short, highly relevant keywords are suggested (e.g., best man, location) before a more in-depth prompt is asked.

[0094] In a preferred embodiment, frequently reusable content is cached locally and preferably output from the cache when user metrics are unfavorable; this reduces network and inference load. RELEFANT Status: 3.1E25 RLE200001EPWO Metric-weighted prioritization (e.g., LRU with extended retention periods) and a quality check (minimum quality) are planned.

[0095] Furthermore, depending on user metrics, it is possible to automatically switch between an active dialogue mode (system initiates suggestions) and a passive participation mode (system reacts to input); Thresholds and hysteresis prevent frequent mode changes.

[0096] Alternatively, this information can be displayed on a kind of "to-do list", and the user can decide when they want to address these questions.

[0097] The information is processed by an algorithm according to the invention using Kl. Both textual and multimedia processing are possible. It can also be provided that new photos and / or videos can be generated from certain relevant contextual information to complete the processing.

[0098] The method according to the invention thus enables, as shown in the exemplary example, a user without technical skills and possibly with physical limitations to comprehensively use audio and, if necessary, Video communication can be structured to encourage the addition of reminders to a system, while these inputs are then used to further optimize the process. Third parties can actively participate in the process and provide relevant contextual information at the beginning or continuously.

[0099] The entered relevant contextual information is then structured and efficiently processed and made available to the user and, if applicable, other users in various technical output formats, such as audio, video and / or text.

[0100] User-metrics-based control shortens inference paths, limits context depth, and adjusts the output rate; this reduces computation time, processor / memory load, and / or energy consumption while maintaining a minimum output quality, thus enabling scalable interaction on given hardware. RELEFANT Status: 3.1E25 RLE200001EPWO

[0101] The features of the invention disclosed in the preceding description, the claims and the drawings may be essential for the realization of the invention in its various embodiments, both individually and in any combination.

Claims

RELEFANT Status: 3.1E25 RLE200001EPWO Patent claims 1. Methods for capturing, processing and presenting data using artificial intelligence (AI), comprising the following steps: a) Determining contextual information relevant to a user, generated based on user data, user input, stored information and / or environmental conditions; b) Analyzing the contextual information using an algorithm with the aid of AI to identify relevant topics, ideas and / or suggestions that are suitable to encourage the user to enter further data; c) Generating one or more personalized suggestions based on the analysis of contextual information, with the suggestions specifically tailored to the identified topics or interests of the user; d) Automatic provision of the generated suggestions to the user in the form of a verbal, textual, haptic and / or visual output, e) Capture of user input in response to the generated output, characterized in that during or between steps a)-e), at least one user-related metric is automatically captured and / or derived from interaction data, and the generation and / or output of suggestions is adaptively controlled depending on the at least one user-related metric in such a way as to reduce computing time, processor / memory load and / or energy consumption of the data processing device, while maintaining a predetermined minimum quality of output.

2. The method according to claim 1, characterized by Determining contextual information from the user by asking questions of a digital avatar that addresses the user via audio and / or video. RELEFANT Status: 3.1E25 RLE200001EPWO 3. Method according to claim 1 or claim 2, characterized in that the user input is provided by means of text, haptic input, speech and / or visual detection, wherein in particular a keyboard, a mouse, a trackpad, a microphone, a camera and / or a sensor device for detecting movements is used.

4. Method according to one of the preceding claims, characterized in that Information about one or more users or a group of users is entered or determined, in particular comprehensively whether it is an individual, a group of persons or an organization, and / or characteristics of the user(s), in particular comprehensively including age, gender, health status.

5. Method according to claim 4, characterized in that An algorithm using AI, with the provision of a photo, an audio file, a video recording and / or audio or audiovisual communication with the user, determines the information and / or characteristics of the user(s).

6. Method according to claim 4 or 5, characterized in that Contextual information for the user, or contextual information about the user entered by another person.

7. Method according to one of the preceding claims, characterized in that Contextual information is obtained through targeted questions from the algorithm to the user, asking about locations, interests and / or people.

8. Method according to one of the preceding claims, characterized in that the user-related metric according to claim 1 comprises at least one of the RELEFANT Status: 3.1E25 RLE200001EPWO The group includes: (i) frequency of retry requests to the computer, (ii) strength of device jitter, (iii) device-set zoom / font size level, (iv) typing error rate and / or autocorrect statistics, (v) color perception parameters, and (vi) a memory performance indicator.

9. Method according to one of the preceding claims, characterized in that an output rate control is implemented which automatically varies the speech or text output rate and / or the length of individual output segments depending on the user-related metric.

10. Method according to one of the preceding claims, characterized in that the number and / or depth of the generated suggestions are dynamically limited depending on the user-related metric, wherein a minimum quality is ensured by means of a confidence-Z quality measure.

11. Method according to one of the preceding claims, characterized in that frequently reusable content is cached locally and preferably output from the cache depending on the user-related metric.

12. Method according to one of the preceding claims, characterized in that, depending on the user-related metric, the system automatically switches between an active dialogue mode (system initiated) and a passive participation mode (system only reacts).

13. System comprising a data processing device as well as an input device, a sensor device and an output device, designed and configured for carrying out a method according to one of the preceding claims.

14. Use of a method according to any one of claims 1 to 12 or of a system according to claim 13 for interacting with a user in a structured manner RELEFANT Status: 3.1E25 RLE200001EPWO Capturing user memories.

15. Use according to claim 14, wherein the captured memories can be displayed in textual or multimedia format, the processing being carried out using algorithms with the aid of AI.