Method of communication with a computer system comprising a knowledge base

Abstract

The invention relates to a method of communication with a computer device (CD) comprising a knowledge base (KB) modeling data in the form of a knowledge graph (KG), said method comprising the following, at the level of said device: - receiving (S2) a question posed using a natural language interface, - generating (S4-S5) an answer to said question using a language model (LNT), said generated answer comprising textual data (TXT) resulting from a textual conversion (S4) of a data table, during which data from the table are enriched (S3) using data from the knowledge graph. Figure 3

Description

Title of the invention: Method for communicating with a computer device comprising a knowledge base. Field of the invention

[0001] The field of the invention is that of querying databases. More specifically, the present invention relates to a method of communicating with a computer system (computer, server, platform, etc.) comprising a knowledge base modeling data in the form of a knowledge graph. During communication, the computer system is queried by a user using a natural language interface and responds to the user in natural language, using knowledge such as tabular data that can be either provided by the user, acquired through training, or obtained from the Web. The present invention also relates to the computer system, a corresponding computer program, and a storage medium. Previous art

[0002] With the development of natural language communication based on Artificial Intelligence techniques, numerous applications now allow a user to converse with a machine in natural language. One such application is the ChatGPT™ conversational agent, which is specifically designed to receive a question posed by a user in natural language, understand that question, and provide a natural language answer as accurately as possible. To generate such an answer, a conversational agent of this type sometimes needs to convert data tables into text. However, current techniques for converting data tables into text are limited by an insufficiently precise conceptualization of tabular data, making the natural language conversion less accurate and detailed, and generally less informative.Indeed, some implicit information in a table is not transcribed into textual form, which impairs the quality of the conversion and sometimes results in text content with undesirable ambiguities. Furthermore, the complexity of the relationships between the data in a table is not always correctly accounted for by previous techniques, given a limited understanding of table structure. Object and summary of the invention

[0003] One of the aims of the invention is to remedy at least one of the drawbacks of the aforementioned prior art by proposing a new technique for converting a data table into text used in the context of a natural language dialogue between a user and a machine, based on the use of data from a knowledge graph.

[0004] To this end, an object of the present invention relates to a method of communication with a computer device comprising a knowledge base modeling data in the form of a knowledge graph, said method comprising the following, at the level of said device:

[0005] - to receive a question posed using a natural language interface,

[0006] - generate an answer to said question using a language model, said generated response including textual data from a textual conversion of a data table, during which data from the table is enriched using data from the knowledge graph.

[0007] Such a communication method has the advantage of contributing to the improvement of the user experience, to the efficiency of data processing in order to develop a response, as well as to the accuracy of the response provided.

[0008] In particular, such a communication method allows a significant improvement in the accessibility of data by a user, by allowing data from a textual conversion of a data table to be injected into the responses provided to the user, such data being more accessible and more understandable than data from a table which proves to be complex and difficult for a user to interpret.

[0009] Furthermore, using data from a knowledge graph to perform this conversion has the advantage of enriching the table data, and therefore the content of the responses provided to the user. Indeed, the communication method according to the invention makes it possible to integrate additional information and relationships between the data, thus enriching the text response.

[0010] The communication method according to the invention also allows users to easily navigate through textual information, which is often more intuitive and easier than browsing data tables.

[0011] According to a particular embodiment, the enrichment of the table data may include the following:

[0012] - semantically annotate at least some of the data in the table using knowledge graph data,

[0013] - for at least one column of the table containing annotated data semantically, to obtain additional information on said semantically annotated data, from the data in the knowledge graph, and to supplement said semantically annotated data in at least one column with the additional information obtained,

[0014] and the textual conversion includes the following:

[0015] - query said language model from a prompt generated in language natural by said device, said prompt asking the language model to generate a textual description of the table from the semantically annotated table data that have been completed,

[0016] - receive said textual description generated in response to the query.

[0017] Thanks to such an embodiment, it is possible to enrich a data table before its conversion into text from a knowledge graph, so as to obtain a text whose content, on the one hand, is enriched and therefore particularly precise, and, on the other hand, in which ambiguities are minimized.

[0018] Such an embodiment also offers the following advantages:

[0019] - The use of an intermediate structure such as the knowledge graph allows to capture and faithfully convert the knowledge contained in tabular data and to augment this knowledge with new related facts from the knowledge graph;

[0020] - Verbalization based on a language model ensures a textual description of high quality.

[0021] According to another particular embodiment, the semantic annotation of at least some of the data in the table may include, for at least one cell and / or at least one column and / or at least one pair of columns in the data table, an extraction, in the knowledge graph, of at least one identifier of an entity in the knowledge graph.

[0022] Using knowledge graph entity identifiers to annotate a data table allows for the retrieval of additional information, thus enriching the table data with more complete and precise details. By using specific identifiers, the communication process can access accurate and verified information, thereby improving the precision of the annotations. Entity identifiers make it possible to link the table data to a broader network of information in the knowledge graph, revealing relationships and dependencies that would not otherwise be apparent. Such semantically annotated data can enable a user to discover new relationships between the table data, predict trends, identify complex correlations, and so on. This embodiment significantly improves the quality of the table's text conversion.

[0023] According to another particular embodiment, the semantic annotation of at least some of the data in the table may include the following:

[0024] - associate at least one cell in the data table with an extracted entity identifier of the knowledge graph,

[0025] - associate at least one column of the data table with an identifier of type entity extracted from the knowledge graph,

[0026] - associate at least one pair of columns from the table with a property identifier of the knowledge graph.

[0027] Such an embodiment allows the data table to be annotated according to several levels of knowledge, the data table being characterized using the entities, types / classes, and properties carried by the knowledge graph, which constitute different types of information within the knowledge graph. This results in a very precise and detailed annotation of the data table, which in turn allows for a very precise and detailed conversion of the data table into text. Furthermore, such an embodiment improves the quality of the table data, as incomplete or incorrect data is corrected and completed. Thus, the quality of the table's conversion into text, and therefore the response provided to the user, is improved.

[0028] According to another particular embodiment, obtaining additional information on said column of the semantically annotated data table may include extracting, in said knowledge graph, knowledge related to said entity identifier, said entity type identifier or said property identifier which was associated during the semantic annotation of at least some of the data in the table.

[0029] Such an embodiment makes it possible to enrich the data in at least one column of the table with additional information from the knowledge graph, thereby improving the quality and depth of the data available in the table. Furthermore, by associating specific entities and well-defined properties, the table data becomes more precise and contextualized, thus facilitating a better understanding and use of the information contained in the table, particularly for converting the table into text.

[0030] According to another particular embodiment, obtaining additional information on said column of the semantically annotated data table may further include an association, using a type relationship, of the additional information extracted in connection with said entity identifier to the additional information extracted in connection with said entity type identifier.

[0031] Such an embodiment makes it possible to optimize the semantic consistency of the data contained in a column thanks to:

[0032] - a data alignment that ensures that the cell annotations in the columns are consistent with the expected data type for that column, thus reinforcing the semantic structure of the table,

[0033] - an explicit typing that clarifies the data type for the column, reducing the ambiguities and errors of interpretation.

[0034] Such an embodiment also makes it possible to optimize the accuracy of the annotations thanks to:

[0035] - the use of a typing context that provides an additional context for the cell annotations, improving the accuracy and relevance of these annotations,

[0036] - a reduction of conversion errors, using as a reference the Additional information is extracted in connection with the entity type identifier to detect and correct potential errors in cell annotations. This implementation thus contributes to improving the quality of the data table's conversion to text, and therefore the quality of the response provided to the user.

[0037] According to another particular embodiment, obtaining additional information on a column of said pair of the semantically annotated data table may include associating additional information extracted in connection with an entity identifier for said column of said pair with additional information extracted in connection with said or another entity identifier for the other column of said pair.

[0038] Such an embodiment makes it possible to further optimize the consistency of the annotations:

[0039] - by ensuring that cell annotations are consistent not only within each column, but also between linked columns,

[0040] - thanks to a strengthening of semantic structures, due to the establishment of explicit relationships between the entities in the columns of the pair.

[0041] Such an embodiment also makes it possible to optimize the enrichment of the semantic annotations of the table:

[0042] - by providing an additional relational context for the annotations of the cells, thus enriching the data with information on the relationships between the entities extracted from the knowledge graph,

[0043] - by augmenting the annotations with additional knowledge from relationships identified between the columns.

[0044] Such an embodiment thus contributes to improving the quality of the conversion to text of the data table, and therefore the response provided to the user.

[0045] According to another particular embodiment, obtaining additional information on said column or on said column of said pair of the semantically annotated data table may include an association with additional knowledge extracted from the knowledge graph, additional information extracted in connection with said entity identifier and / or information additional information extracted in connection with said entity type identifier and / or additional information extracted in connection with said property identifier.

[0046] Such an embodiment makes it possible to improve the enrichment of the table data:

[0047] - by enriching the semantic annotations with additional information derived from the knowledge graph, thus offering a more complete and detailed view of the data,

[0048] - by adding contextual knowledge that can improve the understanding and interpretation of data.

[0049] Such an embodiment also makes it possible to improve the accuracy and relevance of the information contained in the table, thanks in particular to the fact that the additional knowledge from the knowledge graph makes it possible to refine and specify the annotations, thereby increasing their relevance, to the benefit of a reduction of ambiguities and a clarification of the meanings of entities and relations.

[0050] Such an embodiment also makes it possible to significantly improve the quality of the data, and therefore the response provided to the user. Indeed, the additional knowledge from the knowledge graph improves the consistency and reliability of the annotations, thus increasing the overall quality of the data. Moreover, the addition of this additional knowledge facilitates the detection of inconsistencies and anomalies in the table data, allowing for quick and efficient corrections.

[0051] Thanks to such an embodiment, the quality of the conversion to text of the data table is optimized, as is the accuracy and level of precision of the textual information resulting from this conversion.

[0052] According to another particular embodiment, at least some of the data in the table that are semantically annotated are character strings.

[0053] Such an embodiment allows semantic annotation to be targeted at the truly relevant data in the table, namely the character strings. Indeed, character strings often carry rich and varied textual information, making them particularly relevant for semantic annotation. By considering only character strings, the communication process avoids the unnecessary complexity associated with annotating less informative data types, such as numbers, dates, identifiers, etc. The implementation of the communication process is therefore simplified, more resource-efficient, and faster.

[0054] Furthermore, such an embodiment makes it possible to optimize the reduction of annotation errors, as character strings can be more easily associated with Knowledge graph entities are used because character strings contain rich textual information, such as descriptions, names, categories, and other textual details that are essential for semantic annotation. This results in a more accurate matching with the knowledge graph entities.

[0055] The various modes or embodiments mentioned above can be added independently or in combination with each other to the communication process as defined above.

[0056] The invention also relates to a computer system comprising a knowledge base modeling data in the form of a knowledge graph, the computer system being configured to implement:

[0057] - receiving a question posed using a natural language interface,

[0058] - generating an answer to said question using a language model, said generated response comprising textual data from a textual conversion of a data table, during which data from the table is enriched using data from the knowledge graph.

[0059] Such a device is notably configured to implement the aforementioned communication process, according to one or the other of its embodiments.

[0060] The invention also relates to a computer program comprising instructions for implementing the communication method according to the invention, according to any one of the particular embodiments described above, when said program is executed by a processor.

[0061] Such instructions can be stored permanently in a non-transient memory medium of the computer device implementing the communication method according to the invention.

[0062] This program may use any programming language, and be in the form of source code, object code, or code intermediate between source code and object code, such as in a partially compiled form, or in any other desirable form.

[0063] The invention also relates to a recording medium or information medium readable by a computer, and comprising instructions for a computer program as mentioned above.

[0064] The recording medium can be any entity or device capable of storing the program. For example, the medium can include a storage means, such as a ROM, for example a CD ROM or a microelectronic circuit ROM, or a magnetic recording means, for example a mobile medium, a hard disk drive or an SSD.

[0065] On the other hand, the recording medium can be a transmissible medium such as an electrical or optical signal, which can be transmitted via an electrical or optical cable, by radio, or by other means, so that the computer program it contains is executable remotely. The program according to the invention can, in particular, be uploaded to a network, for example, an Internet-type network.

[0066] Alternatively, the recording medium may be an integrated circuit in which the program is incorporated, the circuit being adapted to execute or to be used in the execution of the aforementioned communication process.

[0067] According to one embodiment, the present technique is implemented using software and / or hardware components. In this context, the term "device" or "module" may refer in this document to a software component, a hardware component, or a set of hardware and software components. Brief description of the drawings

[0068] Other features and advantages will become apparent upon reading particular embodiments of the invention, given by way of illustrative and non-limiting examples, and the accompanying drawings, among which:

[0069] [Fig-1] Fig. 1 represents an architecture in which the communication method, according to a particular embodiment of the invention,

[0070] [Fig.2] The [Fig.2] represents a computer device according to a particular embodiment of the invention, as implemented in the architecture of the [Fig.1],

[0071] [Fig.3] Fig.3 represents the main actions implemented in the communication process, according to a particular embodiment of the invention, as implemented in the architecture of [Fig.1].

[0072] [Fig.4] The [Fig.4] represents the main actions implemented within the framework of the execution of a semantic annotation action in the communication process, according to a particular embodiment of the invention, as implemented in the architecture of [Fig.1],

[0073] [Fig.5A] The [Fig.5A] represents an example data table used in the communication process as implemented in the architecture of the [Fig.1],

[0074] [Fig.5B] [Fig.5B] represents an example of a data table resulting from a semantic annotation of the data table of [Fig.5A],

[0075] [Fig.5C] [Fig.5C] represents an example of a data table resulting from a semantic enrichment of the semantically annotated data table of [Fig.5B],

[0076] [Fig.6] [Fig.6] represents an example of a partial representation of an entity in a knowledge graph used in the communication process as implemented in the architecture of [Fig.1],

[0077] [Fig.7] Fig.7 represents the main actions implemented within the framework of the execution of a data enrichment action of the table in [Fig.5B], according to a particular embodiment of the invention, as implemented in the architecture of [Fig.1],

[0078] [Fig.8A] Fig.8A represents an example of a prompt generated during the setup operation of the communication process as implemented in the architecture of [Fig.1],

[0079] [Fig.8B] [Fig.8B] represents an example of a response to the prompt of [Fig.8A], which is generated during the implementation of the communication process as implemented in the architecture of [Fig.1].

[0080] [Fig.9] Fig.9 represents an example of a prompt and a response to that prompt generated using a state-of-the-art communication process.

[0081] Detailed description of an embodiment of the invention

[0082] Fig. 1 represents an architecture in which a communication method is implemented according to an embodiment of the invention.

[0083] Such an architecture comprises:

[0084] - a computer device DI comprising a knowledge base BC modeling data in the form of a knowledge graph (KG);

[0085] - a human-machine interface (HMI) configured to be activated by a user (UT) to communicate with the DI computer system.

[0086] The human-machine interface (HMI) is contained within a communication terminal (CT) which is configured to communicate with the computing device (CD) via a communication network (RC), such as, for example, an Internet, Intranet, etc. The telecommunication terminal (CT) is, for example, a computer, a smartphone, a tablet, etc.

[0087] The DI computer system may include, for example, a computer, a server, a platform, etc.

[0088] In [Fig. 1], the knowledge base BC is integrated into the computer system DI. Of course, such a knowledge base BC can be separate from the computer system DI, the latter then being configured to communicate with the knowledge base BC by any suitable means. The knowledge graph GC is, for example, of the Wikidata™, DBpedia™, Google Knowledge Graph™, Microsoft Concept Graph™, etc. type.

[0089] The UI interface may include, for example, a text-based graphical interface or a sound sensor coupled with a speech recognition interface.

[0090] Such a user interface communicates with a NAV Internet browsing engine. The NAV Internet browsing engine includes a programming interface Web-based INT application allowing the UT user of the TC terminal to communicate in natural language with the DI computer device.

[0091] We will now describe, with reference to [Fig.2], the simplified structure of the DI computer device.

[0092] According to the invention, the DI computer device comprises:

[0093] - a COM communication module configured to, on the one hand, receive queries generated using the UI interface, said query comprising a question in natural language posed by the user UT, and on the other hand, sending responses in natural language to these queries to the communication terminal TC,

[0094] - a CMD command module configured to control the restoration of natural language responses to generated queries, in the form of a web page, PW,

[0095] - an ENR module for enriching the data of a TD data table using of knowledge graph (GC) data, with the ENR module configured to generate an enriched data table (TDE),

[0096] - a CNV module for converting enriched data tables (TDEs), in the form of a TXT text.

[0097] The ENR enrichment module comprises:

[0098] - an ANS submodule for semantic annotation of the TD data table, said The ANS submodule is configured to annotate at least some of the data in the TD table using data from the GC knowledge graph, in order to obtain a semantically annotated data table TDA.

[0099] - an EAS sub-module for enriching all or part of said at least certain The EAS submodule is configured to obtain, for at least one column of the semantically annotated data table (TDA), additional information about the semantically annotated data, or a portion thereof, from the knowledge graph (GC) data, and to supplement the semantically annotated data belonging to at least one column, or a portion thereof, with the additional information obtained. Following such enrichment, an enriched data table (TDE) is generated.

[0100] The CNV data table text conversion module includes:

[0101] - a GPR prompt generation submodule configured to generate automatically, that is, without user intervention, a PRP prompt in natural language,

[0102] - an LNT natural language model configured to be queried from the prompt generated. Such a natural language model can be, for example, an n-grams model, a recurrent neural network (RNN), a large language model (LLM), etc. The natural language model (LNT) has been classically trained to accumulate knowledge.

[0103] Said PRP prompt contains a request to generate a textual description of the enriched data table (TDE), based on the semantically annotated data, or a portion thereof, which has been supplemented by the additional information obtained. Said LNT language model, queried from the PRP prompt, is configured to generate a TXT textual description of the data table, such description advantageously reproducing the content of the table faithfully, and adding further details and clarifications.

[0104] According to the invention, the CMD command module is configured in particular to control the rendering of a Web page containing the TXT textual description, via the user UI interface UT.

[0105] Optionally, said EAS submodule is further configured to obtain additional information on said semantically annotated data, or a part thereof, for at least said column of the enriched data table (TDE), from data originating from at least one external knowledge source (SCE), such as, for example, a database (BD), a corpus (CD) of documents of various types and / or formats (texts, tables, etc.), a web resource (RW), etc. Obtaining such additional information is implemented using, for example, a Retrieval-Augmented Generation (RAG) technique. Since obtaining such additional information is optional, it is represented by dashed lines in [Fig. 2].

[0106] At initialization, the code instructions of the computer program PG are, for example, loaded into RAM (not shown) before being executed by the PROC processor. The PROC processor of the UTR processing unit implements, in particular, the following actions, within the framework of the communication process that will be described below, according to the instructions of the computer program PG:

[0107] - receive a question posed using the natural language UI interface,

[0108] - enrich data in a TD data table using data from the graph of GC knowledge, such enrichment including:

[0109] — an annotation of at least some of the data in table TD using GC knowledge graph data,

[0110] — an enrichment of all or part of said semantically annotated data, based on data from the GC knowledge graph, and possibly from data from said at least one external knowledge source SCE,

[0111] - convert the enriched data table TDE into text format, said conversion including the following:

[0112] — automatically generate a PRP prompt in natural language, asking the natural language model LNT to generate a textual description of the enriched data table TDE, from the semantically annotated data, or a part thereof, which have been supplemented by the additional information obtained,

[0113] — generate by the natural language model LNT an enriched textual description TXT file of the TD data table,

[0114] - command the rendering of a PW web page containing the textual description enriched TXT.

[0115] We now describe, in relation to [Fig.3], together with Figures 1 and 2, the main steps of a communication method with the computer device DI, according to a particular embodiment of the invention.

[0116] During a step SI, the user UT sends a request Q to the computer device DI, via the IU interface, said request Q being able to include a question on a given subject, a request to summarize a data file attached as an appendix to the question Q, a request to convert a data table TD attached as an appendix to the question Q verbatim, etc. This request is received in S2 by the computer device DI, via its COM module.

[0117] The information contained in the query Q may include one or more words. Such a query may be written in natural language or in a particular computer language, for example SQL (Structured Query Language), Python, etc.

[0118] In a non-exhaustive example of an embodiment, the query Q may be the following: "could you give me information on the drug Valsartan™?".

[0119] In another non-exhaustive embodiment example, the query Q may be as follows: "Here is a data table TD containing information about drugs. Could you provide me with a textual description of this table?".

[0120] During step S3, the ENR enrichment module of the DI computing system enriches the data in the TD data table using data from the knowledge graph GC, and optionally using data from the external knowledge source SCE. An enriched TDE data table is then generated at the end of step S3. If the TD data table was provided as an appendix to question Q, the enrichment is performed directly from this table. If the TD data table was not provided as an appendix to question Q, the DI computing system can retrieve the TD data table from the learned data in the knowledge base BC, from the external knowledge source SCE, or from the RW web resources.

[0121] During an S4 step, the CNV conversion module of the DI computer device performs a textual conversion of the enriched data table TDE in order to generate a TXT textual description of the latter.

[0122] During a step S5, the computer device DI sends a response R to the question Q. For this purpose, the control module CMD of the computer device DI commands the rendering, at the level of the communication terminal TC of the user UT, of a Web page PW containing the textual description TXT.

[0123] The TC terminal of user UT receives this web page in S6 which is then rendered in S7 using the UI interface.

[0124] With reference to [Fig.4], together with Figures 1 to 5, we now describe the detailed process of the enrichment step S3, according to one embodiment of the invention.

[0125] An example of a TD data table is shown in [Fig. 5A]. Such a table contains six rows L0 to L5 and five columns Cl to C5. In the example shown, the TD table contains information on five drugs listed respectively in rows L1 to L5 of the TD data table. Columns Cl to C5 are labeled, in row L0, respectively, "Name", "Product", "ephMRA Code" (English abbreviation for "European Pharmaceutical Market Research Association"), "ephMRA Class", and "ATC Class" (English abbreviation for "Anatomical Therapeutic Chemical Classification System"). It goes without saying that the number of rows and columns is not limited to those in [Fig. 5A], and that the TD table could contain information on other, non-exhaustive subjects, for example, information on a company's subsidiaries, information on the vehicles in the catalog of a particular vehicle brand, etc.The data in the TD table can be in various possible formats such as, for example, CSV format (English abbreviation for "comma separated values"), Excel XLS binary format, TSV format (English abbreviation for "Tabulation-separated values"), etc.

[0126] The enrichment step S3 includes a primitive typing substep S30 in which the ANS semantic annotation submodule of the DI computing device associates at least one column of the TD table with a type from among a set of primitive types including a string, an identifier, a date, an integer, a floating-point number, etc.

[0127] In the example in [Fig.5A], columns Cl, C2, C4 and C5 contain character strings and column C3 contains identifiers.

[0128] At the end of substep S30, only columns Cl, C2, C4 and C5 are selected to be semantically annotated.

[0129] During a substep S31, a semantic annotation of some or all of the data in columns C1, C2, C4 and C5 is implemented, such A substep comprising, for at least one cell and / or at least one column and / or at least one pair of columns in the TD table, the extraction of at least one identifier from an entity in the GC knowledge graph. An entity could be, for example, the name of an object, a person, a company, etc. In the example shown in [Fig. 5B], the GC knowledge graph used is Wikidata™ https: / / wikidata.org.

[0130] More specifically, in the example shown in [Fig.5B], such a semantic annotation is of the type:

[0131] - CEA cell-entity annotation (English abbreviation for "Cell Entity Annotation"),

[0132] - CTA column-type annotation (English abbreviation for "Column Type Annotation") »),

[0133] - and column-property annotation CPA (English abbreviation for "Column Property") Annotation").

[0134] To this end, the ANS semantic annotation submodule proceeds:

[0135] - in S310, to a CEA association of at least one cell from the TD table to a entity identifier extracted from the GC knowledge graph,

[0136] - in S311, to a CTA association of at least one column of the TD table to a entity type identifier extracted from the GC knowledge graph,

[0137] - in S312, to a CPA association of at least one pair of columns from the TD table to a property identifier of the knowledge graph GC.

[0138] Other types of semantic annotations can be used as an alternative, such as, for example, the annotation of the lines in their entirety or the thematization information on the table.

[0139] During substep CEA S310, each character string, or some of them, present in columns C1, C2, C4, and C5, with the exception of column headers, is traversed and associated with a value from a corresponding entity in the knowledge graph GC via heuristic techniques combining syntactic distance calculations, table / graph context comparisons, etc., or by machine learning. Thus, for example, in the example in [Fig. 5A]:

[0140] - the cell “VALSARTAN HYDROCHLOROTHIAZIDE CRISTERS” is associated with entity ID Q48566694 "hydrochlorothiazide / valsartan" in the GC knowledge graph and available at https: / / wikidata.org / wiki / Q48566694,

[0141] - the cell “ESOMEPRAZOLE BIOGARAN” column Cl, row L5, is associated with the entity identifier Q553223 "esomeprazole" in the GC knowledge graph and available at the following address: https: / / wikidata.org / wiki / Q553223,

[0142] - the cell “CGP-48933” column C2, row L5, is associated with the entity identifier Q155472 “valsartan” in the GC Knowledge Graph and available at the following address: https: / / www.wikidata.org / wiki / Q155472,

[0143] - etc.

[0144] With reference to [Fig. 5B], an extract from a semantically annotated TDA table is shown, in which the CEA annotations are indicated. As is known per se, for each added entity identifier, a confidence score for the annotation is indicated. In the example shown in [Fig. 5B]:

[0145] - a confidence score of 12% is associated with the cell "VALSARTAN HYDROCHLOROTHIAZIDE CRISTERS”,

[0146] - a confidence score of 14% is associated with the cell “ESOMEPRAZOLE BIOGARAN,

[0147] - a confidence score of 12% is associated with the cell “CGP-48933”,

[0148] - etc.

[0149] During substep CTA S311, for at least one, part, or some of the columns Cl, C2, C4, C5, based on the CEA annotations performed in S310, the heading of said at least one of the columns Cl, C2, C4, C5 is associated with an identifier of an entity in the knowledge graph GC that represents the column type. For this purpose, majority votes on the types of the CEA entities are implemented in a manner known per se, as described, for example, in the document Chabot et al: “DAGOBAH: An End-to-End Context - EREE Tabular Data Semantic Annotation System,” or proximity calculations in a vector space are implemented, as described, for example, in the document Liu et al: “Radar Station: Using KG Embeddings for Semantic Table Interpretation and Entity Disambiguation.” Thus, for example, in the example of [Fig. 5A]:

[0150] - the heading "NAME" in column Cl is associated with entity identifier Q12140 drug” in the GC knowledge graph and available at the following address: https: / / www.wikidata.org / wiki / Q12140,

[0151] - the heading "PRODUCT" in column C2 is associated with the entity identifier Q28885102 "pharmaceutical product" in the GC knowledge graph and available at https: / / www.wikidata.org / wiki / Q28885102,

[0152] - etc.

[0153] With reference to [Fig. 5B], CTA annotations are indicated in the extract of the semantically annotated TDA table. As with CEA annotations, completeness rates are also indicated for CTA annotations.

[0154] During substep CPA S312, the ANS semantic annotation module identifies properties from the knowledge graph GC between at least two columns of the TD table, using well-known majority voting techniques mentioned above. Thus, for example, in the example in [Fig. 5A]:

[0155] - the property identifier P527 "includes" in the GC knowledge graph and available at the following address https: / / www.wikidata.Org / wiki / Property:P527 is identified between columns C1 "NAME" and C2 "PRODUCT",

[0156] - the property identifier P527 is also identified between columns Cl "NAME and C5 "ATC Class".

[0157] With reference to [Fig.5B], such CPA annotations are represented in the semantically annotated TDA table extract.

[0158] With further reference to [Fig.4], the enrichment step S3 further includes a substep S32 in which, for at least one of the columns Cl, C2, C4, C5 semantically annotated in the TDA table of [Fig.5B], the EAS module for enriching semantic annotations implements the following substeps:

[0159] - in S320, obtain, from the knowledge graph GC and possibly from from the external knowledge source SCE, additional IC information on at least one of the CEA, CTA, CPA annotations related to said at least one column,

[0160] - in S321, complete at least one of the CEA, CTA, CPA annotations with the Additional IC information obtained.

[0161] At the end of substep S321, an enriched data table TDE is generated.

[0162] More specifically, during the implementation of substep S320, the following procedure is performed: Starting from at least one identifier in the semantically annotated TDA table, the process retrieves additional knowledge triplets describing the entities, entity types, and properties of the TDA table from the GC knowledge graph. The advantage of such an operation is to determine, from the GC knowledge graph, a sufficient number and quality of additional knowledge triplets.

[0163] In the example of [Fig.5B], for example for column C2, the CEA annotation relating to the identifier Q155472 added to row L5 corresponds to an entity called "Valsartan" in the knowledge graph GC, the term "Valsartan" constituting additional IC information to be added to the identifier Q155472. From the identifier Q155472, other additional IC information can be retrieved.

[0164] For example, the entity corresponding to identifier Q155472 is partially represented in [Fig. 6]. In this example, additional IC information includes, for example, the knowledge triplet: "Valsartan", "is a", "medicine".

[0165] In order for such an enrichment of semantic annotations to be as complete and detailed as possible, while limiting ambiguities, the enrichment of these annotations is implemented according to at least a first dimension which associates each CEA annotation to the CTA annotation of its column of belonging via a type relationship.

[0166] In order to optimize this enrichment of semantic annotations, the enrichment of these annotations can be implemented according to a second dimension associating the CEA annotations of one column with the CEA annotations of another column linked to this column by an identified CPA annotation.

[0167] In order to further optimize this enrichment of semantic annotations, the enrichment of these annotations is implemented according to a third dimension which associates the CEA / CTA / CPA annotations with knowledge from the GC knowledge graph.

[0168] In the non-exhaustive example in [Fig. 5C], several entities from the knowledge graph GC, as supplementary information IC, are retrieved during the semantic annotation enrichment step. For better readability and to allow the LNT language model to use this information correctly, the entity labels are retrieved and used instead of identifiers as follows:

[0169] - https: / / wikidata.org / wiki / Q48566694 = “hydrochlorothiazide / valsartan”,

[0170] - https: / / wikidata.org / wiki / Q553223 = “esomeprazole”,

[0171] - https: / / www.wikidata.org / wiki / Q155472 = “valsartan”,

[0172] - https: / / www.wikidata.org / wiki / Q12140 = "medicine",

[0173] - https: / / www.wikidata.org / wiki / Q28885102 = "pharmaceutical product",

[0174] - https: / / www.wikidata.Org / wiki / Property:P527 = "includes",

[0175] - etc.

[0176] Thus, the additional IC information obtained during the enrichment of semantic annotations according to the first dimension is used to complete the semantic annotations, for example in the following way:

[0177] -<hydrochlorothiazide / valsartan, est un, médicament> ,

[0178] - <esomeprazole is a pharmaceutical product>,

[0179] - cvalsartan, is a pharmaceutical product

[0180] - etc.

[0181] The additional IC information obtained during the enrichment of semantic annotations according to the second dimension is used to supplement the semantic annotations, for example in the following way:

[0182] -<hydrochlorothiazide / valsartan, comprend, valsartan> .

[0183] The additional IC information obtained during the enrichment of semantic annotations according to the third dimension is used to supplement the semantic annotations, for example in the following way:

[0184] - <esomeprazole, mass, 345.114712 dalton>,

[0185] - <esomeprazole, chemical formula, CitHwNsOsS^

[0186] - cvalsartan, mass, 435.22704 dalton>,

[0187] - cvalsartan, chemical formula, C24H29NsO3>.

[0188] With reference to [Fig.7], together with Figures 1, 2 and 4, we now describe the detailed sequence of the text conversion step S4, according to one embodiment of the invention.

[0189] The text conversion step S4 includes a substep S40, during which the prompt generation module GPR automatically generates a prompt PRP requesting the language model LNT to generate a textual description of the data table TD. According to the invention, the PRP prompt includes a verbalization instruction taking as input the semantic annotations that were enriched in S3.

[0190] An example of such a PRP prompt is shown in [Fig. 8A]. The "DATA" tag is intended to be replaced at runtime by the contents of the TD table. The "ANNOTATION" tag is intended to be replaced at runtime by the additional IC information contained in the enriched TDE table.

[0191] The text conversion step S4 includes a substep S41, during which the PRP prompt is subjected to the LNT language model.

[0192] The text conversion step S4 includes a substep S42, during which the LNT language model generates a TXT text description of the TD table. An example of such a text description is shown in [Fig.8B].

[0193] Such an S4 step of textual conversion of tabular data proves to be significantly more efficient than a classic textual conversion of tabular data, such as represented in [Fig.9], which could be requested by a user prompting a conversational agent on the market.

[0194] The communication method described above represents a significant advance in the field of natural language dialogue applications between a user and a machine, offering a precise and efficient solution, particularly in terms of the quality of the text conversion performed, for converting tabular data into detailed and unambiguous natural language text. Verbalizing tabular data notably improves data accessibility and comprehension for a wide range of users. The invention has numerous applications, including: - the generation of automatic reports on tabular data with a rich description of their content,

[0195] - the generation of automatic summaries for presentations or reports,

[0196] - access to table data in question-answering systems based on the RAG,

[0197] - etc.

Claims

Demands

1. Method of communicating with a computer device (CD) comprising a knowledge base (KB) modeling data in the form of a knowledge graph (KG), said method comprising the following, at the level of said device: - receiving (S2) a question posed using a natural language interface, - generating (S4-S5) an answer to said question using a language model (LNT), said generated answer comprising textual data (TXT) resulting from a textual conversion (S4) of a data table, during which data from the table are enriched (S3) using data from the knowledge graph.

2. A communication method according to claim 1, wherein the enrichment of the table data comprises the following: - semantically annotating (S31) at least some of the table data using data from the knowledge graph, - for at least one column of the table (TDA) comprising semantically annotated data, obtaining (S320) additional information (CI) on said semantically annotated data from the knowledge graph data, and supplementing (S321) said semantically annotated data of said at least one column with the supplementary information obtained, and the textual conversion comprises the following: - querying (S41) said language model from a prompt (PRP) generated (S40) in natural language by said device, said prompt requesting the language model to generate a textual description of the table from the semantically annotated table data that has been supplemented,- receive (S6) the said textual description generated in response to the query.

3. A communication method according to claim 2, wherein the semantic annotation (S31) of at least some of the data in the table comprises, for at least one cell and / or at least one column and / or at least one pair of columns in the data table, an extraction, in the knowledge graph, of at least one identifier of an entity in the knowledge graph.

4. A communication method according to claim 3, wherein the semantic annotation (S31) of at least some of the data in the table comprises the following: - associating at least one cell of the data table with an entity identifier extracted from the knowledge graph, - associating at least one column of the data table with an entity type identifier extracted from the knowledge graph, - associating at least one pair of columns of the table with a property identifier from the knowledge graph.

5. A communication method according to any one of claims 3 or 4, wherein the obtaining (S320) of additional information on said column of the semantically annotated data table comprises an extraction, in said knowledge graph, of knowledge related to said entity identifier, said entity type identifier or said property identifier which was associated during the semantic annotation of at least some of the data in the table.

6. A communication method according to claim 5, wherein the obtaining (S320) of additional information on said column of the semantically annotated data table further comprises an association, using a type relationship, of the additional information extracted in connection with said entity identifier to the additional information extracted in connection with said entity type identifier.

7. A communication method according to claim 5 or claim 6, wherein the obtaining (S320) of additional information on a column of said pair of the semantically annotated data table comprises an association of the additional information extracted in connection with an entity identifier for said column of said pair with additional information extracted in connection with said or another entity identifier for the other column of said pair.

8. A communication method according to any one of claims 5 to 7, wherein obtaining (S320) additional information about said column or about said column of said pair of the semantically annotated data table comprises an association with additional knowledge extracted from the knowledge graph, supplementary information extracted in connection with said entity identifier and / or additional information extracted in connection with said entity type identifier and / or additional information extracted in connection with said property identifier.

9. A communication method according to any one of claims 1 to 8, wherein at least some of the data in the table that are semantically annotated are character strings.

10. Computer device (CD) comprising a knowledge base (KB) modeling data in the form of a knowledge graph (KG), the computer device being characterized in that it is configured to implement: - receiving a question posed using a natural language interface, - generating an answer to said question using a language model (LNT), said generated answer comprising textual data from a textual conversion of a data table, during which data from the table are enriched using data from the knowledge graph.

11. Computer program comprising program code instructions for implementing the communication method according to any one of claims 1 to 9, when executed on a computer.

12. Computer-readable information carrier, and containing instructions for a computer program according to claim 11.

Images