Computer-implemented method and system for marking data with a digital watermark

Description

[0001] The invention relates to a computer-implemented method for marking data with a digital watermark and a suitable system for this purpose according to the type defined in more detail in the preamble of claim 9.

[0002] It is common practice to digitally watermark data such as computer-readable files to uniquely identify them. This allows, for example, the verification of a file's authenticity. It makes it possible to trace whether the file was created by an authorized source or modified by unauthorized third parties. Digital watermarks are typically embedded directly into the content to be marked using steganographic methods. However, the content of a file must not be altered so significantly by applying a digital watermark that the file can no longer be used for its originally intended purpose. Ideally, digital watermarks should also be implemented in such a way that their presence in the file is not readily apparent.This prevents third parties from unlawfully removing corresponding digital watermarks from the relevant file.

[0003] With increasing digitalization, the importance of digital data also grows. For example, a vehicle manufacturer can aggregate data describing the usage patterns of its vehicles. This data can be collected from the manufacturer's vehicles during operation and processed on a central computing facility, such as a cloud server. The data generated in this way can, for example, be sold to third parties, allowing them to also carry out vehicle-related technical developments. This enables the development and optimization of vehicle components and / or vehicle-related services.

[0004] There is a risk that a customer of the vehicle manufacturer might unlawfully resell the data acquired by the manufacturer, resulting in lost revenue for the manufacturer. It would therefore be desirable to be able to determine where the data was passed on in order to prevent this in the future. Furthermore, the data generated by the vehicle manufacturer can be manipulated, which can have serious consequences. For example, faulty data could lead to a technical defect or malfunction, potentially injuring or damaging machinery or people. Such an incident could also result in legal action against the vehicle manufacturer. Therefore, it is desirable to provide procedures and tools that make it possible to detect such data manipulation.

[0005] Common methods for marking data with digital watermarks involve, for example, supplementing corresponding computer-readable files with artificially generated content, particularly with a specific format. This supplement then forms the digital watermark. However, if detected, the supplement, and thus the digital watermark, can easily be removed from the corresponding dataset. Therefore, such digital watermarks should ideally not be readily apparent without knowledge of the corresponding implementation scheme.

[0006] Furthermore, it is a known practice in software development to obscure the part of a program's code relevant to its actual functionality by including irrelevant code components. Code components can also be added that alter the program flow, causing irrelevant steps to be performed. This is also known as obfuscation. Should an unauthorized third party read the source code of the program, the underlying code becomes unrecognizable or extremely difficult for humans to read and understand. This makes it more difficult to "reverse engineering" of the program.

[0007] A method for performing obfuscation and a method for evaluating such obfuscation are known, for example, from WO 2007 / 049817 A1. Using the methods described in this publication, it is possible to assess whether the corresponding program code has been sufficiently obfuscated. If this is not the case, a further obfuscation iteration can be performed, thereby improving the obfuscation of the program code. For this purpose, different sections of program code can be subjected to obfuscation and / or different obfuscation methods can be used.

[0008] The present invention is based on the objective of providing an improved computer-implemented method for marking data with a digital watermark, the use of which allows the computer-readable data to be protected more reliably.

[0009] According to the invention, this problem is solved by a computer-implemented method for marking data with a digital watermark having the features of claim 1. Advantageous embodiments and further developments as well as a system for carrying out the method are set forth in the dependent claims.

[0010] A generic computer-implemented method for marking data with a digital watermark is further developed according to the invention by the following process steps executed by a computing unit: a) Providing an initial data set, wherein the initial data set is in a structured format such that the entries of the initial data set have a defined arrangement scheme; b) Reading at least one manipulation scheme from a watermark database communicatively connected to the processing unit; c) Applying a manipulation following the manipulation scheme to each entry of the initial data set to generate a marked data set; d) Providing the marked data set for use; e) Receiving a data set to be checked; f) Determining a probability of match for the data set to be checked as a function of at least one manipulation scheme found in the data set to be checked;(g) if the probability of a match is greater than a specified threshold: comparing the entries of the data record to be checked with a corresponding tagged data record identified for the data record to be checked and identifying entries in the data record to be checked where the manipulations applied in step c) have been removed; and (h) repeating steps (a) to (g), whereby for an entry having the same data type as an entry for which it was determined in step g) that a manipulation had been removed, the processing unit uses a different manipulation scheme from the watermark database.

[0011] Using the computer-implemented method for marking data according to the invention, computer-readable files can be reliably provided with a digital watermark. By iteratively performing process steps a) to g) and h), the processing unit successively employs different manipulation schemes to apply the corresponding markings, thereby gradually making it more difficult to identify the digital watermarks. An entity that is unaware of the underlying manipulation schemes is thus increasingly less able to remove the corresponding manipulations from the marked data set. This allows the initial data set, or the correspondingly marked data set, to be protected ever more reliably against manipulation by unauthorized third parties.An entity executing the method according to the invention is thus reliably able to determine whether a data set obtained from an external source was generated in-house, i.e., corresponds to a corresponding initial data set or is based on such a data set, or originates from another source.

[0012] Since each entry in the initial dataset is imprinted with a corresponding manipulation, it is reliably possible to detect such manipulations in a dataset under review, even if the marked dataset is altered by unauthorized third parties. For example, if some entries are changed or removed, enough other entries with the corresponding manipulation, and thus with a digital watermark, will still be present in the dataset under review. A respective dataset can contain additional content besides the actual entries, such as a header, a "Header" or similar. This additional content can be free of manipulation or also manipulated to be marked with digital watermarks. A single entry can also be manipulated multiple times and thus have several digital watermarks. A pattern with which the processing unit imprints manipulations on the respective entries itself forms a digital watermark.

[0013] The processing unit can generate the initial data set itself or receive it from a trusted source, such as a server, where corresponding values ​​are aggregated. The initial data set can be in a variety of file formats. The only requirement is that the entries in the initial data set adhere to the defined organizational scheme. This allows entries within the data set to be distinguished from one another. For example, it could be a text file, where a new line in the text file corresponds to a new entry. The first line of the text file could also contain additional information. Entries could also be separated by spaces, or similar characteristics.

[0014] The manipulation schemes used are specifically designed for each data type. All common data types can be supported, particularly with automatic detection of standard technical data types such as strings, booleans, num, ints, and the like. For example, mathematical operations can be applied to numerical values, but not to a string of letters.

[0015] After the marked data set is generated, it is released for use. This means that third parties can process and manipulate the marked data set. If such manipulation is suspected, the corresponding data set is submitted to the processing unit as a data set to be reviewed. The processing unit then analyzes whether at least one manipulation corresponding to a manipulation scheme stored in the watermark database can be found in the data set to be reviewed. Such a manipulation could be contained in a corresponding data set to be reviewed by chance. In order to distinguish a coincidental match from an actual manipulation of an entry beyond doubt, the processing unit determines the probability of a match in step g) and compares it with the respective defined threshold.In the simplest case, it might be sufficient if at least one manipulation is found in the dataset to be checked. However, the probability of a match could also be determined, for example, as a ratio of the number of manipulations actually found in the dataset to be checked to the maximum number of manipulations permitted in the respective dataset according to the computing unit's procedure.

[0016] If the probability of a match is greater than the defined threshold, this means that the data set being checked is an initial data set that has been altered by an unauthorized third party, or that the data set being checked contains entries that were originally taken from the flagged data set. This constitutes misuse. The threshold can be of a suitable value, such as 5%, 10%, 20%, 50%, or fractions or multiples thereof. The threshold can be adjusted to suit specific situations.

[0017] The aim of the method according to the invention is to increase the reliability of ensuring that corresponding digital watermarks or manipulations are not removed from the marked data set. To this end, the processing unit identifies the marked data set corresponding to the data set to be checked and determines for which entries the corresponding manipulation has been removed. The manipulation scheme used for these entries or this data type is thus known to relevant third parties, since they are able to remove the corresponding manipulations from the marked data set. Accordingly, in step h), the processing unit uses a different manipulation scheme for corresponding entries or data types, thus increasing the probability that this manipulation scheme is unknown to unauthorized third parties.

[0018] In step f), the processing unit already checks which entries are free of corresponding manipulations and thus digital watermarks. An entry in a data set under review may be free of manipulation because either the manipulation was deliberately removed, or the entry was added by the unauthorized third party without knowledge of the manipulation scheme used to mark data with digital watermarks. To prove beyond doubt that the manipulation has been removed, a corresponding entry from a data set under review should therefore be compared with the corresponding entry from the corresponding marked data set.

[0019] To determine a corresponding tagged data record, the computing unit can compare the content of a data record to be checked with a tagged data record held in a database and designate that tagged data record as "corresponding" Select the record that has the highest content overlap with the record to be checked. Thresholds can also be defined here, specifying the point at which, despite a certain degree of overlap, no correspondence is found, for example, because one or more entries happen to match. The marked records can be added to the database by the processing unit when they are generated.

[0020] It should be noted that the absence of manipulations in a data set under review can also indicate the strength of a manipulation scheme, while the presence of manipulations can indicate its weakness. It is therefore important to distinguish between entries that were already present in the marked data set and entries that were added by unauthorized third parties. Manipulations present in entries added by third parties themselves indicate that the third party has identified the manipulation scheme. According to the invention, the processing unit can also change such a manipulation scheme. Manipulations present in original entries but absent in entries added by third parties indicate a strong manipulation scheme. According to the invention, the processing unit can thus refrain from changing such a manipulation scheme for the future generation of marked data sets.

[0021] An advantageous further development of the method according to the invention provides that the computing unit, for manipulating an entry existing in the initial data set, supplements the respective entry according to a manipulation pattern, in particular by adding a digit obtained by applying a mathematical operation to an entry consisting of a numerical value, and / or by changing at least one formatting element of the respective entry according to a manipulation pattern. Thus, a wide variety of possibilities for manipulating respective entries are available. If the entry contains a numerical value, this can be supplemented by a digit at any position. This digit is generated, in particular, depending on the unmanipulated numerical value. Various mathematical operations can be applied to the corresponding entry for this purpose.For example, the digit can be calculated by adding up the digits. This process is therefore called calculating a checksum. All common methods for calculating checksums can be used.

[0022] If the entry is a text string, a letter or special character could be added at any position. For example, an underscore could be inserted at the fourth position of the text string. If the string contains no numbers, the first three letters could be capitalized.

[0023] Additionally or alternatively, a manipulation pattern can involve changing the formatting of an entry. For example, the font size of a specific letter in the entry, such as the first letter, could be decreased or increased by one. For example, the font color of a character at a specific position could be changed from black to a very dark gray. Several formatting changes can also be applied in combination. For example, a white underline could be added to an entry that is already in black at any point. This makes it even more difficult for an unauthorized third party to detect such manipulations. A particularly preferred manipulation scheme stipulates that such additions or formatting changes are applied according to a defined pattern, such as for every entry or every nth entry (n being an even number).For example, in a text string, a white underscore could be inserted after the first four letters every five entries, while a space could be inserted in the corresponding position every sixth entry. To determine whether such manipulations have occurred in the data set under review, the processing unit checks whether or not it finds the respective patterns that would explain the additions or format changes.

[0024] According to a further advantageous embodiment of the method according to the invention, it is further provided that the computing unit inserts at least one artificially generated entry into the initial data set for the additional characterization of the initial data set and stores the artificially generated entry, in particular comprising arrangement information describing the arrangement of the artificially generated entry in the initial data set, in an artificial entry database; and wherein, in step f), the computing unit checks, in order to determine the probability of a match, whether at least one artificial entry stored in the artificial entry database can be found in the data set to be checked. Thus, not only can existing entries be supplemented or their formatting changed in a given initial data set, but entirely new entries can also be synthetically produced and inserted into the initial data set.Such an artificially generated entry represents a particularly striking example. "Marker" This is done to function as a digital watermark. Ideally, such an artificially generated entry is only created after the existing entries of the initial data set have already been manipulated.

[0025] The generation of an artificially generated entry can follow predefined rules. For example, the generation of an artificially generated entry can be designed to depend on the entries already present in the initial dataset. For instance, an average value can be calculated from some or all of the numbers contained in the dataset and added as an artificially generated entry. Similarly, the artificial generation of entries can follow a specific pattern. For example, an artificially generated entry can be included in the initial dataset every 1000 entries. Adding artificially generated entries also transforms the initial dataset into a marked dataset. In a broader sense, corresponding artificially generated entries can therefore also be added to an already marked dataset.Such an artificially generated entry is a particularly strong indication that the data set under review is an initial data set or that it is based on such an initial data set. This allows for a particularly reliable determination of whether the data set under review originates from within the organization or not.

[0026] A further advantageous embodiment of the method according to the invention provides that the structured format corresponds to a table. In other words, the entries of the initial data record or the marked data record are then sorted by rows and columns. This is a particularly frequently used format, which explains its relevance for adding digital watermarks to such tabular information.

[0027] According to a further advantageous embodiment of the method according to the invention, it is further provided that the computing unit in step f) increases the probability of agreement if: the ratio of the number of manipulations still present in the data set to be checked to the total number of manipulations that can be imprinted on the entries according to the manipulation schemes used is greater than 5%, excluding manipulations in the form of checksums; the ratio of the number of manipulations still present in the data set to be checked to the total number of manipulations that can be imprinted on the entries according to the manipulation schemes used is greater than 10%, for manipulations in the form of checksums; and / or the data set to be checked contains at least one artificially generated entry.

[0028] By modifying a corresponding data record, an unauthorized third party can alter entries in such a way that the manipulations applied by the processing unit in step c) are removed. The unauthorized third party could also add entirely new entries to a flagged data record. In step f), the processing unit checks whether manipulations matching the respective manipulation schemes can be found for each entry in the data record under review. The processing unit can also maintain the respective initial data records and / or flagged data records in a data record database and compare their content with that of a data record under review. Thus, not only can the presence or absence of manipulations in the data record under review be determined, but also matches or discrepancies in the content itself.The more manipulations are found in the data set being examined, the higher the probability of a match. Thresholds of 5% or 10% have proven effective in distinguishing between manipulations actually imposed by the processing unit and a coincidental match. Other percentage values ​​would also be possible.

[0029] Since it is virtually impossible for an unauthorized third party to randomly generate an entry artificially created by the processing unit, the mere detection of such an entry within the dataset under review is sufficient to identify it as having been generated internally. The processing unit can therefore determine a 100% probability of match if at least one such artificially generated entry is found within the dataset under review. Alternatively, the probability of match could simply be increased, for example, by 1%, 5%, 10%, or even fractions or multiples thereof.

[0030] A further advantageous embodiment of the method according to the invention further provides that the watermark database is supplemented by a new manipulation scheme, wherein the new manipulation scheme: is determined manually by a developer; is determined automatically by artificial intelligence, in particular taking into account the frequency of manipulations in datasets to be checked; or is determined automatically by a stochastic combination of predefined, data-type-specific manipulation rules.

[0031] This prevents the manipulation schemes stored in the watermark database from becoming exhausted over time. New manipulation schemes can be defined manually by developers or automatically. Developers can manually check which manipulation schemes currently in use are most frequently identified and neutralized by unauthorized third parties, and which are not. This makes it possible to identify particularly reliable manipulation schemes and transfer the manipulation processes underlying these schemes to newly developed manipulation schemes. This ensures that the newly developed manipulation schemes are also even more reliably unidentifiable by unauthorized third parties.Artificial intelligence is preferably used for the development of these new manipulation schemes, which reduces the effort required. This also lowers the risk of human error. Artificial intelligence is particularly reliable at identifying patterns and thus deducing manipulations that are especially difficult for unauthorized third parties to detect. A suitably pre-trained artificial intelligence is used.

[0032] Data type-specific manipulation rules can also be predefined and randomly combined to create new manipulation schemes. These rules can specify, for example, whether a manipulation should involve adding to an existing entry and / or changing a format. A data type-specific manipulation rule can also specify where within a given entry the manipulation should occur. Furthermore, a data type-specific manipulation rule can define the type of addition and / or format change. All these data type-specific manipulation rules can then be randomly combined in any desired way. This allows for the particularly reliable and simple automated generation of new manipulation schemes.

[0033] A pre-trained, generative artificial intelligence is preferred. "Transformers"This is used. With the help of such a pre-trained generative transformer, it is particularly reliable to implement digital watermarks that are difficult for unauthorized third parties to identify. By analyzing a large number of combinations of marked data records and corresponding data records to be checked, such a transformer is reliably able to adopt the analysis pattern of an unauthorized third party. The transformer is thus able to detect manipulations in data records in the same way that an unauthorized third party would. This, in turn, enables the pre-trained generative transformer to derive manipulation schemes that cannot be traced by said unauthorized third party. This enables particularly reliable protection of the data to be marked with a digital watermark.

[0034] A further advantageous embodiment of the method according to the invention provides that the computing unit generates different marked data records from an initial data record, wherein the computing unit imposes a different manipulation on at least one specific entry by applying various manipulation schemes, and the computing unit provides a first marked data record for use by a first entity and at least one further marked data record for use by at least one further entity. In other words, the computing unit is able to generate different marked data records from the same input data for different customers.If, during the analysis of a data set under review, it is determined that it was generated internally, the entity-specific manipulations found within the data set allow for the identification of the entity—that is, the potential customer—from which the data set originates. This enables the identification of unreliable entities and the implementation of appropriate measures to reduce or even completely prevent the misuse of the company's data by said entity. For example, this allows for targeted searches for security vulnerabilities, the termination or restriction of cooperation with the entity in question, and similar actions. Furthermore, different departments within the company may employ different manipulation schemes to generate differently marked data sets.This allows the origin, recipient and / or seller of the corresponding data to be clearly identified with the help of digital watermarks.

[0035] In a system for marking data with a digital watermark, comprising a processing unit, the processing unit is configured, according to the invention, to execute a method described above. The system can be implemented locally, for example on a desktop computer, or distributed in a cloud environment. The processing unit can accordingly be a server or server cluster, or a component of a server or server cluster.

[0036] The system has access to a computer-readable storage medium in which a computer program product is stored; in turn, it has computer-interpretable instructions whose execution by a processor of the computing unit causes the computing unit to execute a procedure described above.

[0037] Further advantageous embodiments of the computer-implemented method according to the invention for marking data with digital watermarks also result from the exemplary embodiments which are described in more detail below with reference to the figures.

[0038] This shows: Fig. 1 is a schematic representation of a method known from the prior art for marking a tabular data set with a digital watermark; Fig. 2 is a schematic representation of a method according to the invention for marking a tabular data set with a digital watermark; Fig. 3 is a schematic flowchart of a computer-implemented method according to the invention for marking data with digital watermarks; and Fig. 4 is a schematic representation of an exemplary manipulation of an entry of a data network to be provided with a digital watermark in the form of a number.

[0039] With the help of digital watermarks 1, it is possible to prove the authenticity of a computer-readable file and to ensure its traceability. Figure 1An example Table 10 is shown as a computer-readable file. Table 10 includes entries 4 and a header row 11. A digital watermark 1, an artificially generated entry 4* with specific content and / or format, can be added at a suitable location in Table 10. If the artificially generated entry 4* is present in Table 10, this indicates that Table 10 is authentic. The digital watermark 1 was inserted into Table 10 by its creator. Figure 1However, as shown, an unauthorized third party can remove the digital watermark 1, in this case the hatched rows, from Table 10. This would compromise the file's authenticity. For this to occur, the unauthorized third party would need to know the content and format of the digital watermark 1. This is relatively easy to do. Furthermore, there is a high risk that the artificially generated entry 4* will be removed from Table 10 by a table modification already applied by the third party.

[0040] Figure 2The inventive method for marking computer-readable files with digital watermarks 1 is illustrated. Here, the digital watermark 1 is not only artificially generated entries 4* added to a corresponding table 10, but each entry 4 of a corresponding initial data record 3 is provided with a manipulation 6. Even if an unauthorized third party removes the artificially generated entry 4* and at least some of the manipulations 6, enough manipulations 6 remain to allow the authenticity of the corresponding data record to be verified.

[0041] Here too, table 10 or the initial data record 3 can have a header row 11 whose content is not subject to any manipulations 6. However, it would also be conceivable to include manipulations 6 here as well.

[0042] The manipulations 6 should not be detectable or traceable by an unauthorized third party, as they could otherwise be too easily removed or imitated. Furthermore, the respective initial data set 3 must not be altered by the provision of such manipulations 6 to such an extent that it can no longer be used for its intended purpose. In addition, enough entries 4 of the initial data set 3 must be provided with a manipulation 6 that even after the removal of a large number of the corresponding manipulations 6, the authenticity of the data set in question can still be unequivocally established. For this purpose, a computer-implemented method according to the invention for marking data with a digital watermark 1 serves, the process of which is schematically shown in Figure 3 The sub-figures 3a) to h) correspond to the respective process steps.

[0043] In step a), an initial data record 3 is provided, also in the form of a table 10. This initial data record 3 is in a structured format. This means that the entries 4 of the initial data record 3 have, or follow, a defined arrangement scheme. The arrangement scheme can specify that each entry 4 is preceded by a header row 11. Since the initial data record 3 is a table 10, this arrangement scheme stipulates that the respective entries 4 are ordered by row and column. Each entry 4 corresponds to a specific data type. All established and well-known data types can be used, such as Boolean, Byte, Decimal, Long, String, Integer, Float, and the like.

[0044] In step b), a computing unit 2 executing the method according to the invention reads at least one manipulation scheme from a watermark database 5 that is communicatively connected to the computing unit 2.

[0045] In step c), the processing unit 2 imprints a manipulation 6, following a specific manipulation scheme, onto each entry 4 of the initial data record 3 to generate a marked data record 7. Preferably, the processing unit 2 combines several different manipulation schemes. The specific manipulation scheme depends in particular on the data type of each entry 4. For example, different manipulations 6 can be imprinted on texts than on numbers. Each manipulation 6 forms the digital watermark 1. The pattern of manipulations 6 itself also constitutes such a digital watermark.

[0046] As manipulation 6, a respective entry 4 can be added at any specified position and / or the formatting of any specified character can be changed in any specified way. It is also possible to insert an artificially generated entry 4* into the respective initial data record 3 to generate a respective marked data record 7. By implementing the corresponding manipulations 6, the aforementioned marked data record 7 is generated. Figure 3 Each such artificially generated entry 4* is indicated by hatching. Several artificially generated entries 4 are added to form a complete artificially generated row in Table 10.

[0047] In the illustrated embodiment, the data type of the entries 4 differs between the respective columns of Table 10. For example, the first column may contain a vehicle identification number (VIN), the second column the name of a respective vehicle buyer or owner, and the third column a sensor value. For a data type in the form of a string, an underscore can be inserted at a suitable position as a manipulation 6. This underscore can, for example, be added after the first three consecutive letters. Strings containing both letters and numbers can be automatically capitalized. For numbers, the concept of generating a so-called checksum can also be applied.

[0048] For example, an artificially generated entry 4* can be inserted into table 10 every 1000 rows. The artificially generated entries 4*, or the respective initial data records 3 and / or marked data records 7, can be stored completely or partially in a separate database by the processing unit 2. The processing unit 2 can optionally store additional metadata, such as a timestamp, a customer ID, the number of rows and columns in table 10, an access date, a storage size or file size, and the like.

[0049] Artificially generated entries 4* are preferably only generated after the other manipulations 6 have already been implemented. This allows the generation of artificial entries 4* to be made dependent on the manipulations 6. For example, the most frequent value or mode of a given row can be randomly selected for strings. An average value can also be determined for each column and added to a specific number. Alternatively, any table entry could be selected and simply copied to generate the artificially generated entry 4*.

[0050] In step d), the marked data set 7 is then distributed to different users for their use. During this time, unauthorized manipulation or use of the underlying data can occur. For example, an unauthorized third party could resell the data or modify it in an unlawful manner. This third party does not necessarily have to be the entity to which the marked data set 7 was originally distributed. A security vulnerability could also exist within such an entity, allowing the corresponding marked data set 7 to be stolen.

[0051] Later, in step e), a data set 8 to be checked is fed to the processing unit 2. It is suspected that this data set 8 is based on a labeled data set 7, which now needs to be verified.

[0052] For this purpose, in step f), the processing unit 2 determines a probability of match 9 for the data record 8 to be checked, depending on whether at least one manipulation scheme 6 is found in the data record 8 to be checked. For this purpose, the processing unit 2 accesses the aforementioned watermark database 5 and checks whether the respective manipulations 6 can be found in the data record 8 to be checked or not. This also includes checking for the presence of corresponding artificially generated entries 4*.

[0053] Since it is highly unlikely that an unauthorized third party would generate such an artificially created entry 4*, a high probability of match 9 can be determined if such an artificially created entry 4* exists in the data set 8 to be checked. The more manipulations 6 are identified by the processing unit 2, the higher the probability of match 9. A threshold of 10% for manipulations 6 in the form of checksums and 5% for other manipulations 6 has proven effective as a limit for the ratio of manipulations 6 found in the data set 8 to be checked to the generally foreseeable manipulations 6. For example, if 15 manipulations 6 deviating from a checksum were intended, of which seven were removed, there is a match of 8 / 15. This is greater than 5%, so it is very likely that the data set 8 to be checked is based on a marked data set 7 or an underlying initial data set 3.For example, of a total of six planned manipulations, one in the form of checksums might have been removed, resulting in a match of 5 / 6. This is also greater than the aforementioned 10%, so a data set 8 to be checked in this way very likely originates from within the organization.

[0054] To calculate the probability of agreement 9, the processing unit 2 can apply a variety of rules. For example, an individual percentage value can be defined for different manipulation schemes, specifying how many of the manipulations 6 must still be present. All these percentage values ​​can be aggregated to determine the probability of agreement 9. For example, the processing unit 2 can calculate an average from these values, or perform similar actions. For instance, a manipulation 6, particularly the detection of artificially generated entries 4*, can be used as a hard boundary condition, such that if at least one artificially generated entry 4* is present, the probability of agreement can jump to 100%.

[0055] In step g), the processing unit 2 compares, for a data set 8 to be checked whose probability of match 9 exceeds a defined threshold, whether corresponding manipulations 6 have been removed from the respective entries 4. For this purpose, the processing unit 2 compares the content of the data set 8 to be checked with the content of its corresponding tagged data set 7. In this step, the processing unit 2 can identify entries 4 already contained in the tagged data set 7 and entries added by unauthorized third parties. Manipulation schemes in which the corresponding manipulations 6 have been removed are then not used again by the processing unit 2 to apply digital watermarks 1, but are replaced by other manipulation schemes. This can be applied to manipulations 6 that are found in entries created by said third parties themselves.Various criteria can be defined to determine when a particular manipulation scheme should be replaced. It is reasonable to assume that not every entry 4 manipulated according to the same manipulation scheme was altered by an unauthorized third party. Therefore, a threshold can be set for the frequency of removing manipulations 6 from the manipulation scheme, above which the manipulation scheme is replaced. For example, if 75% of all manipulations 6 imposed by a specific manipulation scheme are detected and removed from the data set 8 under review, the corresponding manipulation scheme may need to be changed.

[0056] The method according to the invention is based on the iterative generation of corresponding marked data sets 7 and the analysis of the data sets 8 to be checked, which are fed back to the processing unit 2. Thus, the processing unit 2 learns to gradually employ increasingly reliable manipulation schemes, enabling reliable authentication of data. For this purpose, new manipulation schemes are gradually added to the watermark database 5. These new manipulation schemes can be defined, for example, manually by a developer, automatically by a computer, or automatically based on a stochastic combination of predefined, data-type-specific manipulation rules. This iterative approach is described in Figure 3h ) indicated.

[0057] Figure 4This example shows the process for generating a checksum as manipulation 6. Each entry 4 contains a number or sequence of digits. A suitable digit for the letter "C" must be found. In step 401, each digit of the number is considered individually, and every second digit is multiplied by 2.

[0058] In step 402, the digital root is determined for the respective numbers obtained in this way.

[0059] Now, in step 403, all the numbers obtained in this way are added together.

[0060] The modulo 10 of the result should, for example, be 0. Accordingly, in step 404, 9 is determined for "C".

[0061] The verification of whether such a manipulation 6 has occurred is carried out in the same way. The processing unit 2 itself determines a value for "C" and compares it with the value 9. If both values ​​match, this means that a manipulation 6, as provided for by the processing unit 2, is present in the data set 8 to be checked. The in Figure 4 The flowchart shown is only one example. All other possible approaches for calculating checksums and for implementing respective manipulations are possible.

[0062] A system executing the computer-implemented method according to the invention is also part of the invention.

[0063] The computing unit 2 particularly preferentially generates differently marked data sets 7 for different entities or customers, which makes it possible to find out when and where corresponding data is generated. "gelact"The method according to the invention allows the authenticity of a data set 8 to be verified. Thanks to the iterative execution, the computing unit 2 uses increasingly reliable manipulation schemes.

Claims

1. Computer-implemented method for marking data with a digital watermark (1), characterized by the following method steps, performed by a computing unit (2): a) providing an initial data set (3), the initial data set (3) being in a structured format such that the entries (4) of the initial data set (3) have a stipulated arrangement scheme in the initial data set (3); b) reading at least one manipulation scheme from a watermark database (5) which is communicatively connected to the computing unit (2); c) imposing a manipulation following the manipulation scheme (6) on each entry (4) of the initial data set (3) in order to generate a marked data set (7); d) providing the marked data set (7) for use; e) receiving a data set (8) to be checked; f) determining a probability of agreement (9) for the data set (8) to be checked on the basis of manipulations (6) found in the data set (8) to be checked and following at least one manipulation scheme; g) if the probability of agreement (9) is greater than a stipulated threshold: comparing the entries (4) of the data set (8) to be checked with a corresponding marked data set (7) identified for the data set (8) to be checked and identifying entries (4) in the data set (8) to be checked for which the manipulations (6) imposed in step c) have been removed; and h) repeating steps a) to g), the computing unit (2) using a different manipulation scheme from the watermark database (5) for an entry (4) having the same data type as an entry (4) for which it was determined in step g) that a manipulation (6) was removed.

2. Method according to claim 1, characterized in that the computing unit (2), in order to manipulate an entry (4) existing in the initial data set (3), supplements the relevant entry (4) following a manipulation pattern, in particular by adding a digit obtained by applying a mathematical operation to an entry (4) formed by a numerical value, and / or by changing at least one formatting process of the relevant entry (4) following a manipulation pattern.

3. Method according to either claim 1 or claim 2, characterized in that the computing unit (2), in order to additionally mark the initial database (3), inserts at least one artificially generated entry (4*) into the initial data set (3) and stores the artificially generated entry (4*), in particular comprising arrangement information describing the arrangement of the artificially generated entry (4*) in the initial data set (3), in an artificial entry database, the computing unit (2) checking, in step f), in order to identify the probability of agreement (9), whether at least one artificial entry (4*) held in the artificial entry database can be found in the data set (8) to be checked.

4. Method according to any of claims 1 to 3, characterized in that the structured format corresponds to a table (10).

5. Method according to any of claims 1 to 4, characterized in that the computing unit (2), in step f), increases the probability of agreement (9) if: - the ratio of the number of manipulations (6) still present in the data set (8) to be checked to the total number of manipulations (6) that can be imposed on the entries following the manipulation schemes used is greater than 5%, excluding manipulations (6) in the form of checksums; - the ratio of the number of manipulations (6) still present in the data set (8) to be checked to the total number of manipulations (6) that can be imposed on the entries following the manipulation schemes used is greater than 10%, for manipulations (6) in the form of checksums; and / or - the data set (8) to be checked contains at least one artificially generated entry (4*).

6. Method according to any of claims 1 to 5, characterized in that the watermark database (5) is supplemented by a new manipulation scheme, the new manipulation scheme being stipulated: - manually by a developer; - automatically by artificial intelligence, in particular taking into account the frequency of removal of manipulations (6) in data sets (8) to be checked; or - automatically by a stochastic combination of predefined, data type-specific manipulation rules.

7. Method according to claim 6, characterized in that a pre-trained, generative transformer is used as artificial intelligence.

8. Method according to any of claims 1 to 7, characterized in that the computing unit (2) generates differing marked data sets (7) from an initial data set (3), the computing unit (2) imposing a deviating manipulation (6) on at least one determined entry (4) by applying various manipulation schemes and the computing unit (2) providing a first marked data set for the use of a first entity and at least one further marked data set for the use of at least one further entity.

9. System for marking data with a digital watermark (1), comprising a computing unit (2), characterized in that the computing unit (2) is designed to perform a method according to any of claims 1 to 8.