Machine learning-based user-customized automatic patent document classification method

Abstract

To provide a method that ensures for reducing time / effort / cost / inconvenience of users to spend for classifying patent documents.SOLUTION: Provided is a method for classifying patent documents using a computer, the method causes the computer to execute the steps of: classifying patent documents based on classification input of a user; identifying a patent pool including a plurality of patent documents; learning, based on classification standard, at least one patent document selected by the user from among the plurality of patent documents included in the patent pool; determining whether the classification standard can classify the remaining unclassified patent documents among the plurality of patent documents based on learning results based on the classification standard; displaying, to the user, an indicator indicating whether automatic classification can be executed based on the classification standard; and executing the automatic classification that classifies the remaining patent documents based on the classification standard in accordance with a request of the user.SELECTED DRAWING: Figure 3

Description

[Technical field]

[0001] This specification proposes a method and a system / apparatus for automatically classifying patent documents by learning the classification patterns of a user's patent documents and constructing a classification model, and using the constructed classification model. [Background technology]

[0002] Recently, machine learning, which can be applied to various fields, has been developed. Machine learning is a branch of artificial intelligence and can be divided into supervised learning, unsupervised learning, reinforcement learning, etc. according to its type, and various algorithms such as artificial neural networks, boosting, decision trees, support vector machines, random forests, etc. have been developed.

[0003] Artificial neural networks include an input layer, a hidden layer, and an output layer, and are divided into shallow neural networks and deep neural networks depending on the method of stacking hidden layers. In addition, artificial neural networks that are improved from deep neural networks, such as convolutional neural networks (ConvNet), recurrent neural networks, deep belief networks, and deep Q-networks, have been developed, and the machine learning methods of these artificial neural networks (including deep neural networks) are usually called deep learning. In other words, deep learning is a machine learning method that learns data by utilizing an information input / output layer like the neurons of the brain, and attempts a high level of abstraction by combining various nonlinear transformation techniques. Summary of the Invention [Problem to be solved by the invention]

[0004] Recently, as various patent-based businesses / projects have become more active, there has been a lot of work to classify patents by technology. Patent offices around the world classify technologies by assigning technology classification codes (IPC, CPC, etc.) to each patent document, but these technology classification codes are assigned arbitrarily and often do not match the technologies disclosed in the actual patent documents. Not only do they not match the actual industries and products to a great extent, but they also do not meet the criteria that users want to classify, so they are not often used in actual classification work. As a result, there is a problem in that users have to personally identify all the patents to be classified and perform the classification work, which consumes a lot of time, money, and manpower.

[0005] In addition, patent owners (individuals / companies) may set different patent life periods for different technologies depending on their own patent strategies. In this case, the patent owner must calculate the average duration of patent life / expiration for each technology of the expired patent documents, and then classify the pending patent documents by technology and apply the previously calculated duration for each technology. This not only requires a lot of time, money, and manpower, but also poses the problem that it is very difficult to manage the duration of patents when there are many patents owned. [Means for solving the problem]

[0006] According to one embodiment of the present invention, the method may include the steps of: learning patent documents stored in a patent database using machine learning techniques, establishing a first classification criterion based on similarities between the patent documents, and generating a base classification model for classifying the patent documents based on the first classification criterion; generating a patent pool including a plurality of patent documents based on a user's search and / or selection input; receiving a classification input from the user and classifying the plurality of patent documents included in the patent pool according to the classification input; predicting a second classification criterion for the user to classify the plurality of patent documents based on similarities between the classified patent documents; and classifying the remaining patent documents not yet classified in the patent pool based on the base classification model and the second classification criterion. Effect of the Invention

[0007] According to one embodiment of the present invention, the classification criteria of a user are learned, and the classification criteria desired by the user are predicted to automatically classify patent documents, thereby achieving the effect of reliably reducing the time, effort, cost, and inconvenience required for a user to classify patent documents.

[0008] In addition, according to one embodiment of the present invention, the system learns the expiration patterns of patent rights held by the user by technology, predicts the duration of the patent rights by technology desired by the user, and predicts the expiration period of currently pending patent rights, thereby enabling the user to manage the duration of the patents currently held more easily and efficiently. [Brief description of the drawings]

[0009] There are further advantages associated with various embodiments of the present invention, which will be described in more detail below with reference to the drawings.

[0010] [Figure 1] FIG. 1 is a diagram showing an example of a search result screen for patent documents of a patent search engine according to one embodiment of the present invention. [Diagram 2] FIG. 2 is a diagram illustrating a slide view screen according to an embodiment of the present invention. [Diagram 3] FIG. 3 is a flow chart illustrating a method for classifying patent documents in the patent document classification system according to an embodiment of the present invention. [Figure 4] FIG. 4 is a diagram illustrating a method for classifying patent documents in response to a user's classification input, according to an embodiment of the present invention. [Diagram 5] FIG. 5 is a diagram illustrating a folder setting UI according to an embodiment of the present invention. [Figure 6] FIG. 6 is a diagram illustrating a method for displaying representative information of patent documents according to a folder setting according to an embodiment of the present invention. [Figure 7] FIG. 7 is a diagram illustrating a method for classifying patent documents in response to a user's classification input according to an embodiment of the present invention. [Figure 8] FIG. 8 is an illustration of a classification history window according to one embodiment of the present invention. [Figure 9] FIG. 9 is a diagram illustrating a multi-folder storage method according to an embodiment of the present invention. [Figure 10]FIG. 10 is a flow chart illustrating an embodiment of predicting the expiration period of a patent document, according to an embodiment of the present invention. [Figure 11] FIG. 11 is a diagram illustrating a patent expiration recommendation list according to an embodiment of the present invention. [Figure 12] FIG. 12 is a block diagram of a patent document classification system according to an embodiment of the present invention. [Figure 13] FIG. 13 is a block diagram of a patent document classification system according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] The technology described below can be modified in various ways and can have various embodiments, so a specific embodiment will be illustrated in the drawings and described in detail. However, this does not limit the technology described below to a specific embodiment, and it should be understood that the technology described below includes all modifications, equivalents, and alternatives included in the spirit and scope of the technology described below.

[0012] Terms such as first, second, A, B, etc. may be used to describe various components, but the components are not limited by the terms and are used only for the purpose of distinguishing one component from another. For example, a first component may be named a second component, and similarly, the second component may be named a first component, without departing from the scope of the technology described below. The term "and / or" includes a combination of multiple related listed items or any of multiple related listed items.

[0013] In the terms used in this specification, singular expressions should be understood to include plural expressions unless the context clearly indicates otherwise, and terms such as "comprise" should be understood to mean the presence of a stated feature, number, step, operation, component, part, or combination thereof, but not to exclude the presence or additional possibility of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

[0014] Before describing the drawings in detail, it should be made clear that the components in this specification are merely classified according to the main function of each component. That is, two or more components described below may be combined into one component, or one component may be divided into two or more components according to more specific functions. Furthermore, each of the components described below may additionally perform some or all of the functions of the other components in addition to its own main function, and some of the main functions of each component may be exclusively performed by the other components.

[0015] Additionally, in performing a method or method of operation, the steps of the method may occur in a different order than that stated, unless the context clearly dictates a particular order, i.e., the steps may occur in the same order as stated, may occur substantially simultaneously, or may occur in the reverse order. The present invention is based on websites / servers that provide patent search engines (or computer programs / applications that embody patent search functionality via these websites / servers). Accordingly, the embodiments described herein may be performed by web servers, application servers, and / or client devices, and a system including these may be referred to as a "patent document classification system."

[0016] For convenience of explanation, the subject performing the embodiment will be described as a "patent search engine" or a "patent classification system", but may be substituted with a web server, an application server, and / or a client device. For convenience of explanation, the subject will be described based on a patent search engine provided via a website, but the present invention is not limited thereto and may of course be applied to a patent search engine provided via an application.

[0017] First, we will take a detailed look at the basic functions and GUI (Graphic User Interface) of the patent search engine that is the basis of the present invention, and then we will take a detailed look at the automatic classification method for patent documents.

[0018] FIG. 1 is a diagram showing an example of a search result screen for patent documents of a patent search engine according to one embodiment of the present invention.

[0019] The patent search engine of the present invention can provide a variety of information about patents retrieved in response to a user's search input on a single screen.

[0020] For example, as shown in Fig. 1, the patent search engine may provide, as a result of searching patent documents, the number of patent documents searched by country / status 130 and a list of patent documents searched overall or by country 140. In particular, when providing the list of patent documents 140, the patent search engine may also provide bibliographic information (country of application, application / registration number, type, application date, current status (application / publication / registration, etc.), name of invention, etc.) related to each patent document.

[0021] In addition, the patent search engine may provide an execution button tab 120 for the user to instruct various additional operations on the searched patent documents. For example, as shown in FIG 1, the patent search engine may provide an execution button tab 120 including execution buttons for slide view 110, work save, graph engine, download, and list output operations.

[0022] Here, the slide view 110 provides a function of summarizing and listing representative information of each patent document (e.g., representative drawings, abstract information, independent claims, etc.) on one screen so that the user can more easily understand the patent documents found, which will be described in more detail below with reference to FIG.

[0023] The "Work Save" function provides a function to save patent documents in a specific folder, and can be used for the classification of patent documents proposed in this specification. The "Graph Engine" function collects patent documents selected by the user and generates graphs or calculates statistics according to the criteria desired by the user. The "Download" function provides a function to download patent documents, and the "List Output" function provides a function to output a list of searched patent documents, abstract information, etc.

[0024] In addition, the patent search engine may provide execution buttons for instructing various additional operations using patent documents, and is not limited to the above-mentioned embodiments.

[0025] FIG. 2 is a diagram illustrating a slide view screen according to an embodiment of the present invention.

[0026] Referring to FIG. 2, the slide view screen can be broadly composed of a highlight window 210, a project window 220, a representative drawing window 230, a representative content window 250 of a patent document, and a patent list window 240.

[0027] The highlight window 210 provides a function to highlight specific keywords / search terms entered by the user in each patent document. The project window 220 is a project window where the user can manage (e.g., create and edit) various upper / lower folders created by the user. The user can categorize patent documents by saving them in desired folders and can manage patent documents by folder.

[0028] The representative drawing window 230 provides a function of displaying the representative drawing of the target patent document, and the patent document representative content window 250 provides a function of summarizing representative information necessary for understanding the target patent document. Through the representative drawing window 230 and the patent document representative content window 250, the user can easily understand the characteristics and technical content of the target patent document at once.

[0029] The patent list window 240 is a window that provides a list of patent documents that the user has selected to view in slide view. The patent list window 240 allows the user to know how many and what types of patent documents are included in the current pool, and also to know what type of document the patent document currently being viewed by the user is.

[0030] The patent search engine on which the present invention is based has been briefly described above. The embodiments of the present invention described below can be implemented / provided on the basis of such a patent search engine, and in particular, the patent classification system and method can be provided as a function within a slide view screen. However, the present invention is not limited to being applied to the above-mentioned patent search engine, and can of course be applied in the same / similar manner to various patent search engines / sites.

[0031] FIG. 3 is a flow chart illustrating a method for classifying patent documents in the patent document classification system according to an embodiment of the present invention.

[0032] Referring to Fig. 3, first, the patent document classification system can learn some or all of the patent documents stored in the patent database using machine learning technology. Furthermore, the patent document classification system can establish a first classification criterion based on the similarity between the patent documents learned based on the machine learning technology, and generate a basic classification model for classifying the patent documents according to the established first classification criterion (S301).

[0033] The patent document classification system can learn patent documents using various machine learning techniques. As an embodiment of the machine learning technique, RNNs (Recurrent Neural Networks), DNNs (Deep Neural Networks) and / or CNNs (Convolutional Neural Networks) can be used. The RNNs (Recurrent Neural Networks) technology is an artificial neural network that recognizes patterns from data having a certain sequence (sequence or time series data) such as sentences, genes, handwritten characters, voice signals, data detected by sensors, text, and stock prices. In the RNNs technology, currently input data and previously input data are considered simultaneously, and the sequence information is stored in the hidden layer of the RNNs. The information stored in the hidden layer can be used again when an input that requires the stored information is received after a certain time. DNNs refers to a learning method having two or more such hidden layers. In the DNNs, a computer can derive an optimal division line by repeating the process of creating classification labels by itself, distorting the space, and dividing the data. CNNs technology is an algorithm that applies such DNNs, and will be described in more detail later in step S306.

[0034] The patent document classification system may use a translation engine to translate patent documents into one common language (e.g., Korean or English) before learning the patent documents, and may learn the patent documents in one unified language. However, the invention is not limited to this, and the patent document classification system may use a translation engine to translate patent documents into up to two languages ​​before learning the patent documents.

[0035] When determining the similarity between patent documents, the patent document classification system can make the determination based on the family of each patent document, patent classification code (e.g., IPC, CPC, etc.), prior patent document, patent document cited at the examination stage, patent document cited at the examination stage of other patent documents, keywords (extracted by text mining) and / or text data, etc. In this case, the text data can be vectorized using algorithms such as Word Emdedding and Word2vec.

[0036] The patent document classification system can use cosine coefficient, Jaccard coefficient, Dice coefficient, Euclidean distance, vector inner product, etc. to calculate the similarity between patent documents. Depending on the embodiment, a specific similarity measurement method may need to be used depending on the classification (or clustering) technique, such as the Ward technique that measures the distance between documents using Euclidean distance and the group average technique that calculates similarity by the product of inner products, and the patent documents (or members) belonging to each classification (or cluster) may vary depending on the similarity measurement method.

[0037] Next, the patent classification system can generate a patent pool (Pool) including a plurality of patent documents based on the user's search and / or selection input (S302). The plurality of patent documents included in the patent pool can be the user's classification target documents, as described below.

[0038] Next, the patent document classification system receives a classification input from the user (S303) and can classify a plurality of patent documents included in the patent pool according to the classification input (S304). The patent document classification system can generate at least one folder according to a folder creation input from the user, and can perform classification work by saving (or moving) patent documents in the patent pool to each folder according to the classification input from the user. The classification input can be various user inputs for classifying patent documents, and can be embodied in various ways. In one embodiment, the user can easily perform classification work by performing a motion / touch input of dragging and dropping a patent document into a folder. In another embodiment, the classification input can correspond to a user input for a save now button provided by the patent document classification system. Here, the save now button can correspond to a UI provided by the patent document classification system so that the user can immediately save and classify a patent document in a folder desired by the user. The save now button can be provided for each patent document. In another embodiment, the classification input can correspond to a user input for classifying a target patent document in a toggling manner, for example, an O or an X. A user can classify a target patent document into two categories (e.g., O or X) by selecting / inputting O or X. The category item (O or X) can be set directly by the user or provided by an item already set in the web / application server. An embodiment of the category input will be described in more detail below with reference to Figures 4 and 7.

[0039] Next, the patent document classification system can predict a second classification criterion for the user to classify the patent documents based on the similarity between the classified patent documents (S305). The method of predicting the second classification criterion based on the similarity can be the same / similar to the method of establishing the first classification criterion based on the similarity.

[0040] Next, the patent document classification system may classify the remaining patent documents in the patent pool that have not yet been classified based on the basic classification model and the secondary classification criteria (S306).

[0041] More specifically, in one embodiment, the patent document classification system may build a user-customized classification model by combining the predicted second classification criterion with the basic classification model. The user-customized classification model may be built based on transfer learning. Transfer learning may mainly use a pre-trained model, and the pre-trained model refers to a model that has a similar problem-solving purpose but has already been trained on large-sized data. In particular, the present specification describes a pre-trained model used in transfer learning that has a Convolutional Neural Networks (CNNs) structure. This CNN is composed of the following two parts.

[0042] -Convolutional base: A layer of convolutional layers and pooling layers. The goal of the convolutional base is to efficiently extract features from data.

[0043] -Classifier: It mainly consists of a fully connected layer. A fully connected layer is a layer in which all neurons in the layer are connected to the output nodes of the previous layer without exception. The ultimate goal of the classifier is to learn the extracted features well and classify the data into appropriate categories (image classification).

[0044] The patent document classification system can use these two parts to build a user-customized classification model, and in particular, can build a user-customized classification model using a strategy in which the convolutional base is fixed and only the classifier is newly trained. Such a strategy is advantageously applicable when computer computing power is insufficient or the data set is very small, and / or when the problem to be solved is very similar to the data set on which the pre-trained model has already been trained.

[0045] When this is applied to the present embodiment, the patent document classification system can build a user-customized classification model by fixing the basic classification model as a convolutional base and then newly learning the second classification criterion as a classifier. Here, fixing the basic classification model can be interpreted as fixing the features / criteria for provisionally classifying the patent documents in the patent pool based on the first classification criterion of the basic classification model by the number of folders created by the user. Therefore, the features / criteria of provisional classification based on the first classification criterion are fixed, and the classification features / criteria based on the second classification criterion are newly learned as a classifier, thereby building a user-customized classification model.

[0046] In another embodiment, the patent document classification system may classify patent documents using only the basic model without learning the patent documents classified by the user as in the above embodiment. In this case, the patent document classification system may calculate the technical distance between the patent documents classified by the user based on the basic classification model, and may predict the second classification criterion based on the calculated technical distance (i.e., no learning is required). The patent document classification system may classify the remaining patent documents based on the second classification criterion predicted based on the basic classification model. For example, if a user classifies some patent documents in the patent pool into four classes, the patent document classification system may extract the characteristics of the patent groups belonging to the four classes based on the basic classification model, and may automatically classify the remaining patent documents according to the characteristics of the patent groups in the basic classification model thus extracted. In this embodiment, the second classification criterion may be interpreted as a form in which the classifications of the first classification criterion are grouped according to the user's classification characteristics, thereby reducing the overall classification types / number. For example, if there are 1000 technology categories that can be classified based on the first classification criterion, the second classification criterion can be established to divide the 1000 into 50 technology categories by clustering / grouping those with high similarity based on the user's classification pattern / criteria.

[0047] Furthermore, the patent document classification system may output an input button (e.g., a classification start button) for instructing automatic classification of the remaining patent documents. When receiving a user's input to the input button, the patent document classification system may classify the remaining patent documents using a user-customized classification model. More specifically, when receiving a user's input to the input button, the patent document classification system may classify the remaining patent documents using a user-customized classification model and automatically save them in respective folders that match the classification characteristics, thereby classifying the remaining patent documents.

[0048] That is, this flowchart can be applied as in the following embodiment.

[0049] A user can search for 1,000 patents and set them in a patent pool to generate four folders. The patent classification system can provisionally classify the 1,000 patents into four folders using a basic classification model. A user can first classify and save 100 patents out of the 1,000 into four folders, and the patent classification system can predict the user's classification criteria / characteristics from such user classification information. The patent classification system will generate a model that adds the user's classification criteria / characteristics to the provisional classification criteria / characteristics (i.e., the first classification criteria), and classify and save the remaining 900 patents into four folders.

[0050] On the other hand, although not shown in this flowchart, the patent document classification system can calculate the accuracy of the automatic classification (first accuracy) in real time and output an indicator that displays the first accuracy in real time.

[0051] In one embodiment, the first accuracy may be calculated based on the number of patent documents included in the patent pool, the number of classifications set / specified by the user, and / or the number of patent documents classified by the user to date. More specifically, the patent document classification system may learn a basic classification model and calculate / derive in advance the second accuracy based on the total number of patent documents to be classified (first factor), the number of classifications (second factor), and / or the number of patent documents included in each classification (third factor) as data. For example, the patent document classification system may calculate / derive in advance statistical data that, as a result of learning the basic classification model, when there are 1,000 patent documents to be classified, the number of classifications is 3, and the number of patent documents included in each classification is 100, the classification accuracy is 80%. In this case, the patent document classification system may derive the first accuracy by substituting the user's classification pattern / criteria as a factor into such statistical data. That is, the first accuracy may correspond to the second accuracy calculated in advance when the number of patent documents included in the patent pool by the user, the number of classifications according to the second classification standard, and / or the number of patent documents included in each classification are sequentially applied to the statistical data as the first to third factors. Therefore, if the user includes 1,000 patent documents in the patent pool and classifies 100 patent documents into each of three folders, according to the above example, the user's classification accuracy (i.e., the first accuracy) is derived as 80%. In addition, the patent classification system may derive / display the first accuracy by additionally considering the probability of how close the patent documents included in each classification by the user are (indirect verification of whether the user properly entered them into the classification).

[0052] Based on the above, the first to third factors for deriving the maximum accuracy may be fixed within the statistical data derived based on the basic classification model. In this case, if the patent document classification system determines that the derived first accuracy is lower than the maximum accuracy, it may output a pop-up window to guide the user to additional classification. For example, when classifying patent documents into three classifications in a patent pool including 1,000 patent documents to be classified, the minimum number of patent documents that must be classified into each folder to derive a maximum accuracy of 80% may be set to 100. Therefore, the patent document classification system may output a pop-up window to guide the user to include at least 100 patent documents in each folder to reach the maximum accuracy of 80%. For example, the patent document classification system may output a pop-up window such as "The current classification accuracy is 60%, so please classify more patents" or "The classification accuracy of the XX folder is low, so please classify more patents into the XX folder."

[0053] In another embodiment, the first accuracy may correspond to a degree of matching between a result classified according to a first classification criterion and a result classified according to a second classification criterion.

[0054] The patent document classification system may calculate the accuracy of automatic classification by folder and / or by patent pool, and may output an indicator showing the accuracy by folder and / or by patent pool.

[0055] FIG. 4 is a diagram illustrating a method for classifying patent documents in response to a user's classification input, according to an embodiment of the present invention.

[0056] As mentioned above, the user can easily classify patent documents by dragging and dropping them into the folder in which they are to be classified.

[0057] Based on the slide view screen configuration, for example, as shown in Fig. 4, the patent document classification system can output an icon 450 in a representative content window of the patent document, which allows a user to drag and drop the patent document into a folder. Folders 420 created by the user for classification can be provided in a project window 410 of the slide view, and the user can perform classification and storage in the corresponding folder 410 by dragging and dropping 430 the icon 450 into a desired folder 410.

[0058] In addition, the patent document classification system may provide a setting UI 440 to allow a user to directly set ratings and / or technical themes / keywords for each folder 410, as described below with reference to FIGS. 5 and 6.

[0059] FIG. 5 is a diagram illustrating a folder setting UI according to one embodiment of the present invention, and FIG. 6 is a diagram illustrating a method of displaying representative information of patent documents according to folder settings according to one embodiment of the present invention.

[0060] The patent document classification system may provide a setting UI 520 in which a user can directly input additional information for each folder. For example, as shown in Fig. 5, the patent document classification system may output a setting UI 520 in which a folder 510 rating setting and / or keywords / themes can be set. After selecting a folder 510 to be set, a user can directly set an importance rating and / or keywords / themes for the selected folder 510, thereby more easily managing patent documents by folder 510.

[0061] For patent documents that have additional information entered / set using the setting UI 520, the set / entered additional information can be displayed as representative information and provided to the user.

[0062] 5 and 6, if the grade of folder B is set to A and the keyword / theme is set to #camera, the grades of all patent documents stored in folder B are set to A and the keywords / themes are set to #camera, and are displayed as representative information 610. Therefore, the user does not need to set the grade / keywords / theme, etc. for each patent document individually, but can set them on a folder-by-folder basis using the folder setting UI 520, so that the user can more easily manage patent documents by folder.

[0063] FIG. 7 is a diagram illustrating a method for classifying patent documents in response to a user's classification input according to an embodiment of the present invention.

[0064] As described above, the user can easily perform classification by providing user input to the Save Now button 710 that is output together with the patent document.

[0065] Based on the slide view screen configuration, for example, as shown in FIG. 7, the patent document classification system may output a Save Now button 710 in the patent document representative content window, which can execute a function of immediately saving the corresponding patent document in a folder. The user can immediately save the corresponding patent document in a folder by touching / selecting 720 the Save Now button. If there are multiple folders in which the patent document can be saved, the patent document classification system may provide a selection UI 730 that allows the user to select a folder in which the patent document can be saved when receiving a user input on the Save Now button. The user can immediately save the target patent document in the corresponding folder by selecting and inputting a folder in which the target patent document is to be saved from among the folders displayed in the selection UI 730.

[0066] FIG. 8 is an illustration of a classification history window according to one embodiment of the present invention.

[0067] Referring to FIG. 8, the patent document classification system can record a user's recent classification history and provide the user with a window (ie, a classification history window) 810 in which the relevant record information is displayed.

[0068] In the classification history window 810, patent documents recently classified by the user may be listed in the order in which they were classified, and simple classification information for each patent document (e.g., the application / registration number of the classified patent document, the folder in which it was saved, etc.) may also be provided. In addition, the classification history window 810 may provide an "Undo" icon 820 for each classified patent document. The "Undo" icon 820 corresponds to an icon for canceling the classification operation of the target patent document. Therefore, when a user input is received for the cancel input, the patent document classification system can cancel the classification operation for the target patent document for which the cancel input was received, and the target patent document will return to the state before it was saved / belonged to the corresponding folder (i.e., before classification).

[0069] FIG. 9 is a diagram illustrating a multi-folder storage method according to an embodiment of the present invention.

[0070] The patent classification system supports a multi-folder storage mode, so that one patent document can be stored in multiple folders at the same time, for example, a target patent document can be stored in folders A and B at the same time.

[0071] The patent document classification system may provide an icon 910 for displaying information about the folder in which the target patent document is stored / classified. For example, as shown in FIG. 9, if the target patent document is stored / classified in folders A and B, the patent document classification system may provide a folder icon 910 indicating the number of folders, 2, and when the user touches / selects the folder icon 910, may provide list information (A and B folders) 920 of the folders in which the target patent document is currently stored.

[0072] Since the patent document classification system supports a multi-folder storage method, each patent document can be assigned a rank and / or keywords / themes up to the total number of folders currently created (as ranks and / or keywords / themes can be set by folder).

[0073] Although not shown in this drawing, a patent document classification system that has completed automatic classification can provide a reclassification UI for each folder to allow users to directly check and correct the classification contents.

[0074] The reclassification UI may provide a reclassification target display button for the user to reclassify the patent documents classified in each folder again and a reclassification button for instructing the reclassification of the patent documents to be reclassified. Thus, the user may use the 'reclassification target display button' to display patent documents that are not suitable for classification in the corresponding folder, and may use the 'reclassification button' to delete the patent documents displayed by the reclassification target display button from the corresponding folder or to move them to another folder. The reclassification targets may be displayed in various embodiments, for example, using tags or colors provided by the web / application server (or set directly by the user). In this case, the user may select and input (e.g., touch / click input) tags or colors provided by the web / application server for each patent document to specify the patent documents that need to be reclassified.

[0075] The patent document classification system can evolve by digitizing the content of reclassifications made by users in this way and learning from the reclassification data based on the basic classification model.

[0076] We have discussed in detail above how to build a user-customized classification model based on machine learning (or deep learning) and use it to automatically classify patent documents. In this way, the method of building a new model by combining a user's classification patterns / habits as new features / criteria with a conventional model can be applied to various user-customized patent management embodiments. For example, the method can be used to predict / recommend the expiration date of a patent right, the number of claims to be maintained for each patent, the possibility of right transfer (license), patent maintenance methods for each inventor, and patent maintenance methods based on patent evaluation.

[0077] For example, when the method of the present invention is applied to predicting the expiration period of patent rights, the patent classification system can classify the patents held by the user by technology (at this time, a basic classification model can be used), learn the patent expiration patterns for each classified technical field, and generate a user-customized model. As a result, the patent classification system can automatically recommend the patent expiration prediction period for each technical field at a cycle desired by the user according to the patterns / characteristics of the user's existing patent management (expiration, etc.).

[0078] For example, if the Korea Institute of Life Sciences has a characteristic that stem cell-related patents do not expire and are maintained for more than 15 years, while compound and manufacturing process patents have a characteristic / pattern of expiring all at once within five years, the patent document classification system will identify such expiring characteristics / pattern and recommend that patents in the stem cell technology field currently maintained by the Korea Institute of Life Sciences be maintained for more than 15 years, and guide / recommend that compound and manufacturing process patents be allowed to expire within five years.

[0079] This embodiment will be described in more detail below with reference to FIG.

[0080] FIG. 10 is a flow chart illustrating an embodiment of predicting the expiration period of a patent document, according to an embodiment of the present invention.

[0081] For convenience of explanation, this embodiment has been created as an embodiment following the embodiment of FIG. 3, but it is not limited thereto, and it is of course possible to carry out an embodiment independent of the embodiment of FIG. 3.

[0082] First, the patent document classification system can calculate, by classification, the first average period from the application date or registration date to the expiration date of patent documents classified according to a first classification criterion using a basic classification model for patent documents included in a patent database (DB) (S1001).

[0083] Next, the patent classification system classifies the documents included in the patent pool according to the first or second classification criteria (S1002), and calculates the second average period from the application date or registration date to the expiration date of the documents for each classification (S1003). Here, the documents included in the patent pool may correspond to (include) patents that the user currently owns, or that the user owned but have now expired. Through this step, the patent classification system becomes able to predict the average patent expiration period / pattern / characteristics for each technology of the user.

[0084] Finally, the patent classification system can predict the expiration period of a particular patent document by taking into account the first and second average periods (S1004), where the particular patent document can include patents currently held by the user that have not yet expired and / or patents identified according to user input.

[0085] In order to predict the expiration period, the patent document classification system may first classify a specific patent document according to the first and / or second classification criteria. Then, the patent document classification system may calculate the expiration period of a specific patent document by combining the first and / or second average periods according to the classification of the specific patent document according to the first and / or second classification criteria at a preset ratio. For example, it may be assumed that a first specific patent document belongs to the "bio" classification according to the first classification criteria, and the average lifespan of the "bio" classification is 5 years in the overall patent database. It may also be assumed that a first specific patent document belongs to the "biorhythm" classification according to the second classification criteria, and the average lifespan of the "biorhythm" classification is 10 years in the user expiration patent. If the preset ratio between the first and second classification criteria is set to 3:7, a period of 5 x 0.3 + 10 x 0.7 = 8.5 may be derived as the evaluation index. That is, the expected lifespan of the first specific patent document may be predicted to be 8 years and 6 months.

[0086] The preset ratios can be freely set / changed according to various embodiments, and in terms of a user-customized model, the ratio of the second classification criterion can be set higher than the ratio of the first classification criterion.

[0087] FIG. 11 is a diagram illustrating a patent expiration recommendation list according to an embodiment of the present invention.

[0088] Referring to FIG. 11, the patent document classification system can list the predicted expiration period for each specific patent document as an evaluation index 1120 and output the list by including it in the patent expiration recommendation list 1110.

[0089] The time when the patent expiration recommendation list 1110 is output can be set to a time before the predicted expiration time (for example, 6 months or 1 year before) that has already been set, and can be freely set by the user.

[0090] Using this patent expiration recommendation list 1110, the user can predict the expiration period of the patent rights currently held, and can efficiently manage the duration of each patent right by determining in advance whether to maintain the right or allow it to expire.

[0091] FIG. 12 is a block diagram of a patent document classification system according to an embodiment of the present invention.

[0092] Referring to FIG. 12, the patent document classification system can be largely composed of a web server (or an application server) and a client device.

[0093] The web server (or application server) 1200 may include a processor 1210, a memory unit 1220, and / or a communication unit 1230, and the client device 1300 may include a processor 1240, a sensor unit 1250, a communication unit 1260, and / or a display unit 1270. At least one of the components may be omitted or a new component may be included depending on the embodiment. In addition, each unit may be provided with a hardware circuit / device, etc., to perform the functions described below, and may be embodied in one chip or multiple chips.

[0094] The memory unit 1220 can store various digital data such as video, audio, photos, movies, computer programs, applications, etc. The memory unit 1220 can correspond to various digital data storage spaces such as flash memory, a hard disk drive (HDD), a solid state drive (SSD), etc. In particular, the memory unit 1220 of the web server (or application server) 1200 can correspond to a patent database, and various patent documents from around the world can be stored therein.

[0095] The communication units 1230, 1260 can communicate using at least one communication protocol to transmit / receive various data. In particular, a user input to the client device 1300 can be transferred to the web server (or application server) 1200 via the communication units 1230, 1260, and various output / feedback data / information of the web server (or application server) 1200 can also be transferred to the client device 1300 via the communication unit 1260.

[0096] The sensor unit 1250 may be a generic name for the various sensing means described above, and may sense various user inputs and / or the environment of the client device 1300. In particular, the sensor unit 1250 may sense various user inputs (e.g., touch / motion / selection / search input, etc.) and transfer the sensing results to the processor. In one embodiment, the sensor unit 1250 may include any one of various sensing means such as a gravity sensor, a geomagnetic sensor, a motion sensor, a gyroscope sensor, an acceleration sensor, an infrared sensor, an inclination sensor, a brightness sensor, an altitude sensor, an olfactory sensor, a temperature sensor, a depth sensor, a pressure sensor, a bending sensor, an audio sensor, a video sensor, a Global Positioning System (GPS) sensor, a touch sensor, and a grip sensor. The above-mentioned sensors may be included in the device as separate elements or may be integrated into at least one or more elements.

[0097] The display unit 1270 may display various images / videos. In particular, the display unit 1270 of the client device 1300 may display various web pages / UIs provided by a web server, various application screens / UIs provided by an application server, etc. The processors 1210 and 1240 can control at least one other component unit to perform various embodiments proposed herein. Therefore, the execution subject of the embodiments disclosed herein can be described as being replaced by the processor. The processors 1210 and 1240 can execute various programs / applications stored in the memory unit 1220 and process data / information inside the server / device. The processors 1210 and 1240 can include at least one central processing unit (CPU), micro processor unit (MPU), micro controller unit (MCU), application processor (AP), or any other type of processor well known in the technical field of the present invention. The processor can perform operations for at least one application or program for executing a method according to an embodiment of the present invention.

[0098] FIG. 13 is a block diagram of a patent document classification system according to an embodiment of the present invention.

[0099] The components of this drawing may be embodied using at least one of the components shown in Fig. 12. For example, the load unit may be embodied by a processor and a memory unit, the folder generation unit, the patent document display unit, and the folder display unit may be embodied by a processor and a display unit, and the storage unit may be embodied by a memory unit.

[0100] The loading unit 1310 may load at least one patent document requested by a user. The loading unit 1310 is connected to a storage space for storing various data, commands and / or information, and may store one or more applications for providing a patent classification service providing method according to an embodiment of the present invention. The loading unit 1310 may also store various information received via a communication unit or input via a user input.

[0101] The storage unit 1350 may temporarily or non-temporarily store data transferred from an external device, etc. The storage unit 1350 may include a non-volatile memory such as a read only memory (ROM), an erasable programmable ROM (EPROM), an electrically erasable programmable ROM (EEPROM), a flash memory, etc., a hard disk, a removable disk, or any other type of computer-readable recording medium well known in the art to which the present invention pertains.

[0102] The folder creation unit 1320 can serve to create folders for storing or classifying patent documents.

[0103] The patent document display unit 1330 serves to display at least one of the classification items in the patent document in the information area, and the folder display unit 1340 serves to display the folder in the folder area. It is preferable that the information area and the folder area are separate areas.

[0104] In addition, the storage unit 1350 sets at least a portion of the information area as a draggable area and at least a portion of the folder area as a droppable area, and serves to store patent documents in folders in response to a user's drag and drop input.

[0105] Embodiments of the present invention may be implemented by various means, such as hardware, firmware, software, or a combination thereof. In a hardware implementation, an embodiment of the present invention may be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), processors, controllers, microcontrollers, microprocessors, etc.

[0106] In addition, in the case of implementation by firmware or software, an embodiment of the present invention may be implemented in the form of a module, procedure, function, etc. performing the above-described functions or operations, and may be recorded on a recording medium readable by various computer means. Here, the recording medium may include program instructions, data files, data structures, etc., alone or in combination. The program instructions recorded on the recording medium may be specially designed and configured for the present invention, or may be known and available to those skilled in the art of computer software. For example, the recording medium may include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as compact disk read only memories (CD-ROMs) and digital video disks (DVDs), magneto-optical media such as floptical disks, and hardware devices specially configured to store and execute program instructions, such as ROMs, RAMs, and flash memories. Examples of the program instructions may include not only machine language codes created by a compiler, but also high-level language codes that can be executed by a computer using an interpreter, etc. Such hardware devices may be configured to operate with one or more software programs to perform the operations of the present invention, and vice versa.

[0107] Additionally, the devices and terminals of the present invention may be driven by instructions that cause one or more processors to perform the functions and processes described above. For example, the instructions may include interpreted instructions, such as script instructions, e.g., JavaScript or ECMAScript instructions, executable code, or other instructions stored on a computer readable medium. Additionally, the devices of the present invention may be embodied in a distributed manner across a network, such as a server farm, or may be embodied in a single computing device.

[0108] Also, the computer programs (also known as programs, software, software applications, scripts or code) that are loaded into the apparatus of the present invention and that perform the methods of the present invention can be written in any form of programming language, including compiled or interpreted languages, a priori or procedural languages, and can be deployed in any form, including stand-alone programs, modules, components, subroutines, or other units suitable for use in a computing environment. A computer program does not necessarily correspond to a file in a file system. A program can be stored in a single file provided to a requested program, or in multiple interacting files (e.g., a file storing one or more modules, subprograms or portions of code), or in a portion of a file that holds other programs or data (e.g., one or more scripts stored in a markup language document). A computer program can be deployed to run on multiple computers or on a single computer that is located at one site or distributed across multiple sites and interconnected by a communications network.

[0109] For convenience of explanation, each drawing is described separately, but it is also possible to combine the embodiments described in each drawing to design a new embodiment. In addition, as described above, the present invention is not limited to the configurations and methods of the described embodiments, and the above-described embodiments may be configured by selectively combining all or part of each embodiment so that various modifications are possible.

[0110] Furthermore, although preferred embodiments have been illustrated and described above, this specification is not limited to the specific embodiments described above, and it goes without saying that various modifications may be made by a person having ordinary knowledge in the technical field to which this specification pertains without departing from the gist of the claims, and such modifications should not be understood individually from the technical ideas and perspectives of this specification. [Explanation of symbols]

[0111] 110 Slide Views 120 Run Button Tab 130 patent documents 140 Patent Document List 210 Highlight Window 220 Project Window 230 Representative Drawing Window 240 Patent List Window 250 Patent Document Representative Contents Window 450 Icons 410 Project Window 420 Folder 430 Drag and Drop Input 440 Settings UI 510 Folder 520 Settings UI 610 Representative information 710 Save button 720 Touch / Select 730 Selection UI 810 Classification History Window 820 Cancel execution icon 910 Folder Icon 920 List Information (A and B folders) 1110 Patent Expiration Recommendation List 1120 Rating Index 1200 Web Server 1210 Processor 1220 Memory Unit 1230 Communication Unit 1240 Processor 1250 Sensor Unit 1260 Communication Unit 1270 Display Unit 1300 Client Devices

Claims

1. 1. A computer-implemented method comprising: The method further comprises: classifying the patent documents based on a user classification input; identifying a patent pool comprising a plurality of patent documents; learning at least one patent document selected by a user from among the plurality of patent documents included in the patent pool based on a classification criterion; determining whether the classification criterion is capable of classifying the remaining unclassified patent documents among the plurality of patent documents based on a learning result based on the classification criterion; displaying an indicator to a user indicating whether automatic classification can be performed based on the classification criteria; and performing an automatic classification of the remaining patent documents based on a classification criterion based on a user request.

2. The learning step includes: generating a folder based on a user request; storing at least one patent document from the plurality of patent documents in the created folder; The method of claim 1 , further comprising the step of: studying at least one patent document stored in the folder.

3. The method of claim 2 , wherein the indicator is determined for a folder in which at least one patent document is stored based on the user's classification input.

4. The step of classifying the remaining patent documents based on a classification criterion comprises: The method of claim 1 , further comprising identifying and displaying patent documents from among the remaining patent documents that meet classification criteria.

5. The indicator is The method of claim 1 , wherein the number of the at least one user-selected patent document is determined to be different based on at least one of a degree of similarity between the patent documents.

6. The method of claim 1 , further comprising the step of providing guidance for classifying the user-selected patent document if the patent document is not suitable for automatic classification based on classification criteria.

Images