Semantic temporal visualization of information

Abstract

A computer-implemented method for generating digital information data in a subject area is presented. The method comprises: - providing, at a processing unit (110), digital information corpus data; - extracting, via the processing unit (110), digital information seed data from the digital information corpus data; - performing, via the processing unit (110), a search in at least one database (120) comprising knowledge information, thereby extracting a plurality of text blocks related to the subject area from the at least one database (120); wherein the search is performed based on the digital information seed data, - indexing, via the processing unit (110), the text blocks chronologically; - generating, via the processing unit (110), the digital information data using the chronologically organized text blocks.

Description

TECHNICAL FIELD

[0001] The present invention relates to a computer-implemented method for generating digital information data in a subject area. Moreover, the present invention relates to a computer system for generating digital information data in a subject area. The method and computer system can be used in the innovation chain from research and development to product launch, such as in the field of chemical technology. Other applications are possible. BACKGROUND

[0002] Digitization initiatives in many technical fields are increasingly recognizing the user need to automatically establish causal dependencies in a set of documents accessed via search engines or web drives in order to quickly direct the user's attention to the most important facts across multiple documents. The ranking of search results is not designed for this task at all. Documents processed by semantic information extraction and represented as semantic networks in a knowledge base can do this. However, knowledge base establishment is slow and expensive. More advanced approaches to "logical understanding" of software agents are still in different stages of AI research, so there is an opportunity for pragmatic approximation of causal dependencies in inexpensive technical implementations.

[0003] US 2016 / 0188642 Al discloses a computer-implemented method for combining a primary document with one or more candidate documents. The method comprises extracting process steps disclosed in the primary document and extracting candidate process steps disclosed in the one or more candidate documents; constructing a primary data structure corresponding to the primary document; wherein the primary data structure comprises interconnected nodes and each node corresponds to an extracted process step disclosed in the primary document; identifying one or more candidate processes to combine with the primary data structure; and inserting the one or more identified candidate process steps into the primary data structure.

[0004] US 2016 / 0162486 Al discloses a computer-supported method of assisting in generating an innovation. The method comprises the steps of: retrieving a first set of more than two documents belonging to a first domain from a database; retrieving a second set of more than two documents belonging to a second domain from the database; selecting all possible combinations of documents from the first set with all documents of the second set and, for each combination of documents: determining a composite novelty score, a composite proximity score and a composite impact score; and providing a recommendation that can assist in generating an innovation based on all determined composite novelty scores and / or composite proximity scores and / or composite impact scores.

[0005] US 9,799,040 B2 discloses a method of computer-aided innovation. The method provides a method by which suggested innovation opportunities can be automatically generated, which can then be viewed or otherwise communicated to a user and analyzed by the user. The disclosure provides a method and apparatus for determining an innovation opportunity by selecting one or more terms; determining trend data related to the selected elements; determining an innovation likelihood measure for the selected elements in dependence on the trend data; identifying an innovation opportunity in dependence on the innovation likelihood measure.

[0006] Despite the achievements so far, there is still a need for enhancing information visualization and knowledge management, in particular along the innovation chain from research and development to product launch.

[0007] Problem to be solved

[0008] It is therefore desirable to provide methods and apparatuses that solve the technical challenges mentioned above. In particular, it should be provided for an apparatus and a method for generating digital information data in a subject area via at least one processing unit that allows for enhancing information visualization and knowledge management. SUMMARY

[0009] The problem is solved by a computer-implemented method for generating digital information data in a subject area and a computer system having the features of the independent claims. Advantageous embodiments, which can be realized in isolation or in any arbitrary combination, are listed in the dependent claims.

[0010] In a first aspect of the present invention, a computer-implemented method for generating digital information data in a subject area is presented.

[0011] As used herein, the term "computer-implemented" is a broad term and shall be given its ordinary and customary meaning to a person of ordinary skill in the art and shall not be limited to a special or customized meaning. The term specifically can refer, without limitation, to a process that is fully or partially implemented by using a data processing apparatus, such as a data processing apparatus comprising at least one processing unit. Thus, the term "computer" generally can refer to a combination or network of apparatuses having at least one data processing apparatus, such as at least one processing unit. Further, a computer can comprise one or more further components, such as at least one of a data storage device, an electronic interface, or a human-machine interface.

[0012] As used herein, the term "processing unit" is a broad term and should be given its common and conventional meaning to those skilled in the art, without limitation its specific or customary meaning. Specifically, the term may refer to, but is not limited to, any logic circuit configured to perform basic operations of a computer or system, and / or generally refers to a device configured to perform computational or logical operations. In particular, a processing unit may be configured to process basic instructions that drive a computer or system. As an example, a processing unit may include at least one arithmetic logic unit (ALU), at least one floating-point unit (FPU), such as a math coprocessor or a digital coprocessor, multiple registers specifically configured to provide operands to the ALU and store the results of operations, and memory, such as L1 and L2 caches. In particular, a processing unit may be a multi-core processor. Specifically, a processing unit may be or may include a central processing unit (CPU). Additionally or alternatively, a processing unit may be or may include a microprocessor; therefore, specifically, the elements of a processing unit may be contained within a single integrated circuit (IC) chip. Additionally or alternatively, the processing unit may be or may include one or more application-specific integrated circuits (ASICs) and / or one or more field-programmable gate arrays (FPGAs).

[0013] As used herein, the term "database" is a broad term and should be given its common and conventional meaning to those skilled in the art, without limitation its specific or customary meaning. Specifically, the term may refer to, but is not limited to, any collection of information and / or a physical structure configured to store any collection of information. A database may include at least one storage device configured to store information. A database may be or may include at least one element selected from the group consisting of at least one server, at least one server system comprising multiple servers, at least one cloud server, or cloud computing infrastructure. Methods may be performed using multiple databases, such as at least one document store and at least one knowledge base, as will be outlined in detail below. Methods may be performed using a single database configured to implement multiple functions such as data storage and knowledge storage. For example, a document store may be integrated into a knowledge base or may be an external device.

[0014] As used herein, the term "storage device" is a broad term and should be given its common and conventional meaning to those skilled in the art, without limitation any particular or customary meaning. Specifically, the term may refer to, but is not limited to, the process of recording and / or retraining data.

[0015] As used herein, the term "subject area" is a broad term and should be given its common and conventional meaning to those skilled in the art, without limitation its specific or customary meaning. Specifically, the term may refer to, but is not limited to, branches of knowledge such as medicine, chemistry, and physics.

[0016] As used herein, the term "digital information data" is a broad term and should be given its common and conventional meaning to those skilled in the art, without limitation any specific or customary meaning. Specifically, the term may refer to, but is not limited to, discrete, discontinuous representations of any textual information. Digital information data may include one or more of the following: scientific documents, research-related documents, development-related documents, business-related documents, company-related documents, legal documents, patent documents, regulatory documents, user manuals, instruction manuals, training materials, etc.

[0017] The computer-implemented method includes the following steps, which can be executed in a given order. However, different orders are also possible. Furthermore, one or more steps, or even all steps, can be executed once or repeatedly. Further, the method steps can be executed in an overlapping manner or even in parallel. The method may also include additional method steps not listed.

[0018] The method includes the following steps:

[0019] - Provide digital information corpus data at the processing unit;

[0020] - Extract digital information seed data from digital information corpus data via the processing unit;

[0021] - A search is performed via a processing unit in at least one database containing knowledge information, thereby extracting multiple text blocks related to a topic region from the at least one database; wherein the search is performed based on seed data of digital information.

[0022] - The text blocks are indexed in chronological order by the processing unit;

[0023] - Digital information data is generated using time-organized text blocks via the processing unit.

[0024] As used herein, the term "providing" digital information corpus data is a broad term and should be given its common and conventional meaning to those skilled in the art, and not limited to its specific or customary meaning. Specifically, the term may refer to, but is not limited to, inputting, storing, and / or uploading digital information corpus data.

[0025] Digital information corpus data can be any type of digital information data. For example, digital information corpus data can include complete digital information data, such as a complete document, such as a comment or notice, or a digital information corpus can include at least a portion of digital information data, such as at least one sentence.

[0026] The term "seed data" is a broad term and should be given its common and conventional meaning to those skilled in the art, without limitation any specific or customary meaning. The term can refer to, but is not limited to, data that has already populated the database at the time of its creation. Data seeding is used to provide initial values ​​for lookup lists, for demonstration purposes, proof of concept, etc.

[0027] As used herein, the term "extracting" seed data is a broad term and should be given its common and conventional meaning to those skilled in the art, without limitation any specific or customary meaning. Specifically, the term may refer to, but is not limited to, the digital extraction of data from a given corpus.

[0028] As outlined above, the method includes: performing at least one search via a processing unit in at least one database including knowledge information to extract multiple text blocks related to a topic region from the at least one database; wherein the search is performed based on seed data of digital information. Specifically, the search may be a semantic search performed in the database. As used herein, the term "semantic search" is a broad term and should be given its common and conventional meaning to those skilled in the art, and is not limited to a particular or customized meaning. The term may specifically refer to, but is not limited to, a search considering at least one meaning of the search term. Semantic search may be performed using at least one machine learning tool such as a neural network. Semantic search may include performing document search queries based on seed data.

[0029] As used herein, the term "syntactic search" is a broad term and should be given its common and conventional meaning to those skilled in the art, without limitation its specific or customary meaning. Specifically, the term may refer to, but is not limited to, searching a database for textual matches to search terms. Similarly, as used herein, the term "semantic search" is a broad term and should be given its common and conventional meaning to those skilled in the art, without limitation its specific or customary meaning. Specifically, the term may refer to, but is not limited to, a search considering at least one meaning of search terms. Syntactic and / or semantic searches can be performed using at least one machine learning tool, such as a neural network.

[0030] Semantic search may include performing document search queries based on portions of digital information data. Processing units may be configured to automatically or through user selection identify information within portions of digital information data and perform document searches on that information. Processing units may be configured to identify and resolve ambiguities and / or errors in information provided by the user and perform document searches on that information. For example, processing units may be configured to suggest synonyms, terms, expressions, vocabulary, numbers, formulas, sentences, or addresses, which may be displayed by a user interface for user selection and / or approval. Portions of digital information data may be compared grammatically and / or semantically with digital information data stored in a database. Document searches may include determining grammatical and / or semantic overlap between portions of digital information data and entries stored in the document database. A grammatical and / or semantic search index may be provided by the processing unit. The grammatical and / or semantic search index may include a list of all search results. Presentation of search results may allow users to view content already existing in the database. Furthermore, presentation of search results may allow users to view at least one context in which search terms derived from portions of digital information data they have already entered are currently stored in the database.

[0031] As used herein, the term "document" is a broad term and should be given its common and conventional meaning to those skilled in the art, without limitation any particular or customary meaning. Specifically, the term can refer to, but is not limited to, any numerical representation of a thought. Furthermore, the term "document" can refer to a class of objects including written text and / or at least one drawing. Documents can be scientific documents, research-related documents, development-related documents, business-related documents, company-related documents, legal documents, patent documents, regulatory documents, user manuals, instruction manuals, training materials, etc. Documents can be or may include at least one report, at least one comment, at least one note, at least one scientific paper, at least one figure, at least one user manual, at least one instruction, at least one website, etc. Documents can also be customer feedback related to the product's production process.

[0032] As outlined above, the method includes indexing text blocks chronologically via a processing unit. As used herein, the term "indexing text blocks chronologically" is a broad term and should be given its common and conventional meaning to those skilled in the art, without limitation its specific or customary meaning. Specifically, the term may refer to, but is not limited to, a temporal mapping of text blocks. As used herein, the term "text block" is a broad term and should be given its common and conventional meaning to those skilled in the art, without limitation its specific or customary meaning. Specifically, the term may refer to, but is not limited to, paragraphs of text, such as paragraphs in a text document.

[0033] As outlined above, the method includes generating digital information data using time-organized text blocks via a processing unit. As used herein, the term "generating digital information data using time-organized text blocks" is a broad term and should be given its common and conventional meaning to those skilled in the art, without limitation its specific or customary meaning. Specifically, the term may refer to, but is not limited to, creating digital information data based on time-organized text blocks. The applicant has discovered that the temporal index of data elements is proportional to causal relationships. Therefore, mapping elements in time space can indicate causal relationships of topics within a subject area. This method works best when the text blocks are selected from reasonably defined knowledge sources, or by limiting them to a specific domain.

[0034] This allows for the detection of causal relationships that would otherwise be very difficult to capture. For example, collecting customer feedback has become increasingly common recently. Customer feedback may be an indication of problems that were not detected during the production process. Some, but not all, customer feedback may be an indication of problems that were not detected during the production process. Furthermore, customer feedback often suffers from a lack of standardized format and expression. Assessing whether a customer complaint truly points to an error in production or indicates dissatisfaction from a single customer is very difficult. The use of time-organized text blocks according to the invention allows for the identification of when errors in the production process might occur. This may then trigger an investigation into the root cause. Therefore, the time index is not merely another parameter to be tracked, but can include additional information related to the production process. Thus, the inventive method can allow the detection of hidden patterns and causal relationships.

[0035] The processing unit may be operatively coupled to at least one database. As used herein, the term "operatively coupled" is a broad term and should be given its common and customary meaning to those skilled in the art, without limitation any specific or customary meaning. Specifically, the term may refer to, but is not limited to, a communication connection between the processing unit and at least one database for one or more of the following: transmitting information, accessing storage, or controlling at least one function of another device. The processing unit and the database may include at least one communication interface through which the processing unit and the database are operatively coupled. The processing unit may be configured to access (e.g., read and write) storage in the database via the communication interface. As used herein, the term "communication interface" is a broad term and should be given its common and customary meaning to those skilled in the art, without limitation any specific or customary meaning. Specifically, the term may refer to, but is not limited to, items or elements forming the boundaries configured for transmitting information. In particular, the communication interface may be configured to transmit information from a computing device (e.g., a computer), such as sending or outputting information, for example, to another device. Additionally or alternatively, the communication interface may be configured to transmit information to a computing device, such as to a computer, for example, to receive information. A communication interface may specifically provide means for transmitting or exchanging information. In particular, a communication interface may provide data transmission connections, such as Bluetooth, NFC, inductive coupling, etc. As an example, a communication interface may be or may include at least one port, including one or more of a network or internet port, a USB port, and a disk drive. A communication interface may be at least one network interface.

[0036] Extracting seed data for digital information can include semantic information extraction. In other words, information can be extracted based on the semantic relationships between the seed data.

[0037] The method may further include filtering the extracted seed data of digital information by the processing unit through process attributes. As used herein, the term "process attribute" is a broad term and should be given its common and conventional meaning to those skilled in the art, and not limited to a specific or custom meaning. The term may specifically refer to, but is not limited to, a type of process data variable that specifically relates to the operation of a process, such as a task ID or participant. Many process attributes are available out of the box, but they can also be created manually. By using process attributes for filtering, such as IPC class or project ID, irrelevant topics can be filtered out.

[0038] Extracting multiple text blocks can include selecting portions of the knowledge information from the database to break it down into text blocks. Therefore, text blocks can be created by separating the text into a certain number of text blocks.

[0039] The method may further include: recursively calculating the semantic similarity between the extracted text blocks by the processing unit. Therefore, text blocks can be provided in order of relevance relative to the search query.

[0040] The method may further include: selecting a predetermined number of previous text blocks for each of the indexed text blocks with predetermined timestamps, and identifying a list of candidate concepts in a database for each concept in the text blocks with predetermined timestamps by performing embedding clustering on concept embeddings in all the previous text blocks.

[0041] As used herein, the term "embedding" is a broad term and should be given its common and conventional meaning to those skilled in the art, without limitation its specific or tailored meaning. Specifically, the term can refer to, but is not limited to, a collective term for a set of language modeling and feature learning techniques in Natural Language Processing (NLP), in which words or phrases from a vocabulary are mapped to vectors of real numbers. Conceptually, it involves mathematical embeddings from a space with many dimensions for each word to a continuous vector space with much lower dimensions. Methods for generating this mapping include neural networks, dimensionality reduction of word co-occurrence matrices, probabilistic models, interpretable knowledge base methods, and explicit representations of the context in which words occur. Word and phrase embeddings have been shown to improve the performance of NLP tasks such as parsing and sentiment analysis when used as underlying input representations.

[0042] A database may include at least one knowledge base comprising multiple concepts. As used herein, the term "knowledge base" is a broad term and should be given its common and conventional meaning to those skilled in the art, without limitation a specific or customary meaning. Specifically, the term may refer to, but is not limited to, an ontology comprising at least one hierarchy of classes, subclasses, and instances. Classes are represented herein as concepts. Concepts may be physical and / or chemical concepts, scientific concepts, technical terms, etc. A knowledge base may include a unique identifier for each entry in the document storage. In addition to a unique ID, a knowledge base may include multiple metadata strings. As used herein, the term "metadata string" is a broad term and should be given its common and conventional meaning to those skilled in the art, without limitation a specific or customary meaning. Specifically, the term may refer to, but is not limited to, data that provides information about other data. Specifically, a metadata string may serve as a pointer to at least one other object, which in turn may have at least one additional pointer. Each concept in the knowledge base may be represented by a metadata string. Each concept may be linked to at least one entry in the document storage. Metadata strings may include connecting entries about insights such as documents or document storage, as well as information about connections to other concepts such as higher-level concepts and / or sub-concepts. Since the knowledge base includes a unique identifier for each entry in document storage, the processing unit can determine and provide the corresponding metadata string for entries used for grammatical and / or semantic search indexing. Metadata strings provided in response to at least one grammatical and / or semantic search may include information about at least one concept.

[0043] The method may also include: applying a learning-to-rank model trained on existing digital information corpus data at the processing unit using features that evaluate graph relationships between candidate concepts and assess semantic similarity between text blocks with predetermined timestamps and all previous text blocks.

[0044] As used herein, the term "ranking learning model" is a broad term and should be given its common and conventional meaning to those skilled in the art, without limitation its specific or tailored meaning. Specifically, the term can refer to, but is not limited to, the application of machine learning, typically supervised, semi-supervised, or reinforcement learning, in the construction of ranking models for information retrieval systems. Training data consists of lists of items, with some partial order specified between items in each list. This order is typically induced by assigning a numerical or ordinal score or a binary judgment (e.g., "relevant" or "irrelevant") to each item. The purpose of a ranking model is to rank, i.e., to produce permutations of items in a new, unseen list in a manner similar to the ranking in the training data. Ranking learning is also known as Machine Learning Ranking (MLR).

[0045] The method may further include annotating text blocks with predetermined timestamps using the top k candidate concepts. Therefore, text blocks with predetermined timestamps are evaluated based on candidate concepts to define a specific relevance order.

[0046] The method may further include concatenating a text block with a predetermined timestamp with the top k ranked text blocks of the preceding text block, and labeling them with scores from a ranking learning model. Therefore, the order of relevance of text blocks with predetermined timestamps is defined as having the most relevant concepts at the top.

[0047] The method may also include repeating the following steps: selecting previous text blocks, identifying a list of candidate concepts, applying a ranking learning model, and annotating text blocks with predetermined timestamps, until all text blocks are clustered. Thus, ranking and sorting based on relevance continues until all text blocks are processed to reveal the best quality of potential relevance.

[0048] The method may further include, at the processing unit, transmitting text blocks, specifically writing them to a semantic graph as nodes labeled with predetermined time bins. Thus, semantic relationships are visualized in a predetermined order established using the previously explained method steps.

[0049] The method may further include: forming connections between text blocks at the processing unit, specifically writing them into a semantic graph as traces, particularly as directed edges. Thus, the semantic relationships between text blocks are visualized.

[0050] Generating digital information data can include creating visualizations that indicate the temporal and semantic distances between text blocks. Therefore, temporal and semantic information across text blocks can be readily derived.

[0051] The visualization is an interactive 2D tree visualization, where text block nodes are symbols and traces, particularly edges, are arrows, ordered by time index. By visually tracing text blocks from a document through temporal and semantic space, the semantic temporal tree allows for approximation of the flow of causal reasoning within a document set. Particular attention can be drawn to significant clustering and early truncation of branches, which may respectively indicate over- or under-researched topics. Furthermore, unexpected combinations of terms inspire new analytical directions.

[0052] The distance in the x-direction indicates the temporal distance of the time index step, and the distance in the y-direction indicates the score of the ranking learning model relative to the text block in the previous time index. Therefore, the clear arrangement of temporal and semantic information across text blocks can be readily derived.

[0053] Visualization can easily identify the temporal evolution of semantics. This is particularly useful when dealing with complex issues such as user complaints, as these issues may indicate errors in the production process. 2-D visualization makes it very easy to spot the first occurrence of a chain of semantic similarities. In particular, when used to link customer feedback to errors in production, it is crucial not only to track semantic similarities—customers may use different terms—but also to visualize the time series. A single occurrence of customer feedback on a specific topic may not be relevant; however, if this is followed by various blocks of text with similar semantics, it could be a trigger for an error in the production process that occurred prior to the first customer feedback.

[0054] In another aspect, a computer program for generating digital information data in a subject area is proposed. This computer program includes instructions that, when executed by a computer or computer network, cause the computer or computer network to perform, wholly or partially, the method for generating digital information data in a subject area according to the invention in one or more of the embodiments appended herein. For possible definitions of most terms used herein, reference can be made to the description of the computer-implemented method for generating digital information data in a subject area, which is further described above or below.

[0055] Specifically, computer programs can be stored on computer-readable data carriers and / or computer-readable storage media. As used herein, the terms "computer-readable data carrier" and "computer-readable storage media" can specifically refer to non-transitory data storage devices, such as hardware storage media on which computer-executable instructions are stored. Computer-readable data carriers or storage media can specifically be or can include storage media such as random access memory (RAM) and / or read-only memory (ROM). For example, at least one database, such as a server or cloud server, can be used to store computer programs.

[0056] This document further discloses and proposes a computer program product having program code means so that, when the program is executed on a computer or computer network, it performs the method according to the invention in one or more of the embodiments appended herein. Specifically, the program code means may be stored on a computer-readable data carrier and / or a computer-readable storage medium. As used herein, a computer program product refers to a program that is a tradable product. The product can generally exist in any format, such as in paper format or on a computer-readable data carrier. Specifically, the computer program product may be distributed on a data network.

[0057] This document further discloses and proposes a data carrier having a data structure stored thereon, which, after being loaded into a computer or computer network, such as into the working memory or main memory of the computer or computer network, can perform the method according to the invention in one or more of the embodiments disclosed herein.

[0058] In another aspect, a computer system for generating digital information data in a thematic region is disclosed. The computer system includes at least one database and at least one processing unit. The processing unit is configured to provide digital information corpus data. The processing unit is configured to extract digital information seed data from the digital information corpus data. The processing unit is configured to perform a search in at least one database including knowledge information, thereby extracting multiple text blocks related to the thematic region from the at least one database; wherein the search is performed based on the digital information seed data. The processing unit is configured to index the text blocks in chronological order. The processing unit is configured to generate digital information data using the chronologically organized text blocks.

[0059] At least one processing unit can be operatively coupled to at least one database.

[0060] The proposed method and apparatus allow for enhanced utilization of the inherent consistency and reduced noise levels of document content generated by the workflow, enabling rapid, approximate 2D visualization based on existing information extraction techniques. By intuitively tracking text blocks from documents in both temporal and semantic spaces, semantic temporal trees allow for approximation of the flow of causal reasoning within a document set. Particular attention can be paid to significant clustering and early truncation of branches, which may respectively indicate over- or under-researched topics. Furthermore, unexpected combinations of terms inspire new analytical directions.

[0061] The proposed methods and computer systems allow for enhanced information retrieval and knowledge management through insight capture. Particularly along the innovation chain, from research and development to product launches and customer service, the impact of insight capture could allow for reduced time-to-market and potentially faster problem resolution in response to customer requests. Insights built upon existing insights could allow for triggering new levels of organization-wide learning, which could enhance the effectiveness and impact of new ideas created by users.

[0062] As used herein, the terms “have,” “include,” or “contain,” or any grammatical variations thereof, are used in a non-exclusive manner. Thus, these terms can refer to a situation where an entity described herein has no other features besides those introduced by these terms, or to a situation where one or more other features exist. For example, the statements “A has B,” “A includes B,” and “A contains B” can refer to a situation where A has no other elements besides B (i.e., A consists solely and exclusively of B), or to a situation where entity A has one or more other elements besides B (e.g., element C, elements C and D, or even other elements).

[0063] Furthermore, it should be noted that the terms "at least one," "one or more," or similar expressions indicating that a feature or element may exist once or more often are generally used only once when the corresponding feature or element is introduced. In most cases, the expressions "at least one" or "one or more" are not repeated when referring to the corresponding feature or element, despite the fact that the corresponding feature or element may exist once or more often.

[0064] Furthermore, as used herein, the terms “preferredly,” “more preferably,” “particularly,” “more particularly,” “specifically,” “more specifically,” or similar terms are used in combination with optional features without limiting the possibility of substitution. Therefore, features introduced by these terms are optional features and are not intended to limit the scope of the claims in any way. As those skilled in the art will recognize, the invention can be practiced by using alternative features. Similarly, features introduced by phrases such as “in embodiments of the invention” are intended to be optional features, without limiting alternative embodiments of the invention, without limiting the scope of the invention, and without limiting the possibility of combining features introduced in this manner with other optional or non-optional features of the invention.

[0065] In summary, and without excluding further possible embodiments, the following embodiments are conceivable:

[0066] Example 1. A computer-implemented method for generating digital information data in a subject area, the method comprising:

[0067] - At the processing unit, digital information corpus data is provided;

[0068] - Extract digital information seed data from the digital information corpus data via the processing unit;

[0069] - The processing unit performs a search in at least one database containing knowledge information to extract multiple text blocks related to the topic region from the at least one database; wherein the search is performed based on the digital information seed data.

[0070] - The text blocks are indexed in chronological order via the processing unit;

[0071] - The digital information data is generated using time-organized text blocks via the processing unit.

[0072] Example 2. According to the method described in the foregoing examples, extracting the digital information seed data includes: semantic information extraction.

[0073] Example 3. The method according to any of the foregoing embodiments further includes: the processing unit filtering the extracted digital information seed data by means of process attributes.

[0074] Example 4. The method according to any of the foregoing embodiments, wherein extracting the plurality of text blocks includes: selecting portions for decomposing the knowledge information from the database into text blocks.

[0075] Example 5. The method according to any of the foregoing embodiments further includes: recursively calculating the semantic similarity between the extracted text blocks by the processing unit.

[0076] Example 6. The method according to any of the foregoing embodiments further includes: for each text block having a predetermined timestamp in the indexed text blocks, selecting a predetermined number of previous text blocks, and clustering the embeddings by concept embeddings in all said previous text blocks, and identifying a candidate concept list in the database for each concept in the text blocks having the predetermined timestamps.

[0077] Example 7. The method according to the foregoing embodiments further includes: applying a ranking learning model trained on existing digital information corpus data at the processing unit using features that evaluate graph relationships between candidate concepts and evaluate semantic similarity between the text block with the predetermined timestamp and all the previous text blocks.

[0078] Example 8. The method according to the foregoing embodiments further includes: annotating the text block having the predetermined timestamp with the top k candidate concepts.

[0079] Example 9. The method according to the foregoing embodiments further includes: concatenating the text block having the predetermined timestamp with the top k ranked text blocks in the previous text blocks, and labeling them with the score of the ranking learning model.

[0080] Example 10. The method according to the foregoing embodiments further includes: repeating the following steps until all text blocks are clustered: selecting the previous text blocks, identifying the candidate concept list, applying the ranking learning model, and annotating the text blocks having the predetermined timestamp.

[0081] Example 11. The method according to any of the foregoing embodiments further includes: at the processing unit, transmitting, in particular writing, the text block to a semantic graph as a node marked with a predetermined time bin.

[0082] Example 12. The method according to the foregoing embodiments further includes: at the processing unit, forming connections between the text blocks, particularly writing them into the semantic graph as traces, particularly as directed edges.

[0083] Example 13. The method according to any one of Examples 6 to 12, wherein generating the digital information data includes: generating a visualization indicating the temporal distance and semantic distance of the text block.

[0084] Example 14. The method according to the foregoing embodiments, wherein the visualization is an interactive 2D tree visualization, wherein text block nodes are sorted by time index as symbols and traces, particularly edges, as arrows.

[0085] Example 15. The method according to the foregoing embodiments, wherein the distance in the x-direction indicates the time distance of the time index step, and the distance in the y-direction indicates the score of the ranking learning model relative to the text block in the previous time index.

[0086] Example 16. A computer program comprising computer-executable instructions for performing the method according to any of the foregoing embodiments.

[0087] Example 17. A computer-readable storage medium having stored thereon computer-executable instructions for implementing the method according to any one of Examples 1 to 15.

[0088] Example 18. A computer system for generating digital information data in a subject area, comprising:

[0089] The system includes at least one database and at least one processing unit, wherein the processing unit is configured to provide digital information corpus data, wherein the processing unit is configured to extract digital information seed data from the digital information corpus data, wherein the processing unit is configured to perform a search in the at least one database including knowledge information to extract multiple text blocks related to the topic region from the at least one database; wherein the search is performed based on the digital information seed data, wherein the processing unit is configured to index the text blocks in chronological order, and wherein the processing unit is configured to generate the digital information data using the chronologically organized text blocks.

[0090] Example 19. A computer system according to the foregoing embodiments, wherein the at least one processing unit is operatively coupled to the at least one database.

[0091] Example 20. A computer system according to any of the foregoing embodiments relating to a computer system, wherein the computer system is configured to execute, via the at least one processing unit, a method for generating digital information data in a subject area, according to any of the foregoing embodiments relating to a method for generating digital information data in a subject area. Attached Figure Description

[0092] Further optional features and embodiments will be disclosed in more detail in the following description of embodiments, preferably in conjunction with the dependent claims. As those skilled in the art will recognize, the corresponding optional features can be implemented in isolation and in any feasible combination. The scope of the invention is not limited to the preferred embodiments. Embodiments are schematically depicted in the accompanying drawings. In these drawings, the same reference numerals refer to the same or functionally comparable elements.

[0093] In the attached diagram:

[0094] Figure 1 A flowchart of a computer-implemented method for generating digital information data in a subject area according to the present invention is shown;

[0095] Figure 2 A visualization indicating temporal and semantic distance is shown, given a set of user-selected concepts;

[0096] Figure 3 A visualization indicating the temporal and semantic distances applied to the production process is shown; and

[0097] Figure 4 A system according to the present invention is shown. Detailed Implementation

[0098] Figure 1 A flowchart illustrating a computer-implemented method for generating digital information data in a subject area according to the present invention is shown. According to the invention, the method can be executed by a computer system 100 via at least one processing unit 110. The processing unit 110 may be operatively coupled to at least one database 120.

[0099] Processing unit 110 may be or may include any logic circuit configured to perform basic operations of a computer or system, and / or a device typically configured to perform computational or logical operations. Specifically, processing unit 110 may be configured to process basic instructions that drive a computer or system. As an example, processing unit 110 may include at least one arithmetic logic unit (ALU), at least one floating-point unit (FPU), such as a math coprocessor or a digital coprocessor, multiple registers specifically configured to provide operands to the ALU and store the results of operations, and memory, such as L1 and L2 caches. Specifically, processing unit 110 may be a multi-core processor. Specifically, processing unit 110 may be or may include a central processing unit (CPU). Additionally or alternatively, processing unit 110 may be or may include a microprocessor; therefore, specifically, the elements of the processing unit may be contained in a single integrated circuit (IC) chip. Additionally or alternatively, processing unit 110 may be or may include one or more application-specific integrated circuits (ASICs) and / or one or more field-programmable gate arrays (FPGAs), etc.

[0100] Database 120 may be or may include any set of information and / or a physical structure configured to store any set of information. Database 120 may include at least one storage device configured to store information. Database 120 may be or may include at least one element selected from the group consisting of at least one server, at least one server system including multiple servers, at least one cloud server, or cloud computing infrastructure. The method may be performed using multiple databases 120. Database 120 may include additional sub-units, such as at least one document storage 140, and may additionally or alternatively include at least one knowledge base 160. The method may be performed using a single database 120 configured to implement multiple functions such as data storage and knowledge storage. For example, document storage 140 may be integrated into knowledge base 160 or may be an external device.

[0101] Digital information data can be a discrete, discontinuous representation of any textual information. Digital information data may include one or more of the following: scientific documents, research-related documents, development-related documents, business-related documents, company-related documents, legal documents, patent documents, regulatory documents, user manuals, instruction manuals, training materials, etc.

[0102] Processing unit 110 is operatively coupled to at least one database 120. Specifically, a communication connection exists between processing unit 110 and at least one database 120 for one or more of the following functions: transmitting information, accessing memory, or controlling at least one other device. Processing unit 110 and database 120 may include at least one communication interface through which processing unit 110 and database 120 are operatively coupled. Processing unit 110 may be configured to access (e.g., read and write) storage in the database via the communication interface. The communication interface may be or may include items or elements forming the boundaries configured for transmitting information. In particular, the communication interface may be configured to transmit information from a computing device (e.g., a computer), such as sending or outputting information, for example, to another device. Additionally or alternatively, the communication interface may be configured to transmit information to a computing device, such as to a computer, such as to receive information. The communication interface may specifically provide means for transmitting or exchanging information. In particular, the communication interface may provide data transmission connections, such as Bluetooth, NFC, inductive coupling, etc. As an example, the communication interface may be or may include at least one port, including one or more of a network or internet port, a USB port, and a disk drive. The communication interface may be at least one network interface.

[0103] As by Figure 1As shown in the flowchart, the method begins with step S10, where processing unit 110 is provided. In subsequent step S12, digital information corpus data is provided at processing unit 110. Specifically, in step S12, word and document embeddings for concepts are calculated once over the entire digital information corpus data. Thus, digital information seed data is extracted from the digital information corpus data. In subsequent step S14, a search is performed in at least one database 120, including knowledge information, to extract multiple text blocks related to topic regions from at least one database. The search is performed based on the digital information seed data. For example, a user queries a semantic search engine on the annotated digital corpus data extracted through semantic information. In subsequent step S16, the extracted digital information seed data is filtered by processing unit 110 by process attributes. For example, the user filters the extracted digital information seed data by process attributes (such as IPC class or item ID). The user can be a human user. In subsequent step S18, section headings are extracted from the documents thus found. As shown in subsequent step S20, extracting multiple text blocks includes selecting portions for decomposing knowledge information from the database into text blocks. For example, the user selects portions for decomposing a document into text blocks. Typical examples for portions in a patent are "claims," ​​"background art," and "specification," while for scientific papers they are "introduction," "method," and "conclusion." In subsequent step S22, the processing unit uses the top k result documents from the semantic search to calculate word and document embeddings for concepts in the text blocks.

[0104] In subsequent step S24, the text blocks are indexed chronologically via processing unit 110. In subsequent step S26, this starts from the most recent timestamp. In subsequent step S28, for each of the indexed text blocks with a predetermined timestamp, a predetermined number of previous text blocks are selected. For example, for each text block i,j with timestamp j, text block m,j-1 is selected. In subsequent step S30, for each concept in the text block with a predetermined timestamp, a list of candidate concepts is identified in database 120 by clustering the concept embeddings in all previous text blocks. For example, for each concept in text block i,j, a list of candidate concepts is identified in a database such as a knowledge base by clustering the concept embeddings in all text blocks m,j-1. In subsequent step S32, a ranking learning model trained on existing digital information corpus data is applied at processing unit 110 using features that evaluate graph relationships between candidate concepts and assess semantic similarity between text blocks with predetermined timestamps and all previous text blocks. For example, a ranking learning model trained on existing digital corpus data is applied using features that evaluate graph relationships between candidate concepts and assess semantic similarity between text block i,j and all text blocks m,j-1. In subsequent step S34, text blocks with predetermined timestamps are annotated with the top k candidate concepts. For example, text block i,j is annotated with the top k candidate concepts. In subsequent step S36, text blocks with predetermined timestamps are connected to the top k text blocks of the previous text block and labeled with the score of the ranking learning model. For example, text block i,j is connected to the top k text blocks m,j-1 and the edges are labeled with the score of the ranking learning model. In subsequent step S38, the following steps are repeated: selecting previous text blocks, identifying the candidate concept list, applying the ranking learning model, and annotating text blocks with predetermined timestamps until all text blocks are clustered. In other words, steps S28 to S36 are repeated until all text blocks are clustered.

[0105] In subsequent step S40, at processing unit 110, text blocks are transmitted (e.g., written) to the semantic graph as nodes labeled with predetermined time bins. For example, text blocks are written to the semantic graph as nodes labeled with time bin i. In subsequent step S42, at processing unit 110, connections between text blocks are formed (e.g., written) to the semantic graph as traces, such as directed edges. For example, connections between text blocks are written to the semantic graph as directed edges. As shown in subsequent step S44, generating digital information data includes generating visualizations indicating the temporal and semantic distances of text blocks. The visualization is an interactive 2D tree visualization, where text block nodes are sorted by time index as symbols and traces, particularly edges, as arrows. The distance in the x-direction indicates the temporal distance of the time index step, and the distance in the y-direction indicates the score of the sorting learning model relative to the text blocks in the previous time index. Specifically, an interactive 2D tree visualization is generated, with text block nodes as symbols and edges as arrows, sorted from left to right by time index. This visualization displays a list of given concepts selected by the user from at least one selected text block, text blocks annotated with at least one selected concept, a distance in the x-direction indicating the temporal distance of the time index step, and a distance in the y-direction indicating the score of the ranking learning model relative to text blocks in the previous time index. The method terminates in subsequent step S46.

[0106] Figure 2 A visualization indicative of temporal and semantic distance is shown, given a set of user-selected concepts. Specifically, Figure 2 The results of the method described above are shown. The distance in the x-direction indicates the time distance of the time index step, and the distance in the y-direction indicates the score of the ranking learning model relative to the text block in the previous time index. Utilizing Figure 2The examples shown use the concepts of imidazol and hydrogenation. As an example only, two text blocks 200 and 210 with time index j are shown. Each of the two text blocks 200 and 210 with time index j includes a connection 220 to a text block 230 with time index j-1. Further, each of the two text blocks 200 and 210 with time index j includes a connection 240 to a text block 250 with time index j+1, which includes a lower value in the y-direction, implying a lower score for the ranking learning model relative to the previous text blocks 200 and 210 with time index j. Further, each of the two text blocks 200 and 210 with time index j includes a connection 260 to a text block 270 with time index j+2, which includes a higher value in the y-direction, implying a higher score for the ranking learning model relative to the previous text blocks 200, 210, and 250 with time indices j and j+1. Furthermore, text block 250 with time index j+1 includes a connection 280 to text block 270 with time index j+2. As indicated by reference numeral 290, a user can click on the edges of a text block to view, for example, a ranking score at text block 270 with time index j+2. Further, and by way of example only, text block 250 with time index j+1 includes connections 300, 310 to a first node 320 and a second node 330. As indicated by reference numeral 340, a user can click on nodes 320, 330 to access concept selection and view highlighted text content, metadata, and concepts within the text.

[0107] Figure 3 Another example of the invention is shown. In this example, the method is applied to a production process, particularly in a chemical plant. Maintaining consistent product quality is crucial for the company.

[0108] Recently, collecting customer feedback has become an increasingly common practice. Customer information can be stored in databases. Customer feedback may indicate problems that were not detected during the production process. Some, but not all, customer feedback may indicate problems that were not detected during the production process. Furthermore, customer feedback often lacks standardized formatting and expression. Assessing whether a customer complaint truly points to an error in production or represents dissatisfaction from an individual customer is very difficult.

[0109] As a hypothetical example, customers of car manufacturers might complain in various ways:

[0110] - The color of my car is very dependent on the angle.

[0111] - Inconsistent reflection from the coating

[0112] - The coating looks dull.

[0113] - The engine is very loud.

[0114] -Car accelerator

[0115] -wait.

[0116] It becomes clear that information needs to be clustered according to themes. At the same time, following the chronological order in which text blocks appear is valuable.

[0117] At least a portion of each customer's feedback can be considered corpus data.

[0118] A visualization indicating temporal and semantic distance, given a set of user-selected concepts, in which the concepts are coating and failure. Specifically, Figure 3 The results of the method described above are shown. The distance in the x-direction indicates the time distance of the time index step, and the distance in the y-direction indicates the score of the ranking learning model relative to the text block in the previous time index. Utilizing Figure 3 The example shown uses the concept of coating. As an example only, two text blocks 400 and 410 with time index k are shown. Each of the two text blocks 400 and 410 with time index k includes a connection 420 to a text block 430 with time index k-1. Further, each of the two text blocks 400 and 410 with time index k includes a connection 440 to a text block 450 with time index k+1, which includes a lower value in the y-direction, implying a lower score for the ranking learning model relative to the previous text blocks 400 and 410 with time index k. This indicates that the semantics are similar. Further, each of the two text blocks 400 and 410 with time index k includes a connection 460 to text blocks 480 and 490 with time index k+2, which includes a higher value in the y-direction, implying a higher score for the ranking learning model relative to the previous text blocks 400, 410, and 450 with time indices k and k+1. Further text block 470 with time index k+3 includes a connection 460 to text block 400. Clusters 480, 490, and 470 are relatively consistent in the y-direction, indicating that the text blocks are semantically similar. The x-axis representing the time series indicates and visualizes that the occurrence of similar text blocks is also closely related in time. This representation allows for the direct detection of text block 400 as the first occurrence of something that might trigger customer feedback. This allows for investigation of the production process around time k, targeting potential errors in the production process. Figure 3 Visualizations in the system can infer the causal relationships of errors in production that would otherwise go undetected.

[0119] Figure 4A computer system 100 for generating digital information data in a subject area is illustrated. The processing unit 110 may be or may include any logic circuit configured to perform basic operations of the computer or system, and / or a device typically configured to perform computational or logical operations. Specifically, the processing unit 110 may be configured to process basic instructions that drive the computer or system. As an example, the processing unit 110 may include at least one arithmetic logic unit (ALU), at least one floating-point unit (FPU), such as a math coprocessor or a number coprocessor, multiple registers specifically configured to provide operands to the ALU and store the results of operations, and memory, such as L1 and L2 caches. Specifically, the processing unit 110 may be a multi-core processor. Specifically, the processing unit 110 may be or may include a central processing unit (CPU). Additionally or alternatively, the processing unit 110 may be or may include a microprocessor; therefore, specifically, the elements of the processing unit may be contained within a single integrated circuit (IC) chip. Alternatively or additionally, the processing unit 110 may be or may include one or more application-specific integrated circuits (ASICs) and / or one or more field-programmable gate arrays (FPGAs).

[0120] Database 120 may be or may include any set of information and / or a physical structure configured to store any set of information. Database 120 may include at least one storage device configured to store information. Database 120 may be or may include at least one element selected from the group consisting of at least one server, at least one server system including multiple servers, at least one cloud server, or cloud computing infrastructure. The method may be performed using multiple databases 120. Database 120 may include additional sub-units, such as at least one document storage 140, and may additionally or alternatively include at least one knowledge base 160. The method may be performed using a single database 120 configured to implement multiple functions such as data storage and knowledge storage. For example, document storage 140 may be integrated into knowledge base 160 or may be an external device.

[0121] Digital information data can be a discrete, discontinuous representation of any textual information. Digital information data may include one or more of the following: scientific documents, research-related documents, development-related documents, business-related documents, company-related documents, legal documents, patent documents, regulatory documents, user manuals, instruction manuals, training materials, etc.

[0122] Processing unit 110 is operatively coupled to at least one database 120. Specifically, a communication connection 125 exists between processing unit 110 and at least one database 120 for one or more of the following functions: transmitting information, accessing storage, or controlling at least one device. The processing unit may also be coupled to memory 115. Processing unit 110 and database 120 may include at least one communication interface through which processing unit 110 and database 120 are operatively coupled. Processing unit 110 may be configured to access (e.g., read and write) storage in the database via the communication interface. The communication interface may be or may include items or elements forming boundaries configured for transmitting information. In particular, the communication interface may be configured to transmit information from a computing device (e.g., a computer), such as sending or outputting information, for example, to another device. Additionally or alternatively, the communication interface may be configured to transmit information to a computing device, such as to a computer, such as to receive information. The communication interface may specifically provide means for transmitting or exchanging information. In particular, the communication interface may provide data transmission connections, such as Bluetooth, NFC, inductive coupling, etc. As an example, the communication interface may be or may include at least one port, including one or more of a network or internet port, a USB port, and a disk drive. The communication interface may be at least one network interface. The processing device may be further coupled to the client device 145, particularly via the communication interface 135. In one embodiment, the system may be located in the cloud, and the communication interface 135 may be a network connection.

[0123] List of reference signs

[0124] 100 Computer Systems

[0125] 110 Processing Unit

[0126] 120 Database

[0127] 140 Document Storage

[0128] 160 Knowledge Base

[0129] 200 text blocks

[0130] 210 text blocks

[0131] 220 connection

[0132] 230 text blocks

[0133] 240 connections

[0134] 250 text blocks

[0135] 260 connection

[0136] 270 text blocks

[0137] 280 connection

[0138] 290 clicks on the side

[0139] 300 connections

[0140] 310 connection

[0141] 320 First Node

[0142] 330 Second Node

[0143] 340 Click Nodes

[0144] S10 Start

[0145] S12 calculates once by embedding the words and documents used for concepts into the entire corpus.

[0146] S14 Users query the semantic search engine on a corpus annotated with semantic information extraction.

[0147] S16 Users filter by process attributes

[0148] S18 Extract partial headings from the document

[0149] S20 User selects the portion used to break the document into text blocks.

[0150] S22 uses the top k result documents from the semantic search to compute word and document embeddings for concepts in the text blocks.

[0151] S24 Index the text blocks in chronological order.

[0152] S26 Start from the most recent timestamp

[0153] S28 For each text block i,j with timestamp j, select text block m,j-1.

[0154] S30 For each concept in text block i,j, cluster embedding is performed on the concept embeddings in all text blocks m,j-1 to identify a candidate concept list in the database.

[0155] S32 uses features that evaluate graph relations between candidate concepts and semantic similarity between text block i,j and all text blocks m,j-1, and applies them to a ranking learning model trained on an existing corpus.

[0156] S34 uses the top k candidate concepts to annotate text blocks i,j

[0157] S36 connects text block i,j with the top k ranked text blocks m,j-1, and labels the edges with the scores from the ranking learning model.

[0158] S38 is repeated until all text blocks are clustered.

[0159] S40 writes the text block into the semantic graph as a node labeled with time warehouse i.

[0160] S42 writes the connections between text blocks as directed edges into the semantic graph.

[0161] S44 generates an interactive 2D tree visualization with text block nodes as symbols and edges as arrows, sorted from left to right by time index. It displays a list of given concepts selected by the user from at least a few selected text blocks, text blocks annotated with at least one selected concept, a distance on x indicating the time distance at the time index step, and a distance on y indicating the score of the ranking learning model relative to text blocks in the previous time index.

[0162] S46 End

[0163] 400 text blocks

[0164] 410 text block

[0165] 420 connection

[0166] 430 text blocks

[0167] 440 connection

[0168] 450 text blocks

[0169] 460 connection

[0170] 470 text blocks

[0171] 480 text blocks

[0172] 490 text blocks

Claims

1. A computer-implemented method for generating digital information data in a subject area, wherein, The method is applied to a production process, and the method includes: - At the processing unit (110), digital information corpus data is provided; - Digital information seed data is extracted from the digital information corpus data via the processing unit (110); - A search is performed via the processing unit (110) in at least one database (120) including knowledge information, thereby extracting multiple text blocks related to the topic region from the at least one database (120); wherein the search is performed based on the digital information seed data. - The text blocks are indexed in chronological order via the processing unit (110); - For each text block with a predetermined timestamp in the indexed text blocks, a predetermined number of previous text blocks are selected, and clustering embedding is performed by embedding concepts in all of the previous text blocks, and a list of candidate concepts is identified in the database (120) for each concept in the text blocks with the predetermined timestamp. - Using features that evaluate graph relationships between candidate concepts and semantic similarity between the text block with the predetermined timestamp and all the previous text blocks, a ranking learning model trained on existing digital information corpus data is applied at the processing unit (110). - The digital information data is generated using time-organized text blocks via the processing unit (110), wherein generating the digital information data includes generating a visualization indicating the temporal and semantic distances of the text blocks.

2. The method according to claim 1, wherein, Extracting the seed data of the digital information includes: semantic information extraction.

3. The method according to claim 1 or 2, further comprising: The processing unit (110) filters the extracted digital information seed data by process attributes.

4. The method according to claim 1 or 2, wherein, Extracting the plurality of text blocks includes: selecting portions for decomposing the knowledge information from the database (120) into text blocks.

5. The method according to claim 1 or 2, further comprising: The processing unit (110) recursively calculates the semantic similarity between the extracted text blocks.

6. The method according to claim 5, further comprising: The text block with the predetermined timestamp is annotated using the top k candidate concepts.

7. The method according to claim 6, further comprising: The text block with the predetermined timestamp is concatenated with the top k text blocks in the previous text blocks, and labeled with the score of the ranking learning model.

8. The method according to claim 7, further comprising: Repeat the following steps until all text blocks are clustered: select the previous text block, identify the candidate concept list, apply the ranking learning model, and annotate the text block with the predetermined timestamp.

9. The method according to claim 1 or 2, further comprising: At the processing unit (110), the text block is transmitted to the semantic graph as a node marked with a predetermined time slot.

10. The method of claim 9, further comprising: The connections between the text blocks are formed into the semantic graph at the processing unit (110) as traces.

11. The method according to claim 1 or 2, wherein, The visualization is an interactive 2D tree visualization, in which text block nodes are used as symbols and traces as arrows, and are sorted by time index.

12. The method according to claim 11, wherein, The distance in the x-direction indicates the time distance of the time index step, and the distance in the y-direction indicates the score of the ranking learning model relative to the text block in the previous time index.

13. The method according to claim 1, wherein, The database (120) also includes customer information, and the customer information includes customer feedback on the production process.

14. The method according to claim 1, wherein, The production process includes production processes within a chemical plant.

15. The method according to claim 9, wherein, Transmitting the text block to the semantic graph includes writing the text block to the semantic graph.

16. The method of claim 10, wherein, Forming connections between the text blocks into the semantic graph includes: writing the connections between the text blocks into the semantic graph.

17. The method according to claim 10, wherein, The trace includes directed edges.

18. The method according to claim 11, wherein, The trace includes edges.

19. A computer program comprising computer-executable instructions for performing the method according to any one of claims 1 to 18.

20. A computer-readable storage medium having stored thereon computer-executable instructions for implementing the method according to any one of claims 1 to 18.

21. A computer system (100) for generating digital information data in a subject area, comprising: The system includes at least one database (120) and at least one processing unit (110), wherein the processing unit (110) is configured to provide digital information corpus data, wherein the processing unit (110) is configured to extract digital information seed data from the digital information corpus data, wherein the processing unit (110) is configured to perform a search in the at least one database (120) including knowledge information to extract multiple text blocks related to the topic region from the at least one database (120); wherein the search is performed based on the digital information seed data, wherein the processing unit (110) is configured to index the text blocks in chronological order, wherein the processing unit (110) is configured to: for each text block with a predetermined timestamp among the indexed text blocks, select a predetermined number of previous text blocks. The processing unit (110) is configured to: perform clustering embedding by embedding concepts in all the previous text blocks, identify a list of candidate concepts in the database (120) for each concept in the text block having the predetermined timestamp, and apply a ranking learning model trained on existing digital information corpus data at the processing unit (110) using features that evaluate graph relationships between candidate concepts and evaluate semantic similarity between the text block having the predetermined timestamp and all the previous text blocks; and wherein the processing unit (110) is configured to generate the digital information data using text blocks organized by time, wherein generating the digital information data includes: generating visualizations indicating the temporal and semantic distances of the text blocks, wherein the method steps performed by the processing unit (110) are applicable to a production process.

22. The computer system according to claim 21, wherein, The production process includes production processes within a chemical plant.

23. The computer system according to claim 21 or 22, wherein, The at least one processing unit (110) is operatively coupled to the at least one database (120).

24. The computer system according to any one of claims 21 to 23 relating to a computer system, wherein, The computer system is configured to execute, via the at least one processing unit (110), a method for generating digital information data in a subject area according to any one of claims 1 to 18 relating to a method for generating digital information data in a subject area.

Images