PROMPT-BASED SUMMARIZATION FOR AUTOMATIC ROOT CAUSE ANALYSIS (RCA) The present invention relates to a method, system and computer-readable medium for prompt-based summarization for automatic root cause analysis, RCA. Root Cause Analysis (RCA) is a problem solving process used in different technical domains such as software development, telecommunication, manufacturing and digitization. For instance, when an electronic or digital infrastructure in a smart city is shut down, the maintainer or developer would first investigate in which part of the equipment the error occurs, then explore what are the causes and finally conduct certain actions to address the issue. This problem-solving process will normally generate a large amount of RCA reports in text format, which describe the details of the problems, causes and solutions. Text based RCA reports efficiently store the valuable historical information, but it is a challenge for humans to read the comprehensive reports and extract the important information. To address this issue, tools have been developed (for an overview, see SAP, “Root Cause Analysis Tools”, 2022, https://support.sap.com/en/alm/solution- manager/expert-portal/root-cause-analysis-overview.html) that, instead of exploring information from existing RCA reports, organize the day-to-day RCA operations and aim at helping human users to explore the solution, such as Dynatrace (for reference, see Dynatrace, “Automatic Root-Cause Analysis”, 2022, https://www.dynatrace.com/support/help/how-to-use-dynatrace/ problem-detection- and-analysis/problem-analysis/root-cause-analysis). In contrast, embodiments of the present disclosure focus on generating automatically interest-oriented summaries from the lengthy RCA reports to accelerate the speed of RCA analysis. Different from normal NLP-based text summaries, the interest-oriented summaries are expected to only contain information, which is of interest for the user. The current state-of-art NLP technology for this task is prompt-based summarization. The prompt is certain textual input in the form of a set of keywords or a descriptive sentence, such as “The purpose of present invention is”. Via prompts, it is possible to interact with the respective summarization system to control the model to generate the summary only about the ‘invention purpose’ (as disclosed in Junxian He, Wojciech Kryscinski, Bryan McCann, Nazneen Rajani, Caiming Xiong, “Ctrlsum: Towards generic controllable text summarization”, arXiv preprint arXiv:2012.04281, 2020. To summarize a RCA report, it is important to retain the correct relation between different problems and corresponding causes and solutions. However, the Ctrlsum framework cited above, like other existing prompt-based approaches, works with a single prompt only and, hence, focusses only on an individual and fixed topic, i.e., analyses the topics independently, which cannot catch relationships among problems, causes and solutions in one document. Additionally, the control of the summary length is also a common challenge for the current abstractive summarization. It is therefore an object of the present invention to improve and further develop a computer-implemented method, a system and a computer-readable medium for prompt-based summarization for automatic RCA by processing lengthy RCA reports in such a way that the speed and effectiveness of RCA is enhanced. In accordance with the invention, the aforementioned object is accomplished by a computer-implemented method of prompt-based summarization for automatic root cause analysis, RCA, the method comprising: receiving, from a user, an input textual RCA report of interest; determining a problem prompt; creating, based on the determined problem prompt, a set of chained problem, cause and solution prompts; using a text summarizer model to iteratively generate summaries of the problems, causes and solutions mentioned in the input textual RCA report that correspond to the created set of chained problem, cause and solution prompts; and integrating the generated summaries to construct a RCA Summary such that dependencies between problems and their corresponding causes and/or solutions are maintained. Furthermore, the aforementioned object is accomplished by a prompt-based summarization system for automatic RCA and by a tangible, non-transitory computer-readable medium as specified in the independent claims. Embodiments of the present invention provide prompt-based summarization systems and methods for automatic RCA that generate chained prompts, thus to generate a RCA summary which retains the relationship among problems, causes and solutions. Compared to prior art systems that generate a summary directly from an entire long RCA document, methods and systems according to embodiments of the present disclosure, which generate the summary based on a content meeting users’ preference and clearly retain the relationship among different problems, causes and solutions, achieves a better efficiency. This is simply because the common summarization approach aims to compress as much information as possible into a shorter paragraph. Giving this kind of summary, users still need to read through the text to find out causes and solutions for each problem. In contrast, a system according to embodiments of the present disclosure can summarize only the information, which are of interest to the respective user. Besides, an automatic RCA summarization system according to embodiments, may also provide the summary in a readable ordered lists and the summary may clearly present each problem with corresponding causes and solutions in tree-structure, which is extremely efficient for users to do the RCA analysis and solve a given technical problem. Assuming a system generates the RCA summary by using only a normal prompt- based summarization model, the system requires that the user manually provides problem, cause and solution prompts every time when the user input a new document. Embodiments of the present disclosure provide a RCA Prompt Constructor that helps in overcoming this shortcoming by automatically constructing the chained prompts for text summarization. The SAP RCA Analysis Tool as described in SAP, “Root Cause Analysis Tools”, 2022, https://support.sap.com/en/alm/solution-manager/expert-porta l/root-cause- analysis-overview.html) works towards simplifying the problem resolution process specifically in an IT environment. The document provides a systematic top-down approach helping users to quickly isolate a component causing a problem. However, the disclosed system is dedicated to the IT RCA analysis and aims to help users to identify the root cause. Different from this approach, embodiments of the present disclosure are generic to different domain from R&D to manufacturing. Instead of helping users to solve the current individual problem, systems as disclosed herein focus on processing numerous existing RCA reports to help users to summarize the valuable information from historical experiences. Based on summarized valuable experience, users are able to either solve an individual problem or systematically optimize their processes. In an embodiment, it may be provided that a final RCA summary is built by automatically adjusting the length of the RCA Summary based on physical properties and/or specifications of the user. In this way, namely by relying on physical properties and/or specifications, which may be provided by or derived from the user, to adjust the length of the RCA Summary the user can control the behaviour of the RCA summarization system. In an embodiment, the physical properties and/or specifications of the user may include non-static physical parameters of the user. For instance, such non-static physical parameters of the user may include or may be determined from a touch force or pressure that is exerted by the user on a respective hardware device on which the RCA summarization system is implemented. Alternative properties or parameters may include, e.g., the user’s heart rate, an act of looking, or a predefined interaction with the RCA summarization system. In an embodiment, the physical parameters of the user may be provided by or may be derived from the user in the context of a biometric identification process of the user. In this context, a biometric processing module may be provided that is configured to recognize the user based on the sensed biometric information. Specifically, the summary length can be automatically adjusted based on user physical properties such as the touch pressure, or the user can control the summary length by touching biometric recognizer/identification device with different force, such that the summary length can be controlled based on human physical properties. In an embodiment, the present disclosure provides a chained prompt-based text summarization method in which the RCA Summary is generated by summarizing each problem into a problem summary and organizing problems in an ordered list, by summarizing, for each problem of the problem summary, all related causes and organizing them in an ordered list, and by summarizing, for each cause of each problem of the problem summary, all related solutions and organizing them in an ordered list. The summary (output) may be in format of an ordered list, wherein the first sentence of each summary chunk is the same as one of the prompts. In an embodiment, a prompt database may be provided that stores histories of problem prompts previously deployed by users. Based thereupon, if a respective user exists already in the prompt database, the problem prompt may be determined by retrieving it from the prompt database. On the other hand, if the user does not yet exist in the prompt database, the problem prompt may be determined by generating a new one from the scratch. In an embodiment, the text summarizer model may be trained on a training dataset that contains sequential text summaries as ground truth. For example, the text summarizer model may be implemented in form of a CTRLsum model. In this context, it is noted that the CTRLsum model (as described in Junxian He, Wojciech Kryscinski, Bryan McCann, Nazneen Rajani, Caiming Xiong, “Ctrlsum: Towards generic controllable text summarization”, arXiv preprint arXiv:2012.04281, 2020) presents a novel CTRLsum framework for controllable summarization which connects the summary output and the users’ preference and expectations. It is a summarization framework that enables users to control aspects of generated summaries by interacting with the system through textual input in the form of a set of descriptive prompts. However, in the CTRLsum framework as disclosed in the document cited above, all prompts are manually created. Moreover, the summarization system always works with a single prompt, which generates a summary focusing on a single and fixed aspect. In contrast, the system according to embodiments disclosed herein provides an automatic prompt method that iteratively constructs three-aspects prompts (i.e., problem, cause and solution) based on generated summaries in intermediates steps. In this way, the prompts are automatically customized for each input document. In an embodiment, it may be provided that building the final RCA summary includes the steps of generating a problem dependency graph that represents relations between different problems mentioned in the input textual RCA report, estimating an importance of each problem based on a length of a dependency chain of the respective problem in the problem dependency graph, and sorting the problems according to their estimated importance and building the final RCA summary by considering, from the sorted problems, only the top k problems, wherein the parameter k is determined based on the user specific parameters. In an embodiment, feedback may be collected from the user based on a click- through rate and/or reading time of the final RCA summary, wherein the collected feedback may be used to personalize the estimation of an importance of each problem. There are several ways how to design and further develop the teaching of the present invention in an advantageous way. To this end, it is to be referred to the dependent claims on the one hand and to the following explanation of preferred embodiments of the invention by way of example, illustrated by the figure on the other hand. In connection with the explanation of the preferred embodiments of the invention by the aid of the figure, generally preferred embodiments and further developments of the teaching will be explained. In the drawing Fig.1 is a diagram illustrating the basic architecture of a prompt-based summarization system for Automatic Root Cause Analysis, RCA, according to an embodiment of the present invention, Fig.2 is a diagram illustrating a basic architecture of an automatic chained RCA prompt constructor of a prompt-based RCA summarization system according to an embodiment of the present invention, Fig.3 is a diagram illustrating pseudocode of the functionality of a chained RCA prompt constructor of a prompt-based RCA summarization system according to an embodiment of the present invention, and Fig.4 is a diagram illustrating a structure of a problem dependency graph used in a prompt-based RCA summarization system according to an embodiment of the present invention. Root cause summarization aims to summarize each problem and its corresponding causes and solutions from usually rather long documents of the Root Cause Analysis (RCA). As an example, the RCA summary of problems and solutions of electronic city infrastructures is beneficial for a smart city to improve their automation. One challenge is to clearly retain the relationship among different problems, causes and solutions. Additionally, an automatically length-controllable summary is another common challenge in text summarization domain. Embodiments of the present disclosure addresses both issues via: 1) automatically generating chained prompts for RCA summarization; 2) adjusting the text summary length based on physical properties (e.g., eye or touch perception) of a human user. Fig.1 schematically illustrates a basic system architecture of a prompt-based RCA summarization system 100 in accordance with embodiments of the present disclosure. More specifically, Fig.1 illustrates a comprehensive procedure of how the prompt-based summarization system 100 automatically generates an adaptive summary for an RCA analysis in accordance with embodiments of the present disclosure. According to an embodiment, it may be provided that the system 100 is embedded in a computation device of a user 102, e.g., in a mobile device such as a smart phone, a tablet or the like. In the following, the individual components and functionalities of the system 100 and their interactions will be described in detail with reference to Fig.1. 1. Triggering the system 100 In accordance with embodiments of the present disclosure, the process may be started by the user 102 by interacting with a biometric processing module 104, as shown at S10 in Fig.1, and by providing at least one textual RCA report of interest 106 as input, as shown at S12 in Fig.1. 2. Constructing chained RCA prompts After sensing the user’s 102 biometric information (e.g., fingerprint and/or eye iris) through the biometric processing module 104, the system 100 may apply a Biometric Recognizer 108 (which may be implemented as a functional part of the biometric processing module 104) to recognize the individual user 102 based on the sensed biometric information. The output of the Biometric Recognizer 108 (i.e., the identified user 102) may be forwarded as input to an RCA Prompt Constructor module 112, as shown at S14 in Fig.1. Fig.2 illustrates a functionality of the RCA Prompt Constructor 112 used in the system of Fig.1 in accordance with embodiments of the present disclosure. Basically, as will be described hereinafter in detail, this constructor 112 may be configured to automatically compute chained RCA prompts by recursively utilizing outputs of a text summarization model 114. The RCA Prompt Constructor 112 may be triggered, e.g., when it receives the Biometric Identification from the Biometric Recognizer 108, as shown at S14 in Figs.1 and 2. Then, as shown at S16, the RCA Prompt Constructor 112 checks first if the user 102 exists in a Prompt Database 116. If yes, the RCA Prompt Constructor 112 will retrieve the prompt history of that user 102 from the database 116, as shown at S20, and output a Problem Prompt 118 extracted from the user’s 102 prompt history, as shown at S22. The Prompt Database 116 may be configured to store all the historical prompts of all users. For instance, a developer would focus more on technical problems, while a product owner would focus more on systematic issues. For example, when a particular developer reuses the system 100, the constructor 112 can simply retrieve the ‘technical problem’ prompt, which that user used before. The same applies for the product owner user or for any other previous user focusing on a particular kind or class of problems. If the user 102 does not exist in the database 116 (S24), a Chained Prompt Constructor module 120 of the RCA Prompt Constructor 112 will generate a new Problem Prompt 118 from scratch, as shown at S26. As shown in Table 1, the Problem Prompt 118 may be provided/generated in the format of [“The <(optional)domain> problems mentioned in this document include: 1)], where the “domain” information is optional. For instance, for developer users, a possible prompt could be [“The technical problems mentioned in this document include: 1)”]. The last string “1)” is used to guide the text summarization model 114 in generating the summary in an ordered list format. Table 1: The RCA Prompts As shown at S28, the Problem Prompt 118 is then used as an input to the Text Summarizer Model 114. The summarizer model 114 is designed to generate the summary of each problem mentioned in the original text 106 that was provided by the user 102 and to form a problem summary in an ordered problem list. As shown at S30, this ordered problem list is then forwarded to the Chained Prompt Constructor 120 as input to next generate a Cause Prompt 122. Given the ordered problem list, the Constructor 120 may be configured to automatically generate the Cause Prompt 122 in a chained manner. That is, as shown at S32, for each problem in the problem summary list, the Constructor 120 may construct a Cause Prompt 122 as [“The problem of <summary(Problem)> is caused by: 1)] (see Table 1 above), where the summary(Problem) is the problem summary generated in the previous step. Each Cause Prompt 122 may then be provided as input to the Text Summarizer Model 114, as shown at S34. The summarizer model 114 may be designed to generate, for each problem, the cause summary based on the original text 106 in an ordered cause list, as shown at S36. Given the ordered cause list, the Constructor 120 then continually generates a Solution Prompt 124 for each cause, as shown at S38. As shown in Table 1 above, the cause specific Solution Prompt may be provided/generated in the format of [“The problem of <summary(Problem)>, which is caused by <summary(Problem, Cause)>, can be solved via: 1)], where the summary(Problem) is the problem summary and summary(Problem, Cause) is the cause summary of the corresponding problem. Alternatively, the constructor 120 may be configured to construct Solution Prompts for the extracted problems, having the format of [“The problem of <summary(Problem) can be solved via: 1)]. In accordance with embodiments of the present disclosure, it may be provided that the decision of generating either cause-level or problem-level Solution Prompt can be made by the user 102. Same as for the Cause Prompt 122, each Solution Prompt 124 is then the input to the Text Summarizer model 114, as shown at S40. For each cause of each problem, the summarizer model 114 will then generate a solution summary based on the original text 106 in an ordered solution list. The functionality of the Chained Prompt Constructor 120 according to an embodiment of the present disclosure can be summarized by the pseudocode displayed in Fig.3. 3. Generating RCA summary After having generated the chained RCA prompts as described above, the system 100 proceeds by combining all problem, cause and solution summaries and outputting a united summary as a final comprehensive RCA Summary 126, as shown at S42. In the RCA Summary 126, all problems mentioned in the original long text 106 are summarized and organized in an ordered list. For each problem, all related causes are summarized in an ordered list. For each cause of each problem, the solutions are summarized in an ordered list. In this way, the dependencies between problems and their corresponding causes and solutions are maintained clearly in the final RCA summary 126, which is more efficient for further analysis. To generate summaries in an ordered list, the prompt-based Text Summarizer model 114, which may be implemented, e.g., in form of the CTRLsum model (as described in Junxian He, Wojciech Kryscinski, Bryan McCann, Nazneen Rajani, Caiming Xiong, “Ctrlsum: Towards generic controllable text summarization”, arXiv preprint arXiv:2012.04281, 2020, which is hereby incorporated by reference herein), may be trained on a training dataset 128 that contains sequential summaries as ground truth. During test, the Text Summarizer 114 therefore will generate problem, cause and solution summaries in ordered list by giving prompts which contain the trigger term “:1)” (shown in Table 1 above). 4. Adapt summary length to user physical properties According to embodiments of the present disclosure, the system 100 further includes an Adaptive Summary Generator 130. It may be provided that, given the raw RCA Summary 126 from the Text Summarizer model 114, the Adaptive Summary Generator 130 automatically adjusts the length of the RCA summary based on one or more physical parameters of the user 102 and provides a final RCA summary 140 with an adapted reduced length (as shown at S46), which is then provided as final output to the user 102 (as shown at S48). The physical parameters may include, but are not limited to, the user’s 102 heart rate, the act of looking (eye) and pressure on the respective hardware device on which the system 100 is implemented (e.g., a tablet or the like) when holding it. For example, when activating the system 100, the user 102 may press/touch a sensing area (e.g., a touch button or a touch screen) of the respective hardware device (e.g., a tablet or the like) naturally with certain physical force. To this end, the Biometric processor 104 of the system 100 may include a Haptic Interpreter 110 that is configured to detect different sense of the touch and convert this physical parameter to a numerical parameter as input to the Adaptive Summary Generator 130, as shown at S44 in Fig.1. The interpretation by the Haptic Interpreter 110 may be personalized, which means given the same touch pressure the system 100 will interpret a detected physical force differently based on different users. The idea is that the physical parameter of the human reflects the stress level. According to embodiments of the present disclosure, the Adaptive Summary Generator 130 may be configured to maintain a Problem Dependency Graph 131, which represents relations between different problems mentioned in the original text 106. The Problem Dependency Graph 131 may be created manually based on each RCA report. According to an embodiment, it may be provided that for each node 133 (i.e., problem) in the graph 131, a Node Importance Estimator 132 measures its importance based on the length of the corresponding dependency chain. Starting from a target node as a root node, the Estimator 132 may explore each path and sum up the amount of edges as a final importance score of this node. For example, Fig.4 depicts an exemplary dependency graph 131 corresponding to a specific input RCA report 106, where the nodes 131 present different problems mentioned in RCA report 106 and the edges present dependencies between the respective problems. For instance, the edge between problem 0 and 2 indicates that problem 0 leads to problem 2 (i.e., if problem 0 occurs, problem 2 will occur as well). As shown in the example of Fig.4, the corresponding original RCA report 106 describes nine problems, wherein problem zero is the most important one since it results in another five problems (as indicated by the dotted line arrows, starting from problem zero, the amount of linked edges is five). The most important problem therefore means the most influential problem that leads to most consequences. According to an embodiment of the present disclosure, it may be provided that, based on the parameter(s) received from Haptic Interpreter 110, the Adaptive Summary Generator 130 may be configured to shorten the raw RCA summary 126, e.g., by selecting the top k important problems including its corresponding causes and solutions. For example, a stronger touch force on the tablet screen may be interpreted to be indicative of the user 102 having a higher mental pressure (e.g., time-pressure). Adaptively, the system 100 may then generate a shorter RCA summary 140. Additionally, for example, by pressing the touch sensing area with different force, the user 102 is able to proactively control the length of the summary 140. Moreover, the estimator module 132 of the Adaptive Summary Generator 130 may be configured to continuously improve the importance estimation based on user feedback (as shown at S50), such as a click-through rate and/or reading time. As already mentioned above, finally, the adaptive RCA Summary 140 may be returned to the user 102 as final output of the system 100, as shown at S48. Once the adaptive RCA summary 140 is generated and sent to the user 102, the process ends and may be triggered by again by receiving a new textual RCA report of interest from a user. In an embodiment, the system may be setup by execution of one or more of the following steps (cold start): 1) Create a Biometric Processor 104 to detect physical properties (e.g., touch force) of users 102 and recognize user identity via biometric information (e.g., fingerprint). 2) Create a Prompt Database 116 to store prompts used by previous (regular) users 102. 3) Create a RCA Prompt Constructor 112 to automatically construct the chained Problem, Cause and Solution prompts. 4) Prepare a Training Dataset 128 and train a Text Summarizer model 114, such as CTRLsum model, which generates a shortened version of a long RCA document 106 while preserving the important information that meets the users’ preference. 5) Provide a Problem Dependency Graph 131 such as in a knowledge graph format, which represents the relationships among all problems mentioned in the long RCA document 106. 6) Create an Adaptive Summary Generator 130 that is configured to decide to return how many and which important problem with its corresponding cause and solution summaries as final output 140, optionally based on user physical properties. According to some embodiments, when new data arrive, i.e., when a user 102 provides or inputs a new RCA document 106 for summarization, one or more of the following steps may be executed: 1) Optional: The user 102 may touch a sensing area of the hardware (e.g., a tablet) and activate the system 100. 2) The user 102 may input the (long) RCA document 106 of interest into the system 100. 3) The system 100 generates the RCA summary 140 via: a. The RCA Prompt Constructor 112 constructs problem, cause and solution prompts. b. The Text Summarization Model 114 generates an initial RCA summary 126. c. The Adaptive Summary Generator 130 adjusts the final summary length based on user physical properties (e.g., touch pressure) interpreted by the Biometric Processor 104. 4) Collect the returned RCA summary output, which contains one or multiple or all Problems with corresponding Causes and Solutions. 5) Optional: User 102 may provide feedback to the system 100 to indicate whether the Problem(s) shown in the output are real crucial Problems. One or more aspects of the present disclosure provide a method to automatically generate a tree-structured Root Cause Analysis summary by using prompt-based summarization technology, which retains the relationship between different problems and their corresponding causes and solution. When generating the RCA summary with prompt-based summarization technology, an automatic method may be provided that iteratively constructs the chained Problem, Cause and Solution prompts based on the text summary. One or more aspects of the present disclosure provide a method to determine the length of the final summary output based on human physical properties (e.g., touch force or look in one’s eye) and the importance of each problem. In this context, the system may shorten the final output by showing the RCA summary of top k important problems based on human physical properties (e.g., stress level). While using the system, an estimation of the problem importance may be personalized and harmonized with the problem dependency graph based on user feedback, which may be collected when users are utilizing the system. Systems and methods according to embodiments disclosed herein can be used in a variety of different technical use cases, some of them will be exemplarily described hereinafter in some more detail. In particular, the disclosed techniques can be used in many applications where AI predictions are presented to human users. For example, it can be used in any instance where text summary based RCA analysis can be used, such as in digital government, public safety or carbon neutrality domains. However, as will be appreciated by those skilled in the art, the described use cases are only exemplarily and the present invention is not restricted to the described use cases. Differentiators to common RCA systems include a tree- structured RCA summary, an automatic prompt construction, and/or an adaptive summary length, i.e., 1) the system may generate RCA summaries, which clearly retain the relationship among problems, causes and solutions in a tree-structure; 2) the system may automatically construct chained prompts for a summarization model; 3) the system may automatically adapt the summary length based on human physical properties (e.g., touch pressure on a tablet screen). Use Case 1: AI-enhanced City Maintenance Use Case: A smart city may reduce carbon emissions by increasing the share of renewable energies, automation and digitalization of the city management. On the other hand, this requires an increasing amount of electronic equipment, such as EV chargers, digital trash bins, smart stations used in cities, which result in an increasing demand of the city maintenance. As a result, large amount of RCA reports are generated, which record city infrastructure damages, causes and solutions. Therefore, it may be desirable in a smart city to have an efficient approach to process those RCA reports to improve the city maintenance work. Data Source: In this context, a prompt-based summarization system for RCA in according with embodiments of the present application could be suitably applied to RCA reports as text, which describe the city infrastructure issues and the problem- solving process. Such RCA reports could be derived, e.g., from sensor networks/systems of or within the smart city. Prompt-based RCA summarization system according to the present disclosure: According to an embodiment of the present disclosure, the system 100 may be configured to summarize those reports in terms of problems, causes and solutions. The length of the summary may be adapted, e.g., to a stress-level of a worker who operates the system 100 (as they might have a tight schedule); hence, the most urgent problems are addressed first. Output: For each input report, the output may include: summaries of all problems in an ordered list; for each problem, summaries of causes in an ordered list; and, for each problem, summaries of solutions in an ordered list. Physical Changes (Technicity): After a city maintenance worker activated the system according to embodiments of the present disclosure, e.g., by pressing/touching a given sensing area, or looking into a camera of a tablet device, the system may display the generated summary. Parameters for controlling the summary length may be computed based, e.g., on the worker’s touch force or look in their eyes. In this way, the generated summary is optimized in length, hence more effective for the worker to read. This helps the worker to manage the information load when having a tight schedule. Use Case 2: Emission Reduction in Smart Cities: Waste-To-Energy Use Case: Waste-to-energy (WtE) can be a solution to reduce emissions as burning waste instead of oil or gas reduces the overall emissions; however, burning waste in an improper way actually results in a worse climate balance. In this context, city- level RCA reports may provide crucial insides from past experience about how to design an efficient waste separation schema and how to adopt proper WtE generation technology to reach a better climate / emission balance. Data Source: In this context, a prompt-based summarization system for RCA in according with embodiments of the present application could be suitably applied to RCA reports as text, which describe different problem-solving cases such as the waste separation schema and WtE generation technology adaptation. Prompt-based RCA summarization system according to the present disclosure: According to an embodiment of the present disclosure, the system 100 may be configured to summarize those reports in terms of problems, causes and solutions. The length of the summary may be adapted to, e.g., a stress-level of a worker who operates the system 100 (as they might have a tight schedule); hence, the most urgent problems can be addressed first. Output: For each input report, the output may include: summaries of all problems in an ordered list; for each problem, summaries of causes in an ordered list; and, for each problem, summaries of solutions in an ordered list. Physical Changes (Technicity): The summary, for which the parameters controlling the RCA summary length are computed based on physical properties of the human user (e.g., touch force of the fingers on a device screen, or the look in one’s eyes) in order to achieve an effective reading, may be shown to the user on a screen or display. Use Case 3: Digitalized Root-Cause Analysis for Continuous Quality Improvement of Green Manufacturing Use Case: Green manufacturing processes increasingly use robots to achieve sustainable manufacturing. For instance, robotic vehicles may be used instead of gas-powered trucks to transport parts, robots may be used to sort tedious material for recycling, or robots may be used to clean tanks filled with toxic chemicals. However, the increasing number of robots results in a larger demand of robotic maintenance work. Consequently, large amount of RCA reports are generated in this context, which record the robotic issues, causes and solutions. Maintainer need an efficient approach to process those reports to maintain and improve the robotics. Data Source: In this context, a prompt-based summarization system for RCA in according with embodiments of the present application could be suitably applied to RCA reports as text, which record the robotic issues, cause and solutions. Prompt-based RCA summarization system according to the present disclosure: According to an embodiment of the present disclosure, the system 100 may be configured to summarize those reports in terms of problems, causes and solutions. The length of the summary may be adapted to, e.g., a stress-level of a worker who operates the system 100 (as they might have a tight schedule); hence, the most urgent problems are addressed first. Output: For each input report, the output may include: summaries of all problems in an ordered list; for each problem, summaries of causes in an ordered list; and, for each problem, summaries of solutions in an ordered list. Physical Changes (Technicity): After a robot engineer activated the system 100 according to embodiments of the present disclosure, e.g., by pressing/touching a sensing area, or looking into a camera of a tablet device, the system 100 may display the generated summary. The parameters for controlling the summary length may be computed, e.g., based on a worker’s touch force or look in their eyes. In this way, the generated summary has an optimized length, hence more effective for this engineer to read. This helps the engineer to manage the information load when having a tight schedule. Many modifications and other embodiments of the invention set forth herein will come to mind to the one skilled in the art to which the invention pertains having the benefit of the teachings presented in the foregoing description and the associated drawings. Therefore, it is to be understood that the invention is not to be limited to the specific embodiments disclosed and that modifications and other embodiments are intended to be included within the scope of the appended claims. Although specific terms are employed herein, they are used in a generic and descriptive sense only and not for purposes of limitation.