• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    geïmplementeerde werkwijze voor het gericht zoeken in een databank en het rapporteren van relevante octrooidocumenten om op die manier de octrooibaarheid van een concept te onderzoeken. De computer-geïmplementeerde werkwijze omvat volgende opeenvolgende stappen: - Het inlezen van een basistekst die het concept beschrijft; - Het bepalen van de IPC classificatie; - Verwerken van de basistekst in zoekmachinetaal; - Het zoeken in een databank naar één of meerdere elektronische octrooidocumenten met de corresponderende IPC classificatie; - Het rapporteren van gegevens van de genoemde één of meerdere octrooidocumenten aan de gebruiker, waarbij neurale netwerken met woordvectoren worden gebruikt om het zoekproces te optimaliseren. In het rapport worden een reeks kenmerken uit de basistekst op een overzichtelijke en interactieve manier afgetoetst per genoemd elektronisch octrooidocument, om op die manier de octrooieerbaarheid van een concept ten opzichte van de gevonden octrooidocumenten te onderzoeken. The present invention relates to a computer fig. I implemented method for targeted searching in a database and reporting relevant patent documents in order to investigate the patentability of a concept. The computer-implemented method comprises the following consecutive steps: - Reading in a basic text describing the concept; - Determining the IPC classification; - Processing the basic text in search engine language; - Searching in a database for one or more electronic patent documents with the corresponding IPC classification; - Reporting data from said one or more patent documents to the user, whereby neural networks with word vectors are used to optimize the search process. In the report, a series of characteristics from the basic text are checked in a clear and interactive way for each electronic patent document mentioned, in order to investigate the patentability of a concept in relation to the found patent documents.
    公开号:BE1022868B1
    申请号:E2015/5571
    申请日:2015-09-14
    公开日:2016-09-29
    发明作者:Johan Brants；Sabine Eeckhaoudt；Cauteren Tim Van；Ellen Crabbe；Rim Ghedira；Stijn Monsaert；Pieter GREGOIR；Ken HOUBEN；Jenthe TACK；Joachim VANDERBEKE；Eline DEPOORTERE；Den Berge Wendy Van；Lien DUPONT；Cate Swift；DE MOORTEL Dennis VAN；Dina VERGEYLEN；Rynck Jelle De
    申请人:Brantsandpatents Bvba；
    IPC主号:
    专利说明:

    COMPUTER IMPLEMENTED METHOD, SYSTEM AND COMPUTER PROGRAM PRODUCT FOR DIRECTLY SEARCHING IN A DATABASE AND EXAMINING THE PATENTABILITY OF A CONCEPT
    TECHNICAL DOMAIN
    The invention relates to a computer-implemented method, system and computer program product for the targeted search in a database, the reporting of relevant patent documents and the investigation of the patentability of a concept, more particularly for providing a well-arranged and compact report document with detailed data from relevant patent documents.
    BACKGROUND ART
    It is known to search for data from patent documents in databases, where a user enters search terms and the search engine reports various patent documents to the user.
    A problem with the known search methods is that they are hardly accessible to the majority of researchers or inventors, because they have difficulty translating the technical terms from their field to the partially legal professional jargon used in patents. This often deter researchers from testing if their idea or invention is eligible for a patent, which can lead to missed opportunities and the loss of the rights to their invention. Often people want to be able to make a quick judgment about the patentability of a concept, but the large amounts of data in the search results confuse the user quickly, so that they lose the overview and the time to make a quick judgment increases enormously. Moreover, the current software does not indicate the characteristics of their invention that differ and / or resemble the existing patent documents.
    In current search algorithms, many techniques are used to optimize the search process, including assigning weights to words, tracing words to their stem and using n-grams of at least 2 words. The embedding of words ("word embedding") in structures based on vectors, in which neural networks are trained to image words on vectors of real numbers in a low-dimensional continuous space, is thereby overlooked, despite the fact that it is precisely that method that can make the link between the basic text describing the concept and the database of patent documents, provided the correct training for the neural network.
    It is an object of the present invention to find a solution to at least some of the aforementioned problems in examining patentability and the targeted search and reporting of relevant patent documents.
    There is a need for a more adapted form for investigating patentability, whereby targeted searches can be made in a database on the basis of a text that describes the concept on the basis of a few characteristics. There is also a need for user-friendly and interactive software that displays the relevant patent documents in a clear and easy-to-interpret manner, whereby the patentability of the concept is checked against the available State of the art.
    The invention contemplates an improved method which offers a solution to at least one of the aforementioned disadvantages in the investigation of patentability and the targeted search and reporting of relevant patent documents, as defined by claim 1. The present invention also has an improved system and an improved program product as its object. as defined by claims 11 and 12 respectively.
    SUMMARY OF THE INVENTION
    The invention relates in a first aspect to a computer-implemented method for targeted searching in a database and reporting relevant patent documents, in order to determine the patentability of a concept on the basis thereof. The computer-implemented method comprises the following consecutive steps:
    Reading in a basic text describing the concept on the basis of a preamble, which outlines the State of Technology on the basis of old related inventions, and, a text describing the alleged new and inventive features of the concept Determining the IPC classification - Processing the basic text in search engine language
    Searching a database for one or more electronic patent documents
    Reporting data from said one or more patent documents to the user, whereby neural networks with word vectors are used to optimize the search process and a series of characteristics from the basic text are checked in a clear and interactive way per said electronic patent document, for that purpose examine the patentability of a concept in relation to these documents.
    In a second aspect, the invention relates to a system for querying a database and examining the patentability of a concept, which system comprises a central processing unit, which is configured to carry out the computer-implemented method as defined by claims 1- 10.
    In a further aspect the invention relates to a computer program product for querying a database, the targeted search and reporting of relevant patent documents with the aim of examining the patentability of a concept, which computer program product comprises at least one computer-readable medium on which computer readable program code portions are stored, which program code portions include instructions for performing the computer-implemented method as defined by claims 1-10.
    DESCRIPTION OF THE FIGURES
    Figure 1 shows a table illustrating a preferred form of a result of the method according to the present invention.
    This table comprises a column that lists the various characteristics Kl-Km (1) from the basic text and a row that arranges different patent documents D1-Dn (2) according to relevance. Dl-Dn contain a hyperlink to the patent document from the database. In each table cell (3) is displayed if a specific feature occurs in patent document D *, where a hyperlink is provided to the page and line (p, I) where the specific feature occurs in the patent document. The rows and columns can be arranged interactively, with the rest of the table automatically adjusted.
    DETAILED DESCRIPTION
    The invention relates to a computer-implemented method for specifically searching a database and reporting relevant patent documents, in order to investigate the patentability of a concept. "A", "de" and "het" in this document refer to both the singular and the plural unless the context clearly assumes otherwise. For example, "a segment" means one or more than one segment.
    The terms "include", "comprising", "consist of", "consisting of", "provided with", "contain", "containing", "include", "including", "contents", "contents" are synonyms and are inclusive or open terms indicating the presence of what follows, and which do not preclude or preclude the presence of other components, features, elements, members, steps, known from or described in the prior art.
    Unless defined otherwise, all terms used in the description of the invention, including technical and scientific terms, have the meaning as generally understood by those skilled in the art of the invention. For a better assessment of the description of the invention, the following terms are explicitly explained.
    A neural network is software that is used for simulating, researching and developing. It applies artificial neural networks and adapts software concepts derived from biological neural networks. Furthermore, it can be used in a wider range of adaptive systems such as artificial intelligence and machine learning.
    A bigram is a combination of two words. An n-gram is a combination of n words, with a random natural number, in most cases less than 10.
    The embedding of words ("word embedding") is a collective name for a series of techniques for modeling a language and learning language functions, where words of the vocabulary (and possibly sentences thereof) are mapped onto vectors of real numbers in a low-dimensional space relative to the extensiveness of the vocabulary. Various methods are known for this image, including neural networks, dimensionality reduction of the matrix describing the occurrence of words together, and, explicit representation depending on the context in which the words occur.
    An algorithm in mathematics and computer science is a stand-alone step-by-step series of operations that are performed. There are algorithms for performing calculations, data processing and automated reasoning.
    A hyperlink (or simply link) - with a Dutch word connection or link - is a computer and internet term that indicates a reference (reference) in a hypertext (for example a website) that the user can follow.
    A computer is a device with which data can be processed according to formal procedures (algorithms). Usually the word computer means an electronic, digital device, but there are also mechanical and analogue computers. In addition, a computer can work in different number systems such as the decimal (tensile) or binary (bivalent) system. The current computers all work in the binary system, but a quantum computer is also possible.
    In a first aspect the invention relates to a computer-implemented method for querying a database and the targeted search and reporting of relevant patent documents, which computer-implemented method comprises the following consecutive steps: - reading in a basic text comprising a preamble comprising the preamble Stand der Techniek outlines and includes a text that describes the concept in one or more characteristics; - determining the IPC classification; - the targeted search in a database for one or more patent documents with the aforementioned IPC classification; - reporting data from said one or more patent documents to the user, whereby "word embedding" is integrated in the search process in addition to the known search procedures.
    According to a preferred form, the reporting of the search results is based on a structure.
    According to a preferred preferred form, said structure is a table, wherein said characteristics from the basic text are arranged in a column, preferably in order of relevance. The patent documents mentioned are arranged in a row of the table. The ranking of the patent documents is preferably done according to their respective relevance with regard to the mentioned characteristics. The raster elements corresponding to a specific characteristic and a specific patent document contain information about the location (page, line) where the specific characteristic occurs in the specific patent document.
    According to a more preferred preferred form, it is possible to link back from the structure / table to the database. For example, a (hyper) link can be provided in each raster element, which refers directly to the page and line of the patent document. The specific characteristic can be marked in the linked patent document.
    According to an alternative preferred preferred form, it is possible to adjust the structure interactively. For example, it is possible to move the characteristics from the basic text up or down in the table, whereby the arrangement of the patent documents is automatically adjusted. In this way it is possible to better define the scope of protection of the concept with regard to the current state of technology.
    According to a more preferred preferred form, it is possible to load one or more documents into the program to train the search algorithm. As a non-limiting example, said documents may contain lists of synonyms that link said base text to the vocabulary used in the State of the Art.
    According to a further preferred preferred form, the structure from a first search iteration can be used for a new search. For example, it is possible to make a selection of characteristics from the first search iteration and to perform the search process again based on this selection. For example, one can search for harmful documents from the State of the Art, specifically for this selection of characteristics. This is useful when investigating the novelty and inventiveness of a concept.
    According to a further preferred preferred form, the user can select features / claims in the structure that displays the search results, to have a patent application written on the basis of the intended type of patent (ie BE, EP, PCT, US, etc.). ..). Depending on the intended patent, the written language can also be chosen and templates can be used to indicate the choice of words for certain terminology. Furthermore, online dictionaries and encyclopedias can be used to further explain the specific terminology from the domain of the concept.
    In a second aspect, the invention relates to a system for the targeted search in a database, the reporting of relevant patent documents and the investigation of patentability of a concept, which system comprises a central processing unit, which is configured for carrying out the computer-implemented method as defined by claims 1-10.
    In a third aspect, the invention relates to a computer program product for targeted searching in a database, to report relevant patent documents and to investigate the patentability of a concept, which computer program product comprises at least a computer-readable medium on which computer-readable program code- portions are stored, which program code portions include instructions for performing the computer-implemented method as defined by claims 1-10.
    In the following, the invention is described a.d.h.v. non-limiting examples illustrating the invention, and which are not intended or may be interpreted to limit the scope of the invention.
    EXAMPLES EXAMPLE 1:
    Figure 1 shows a table illustrating a preferred form of a result of the method according to the present invention.
    This table comprises a column that lists the various characteristics Kl-Km (1) from the basic text and a row that arranges different patent documents D1-Dn (2) according to relevance. Dl-Dn contain a hyperlink to the patent document from the database. In each table cell (3) is displayed if a specific feature occurs in patent document D *, where a hyperlink is provided to the page and line (p, I) where the specific feature occurs in the patent document. The rows and columns can be arranged interactively, with the rest of the table automatically adjusted.
    It is believed that the present invention is not limited to the embodiments described above and that some modifications or changes can be added to the described examples without re-evaluating the appended claims.
    权利要求:
    Claims (12)
    [1]
    CONCLUSIONS
    A computer-implemented method for targeted searching in a database, reporting relevant patent documents and examining the patentability of a user's concept, which computer-implemented method comprises the following successive steps: Reading in a basic text in one or more national languages that describe the concept in one or more characteristics; Determining an IPC classification; Searching the database on the basis of one or more search procedures for one or more electronic patent documents with the aforementioned IPC classification; Reporting one or more search results, comprising the data of said one or more electronic patent documents to the user, characterized in that "word embedding" is integrated on the basis of a neural network in said search procedures.
    [2]
    Method according to the preceding claim 1, characterized in that the data is reported on the basis of a structure.
    [3]
    Method according to the preceding claim 2, characterized in that said structure comprises a table.
    [4]
    Method according to the preceding claim 3, characterized in that said features are in a column of said table and that said electronic documents are arranged in a row of said table, such that said features and the electronic documents are a grid to shape.
    [5]
    Method according to claims 2 to 4, characterized in that the structure can be adjusted interactively.
    [6]
    Method according to one of the preceding claims 1 to 5, characterized in that feedback from the said data of said one or more patent documents can be made to the database.
    [7]
    Method according to one of the preceding claims 1 to 6, characterized in that one or more documents can be loaded to train the neural network.
    [8]
    Method according to one of the preceding claims 1 to 7, characterized in that a number of patent documents are displayed, the number of which is at most a predetermined natural number.
    [9]
    Method according to one of the preceding claims 1 to 8, characterized in that the search results are used as the basis for a new search in the database.
    [10]
    Method according to one of the preceding claims 1 to 9, characterized in that a patent application is generated in one or more national languages on the basis of the one or more features selected by the user, this on the basis of templates and / or or online and / or offline dictionaries and / or encyclopedias.
    [11]
    A system for the targeted searching of a database, the reporting of patent documents and the investigation of patentability of a concept, which system comprises a central processing unit, which is configured for carrying out the computer-implemented method as defined by claims 1-10 .
    [12]
    A computer program product for the targeted search in a database, the reporting of patent documents and the investigation of the patentability of a concept, which computer program product comprises at least one computer-readable medium on which computer-readable program code portions are stored, which program code portions include instructions for performing the computer-implemented method as defined by claims 1-10.
    类似技术:
    公开号 | 公开日 | 专利标题
    US10789652B2|2020-09-29|Dimensionality reduction of multi-attribute consumer profiles
    US20180336640A1|2018-11-22|Rate analyzer models and user interfaces
    CN104199833B|2017-09-01|The clustering method and clustering apparatus of a kind of network search words
    Falessi et al.2010|A comprehensive characterization of NLP techniques for identifying equivalent requirements
    US9514412B2|2016-12-06|Techniques for detecting deceptive answers to user questions based on user preference relationships
    CN108733675B|2021-04-09|Emotion evaluation method and device based on large amount of sample data
    Gottapu et al.2016|Entity resolution using convolutional neural network
    Zeller et al.2017|Is the item-position effect in achievement measures induced by increasing item difficulty?
    Eder et al.2015|Configuring latent semantic indexing for requirements tracing
    Verma et al.2017|Evaluation of unsupervised learning based extractive text summarization technique for large scale review and feedback data
    Zosimov et al.2015|Inductive building of search results ranking models to enhance the relevance of text information retrieval
    Cavalcanti et al.2012|Detection and evaluation of cheating on college exams using supervised classification
    BE1022868B1|2016-09-29|COMPUTER IMPLEMENTED METHOD, SYSTEM AND COMPUTER PROGRAM PRODUCT FOR DIRECTLY SEARCHING IN A DATABASE AND EXAMINING THE PATENTABILITY OF A CONCEPT
    ISLJAMOVIC et al.2014|PREDICTING STUDENTS’ACADEMIC PERFORMANCE USING ARTIFICIAL NEURAL NETWORK: A CASE STUDY FROM FACULTY OF ORGANIZATIONAL SCIENCES
    CN109684627A|2019-04-26|A kind of file classification method and device
    Hassan2019|Enhancing supply chain risk management by applying machine learning to identify risks
    Ratna et al.2017|Cross-language plagiarism detection system using latent semantic analysis and learning vector quantization
    Rajmohan et al.2021|Cov2ex: a covid-19 website with region-wise sentiment classification using the top trending social media keywords
    JP6894315B2|2021-06-30|Document processing system, document processing method, and program
    CN110688405A|2020-01-14|Expert recommendation method, device, terminal and medium based on artificial intelligence
    Giménez et al.2017|Personality recognition using convolutional neural networks
    Pawar et al.2015|Deciphering review comments: Identifying suggestions, appreciations and complaints
    Kuta et al.2016|Authorship attribution of polish newspaper articles
    Flanagan et al.2015|Predicting and visualizing wine characteristics through analysis of tasting notes from viewpoints
    CN109299007A|2019-02-01|A kind of defect repair person's auto recommending method
    同族专利:
    公开号 | 公开日
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    法律状态:
    优先权:
    申请号 | 申请日 | 专利标题
    [返回顶部]