WO2021120979A1 - 生成专利概述信息的方法、装置、电子设备和介质 - Google Patents

生成专利概述信息的方法、装置、电子设备和介质 Download PDF

Info

Publication number
WO2021120979A1
WO2021120979A1 PCT/CN2020/130300 CN2020130300W WO2021120979A1 WO 2021120979 A1 WO2021120979 A1 WO 2021120979A1 CN 2020130300 W CN2020130300 W CN 2020130300W WO 2021120979 A1 WO2021120979 A1 WO 2021120979A1
Authority
WO
WIPO (PCT)
Prior art keywords
sentence
information
candidate
technical solution
patent document
Prior art date
Application number
PCT/CN2020/130300
Other languages
English (en)
French (fr)
Inventor
马库斯 汉斯
张�成
蔡洁
袁明
陆蕴芳
Original Assignee
智慧芽信息科技(苏州)有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 智慧芽信息科技(苏州)有限公司 filed Critical 智慧芽信息科技(苏州)有限公司
Priority to EP20901335.8A priority Critical patent/EP4080381A4/en
Publication of WO2021120979A1 publication Critical patent/WO2021120979A1/zh
Priority to US17/844,822 priority patent/US20220365956A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/34—Browsing; Visualisation therefor
    • G06F16/345—Summarisation for human users
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/35—Clustering; Classification
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/35—Clustering; Classification
    • G06F16/353—Clustering; Classification into predefined classes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/38—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
    • G06F16/383—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/30—Semantic analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/10—Services
    • G06Q50/18—Legal services
    • G06Q50/184—Intellectual property management
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/20—Natural language analysis
    • G06F40/279—Recognition of textual entities
    • G06F40/284—Lexical analysis, e.g. tokenisation or collocates
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/20—Natural language analysis
    • G06F40/279—Recognition of textual entities
    • G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking

Definitions

  • the present disclosure relates to the field of computer technology, for example, to methods, devices, electronic devices, and media for generating patent summary information.
  • the text summarization technology is a common technology for extracting effective information. It usually uses a computer to process natural language text, and automatically extracts part of the content that can accurately reflect the central content of the text from the natural language text. Such technology for extracting effective information helps reduce the information overload problem faced by Internet users, and helps users locate the information they need from the Internet faster and more effectively.
  • the present disclosure proposes methods, devices, electronic equipment, and media for generating patent summary information to obtain effective patent information from massive amounts of patent information.
  • a method for generating technical problem information including:
  • a target candidate sentence from the above at least one candidate sentence, wherein the target candidate sentence includes at least one of the following: a sentence that is successfully compared with the subject of the target patent document, and a sentence of a predefined category;
  • a method for generating solution information includes:
  • For each candidate technical solution sentence in the at least one candidate technical solution sentence determine the dominant feature group of the candidate technical solution sentence, and score or classify the candidate technical solution sentence according to the dominant feature group;
  • solution information is generated.
  • a method for generating content information of parts includes:
  • component content information is generated.
  • a device for generating patent summary information includes:
  • the technical problem extraction unit is configured to extract technical problem information from the target patent document
  • the solution extraction unit is configured to extract solution information from the aforementioned target patent document;
  • the patent summary information generation unit is configured to generate patent summary information based on the aforementioned technical problem information and the aforementioned solution information.
  • a device for generating technical problem information includes:
  • the candidate sentence extraction unit is configured to extract at least one candidate sentence related to the technical problem from the target patent document;
  • the candidate sentence determining unit is configured to determine a target candidate sentence from the at least one candidate sentence, wherein the target candidate sentence includes at least one of the following: a sentence that is successfully compared with the subject of the target patent document, and a sentence of a predefined category ;
  • the technical problem information generating unit is configured to generate technical problem information based on the determined target candidate sentence.
  • a device for generating solution information includes:
  • the candidate technical solution sentence extraction unit is configured to extract at least one candidate technical solution sentence from the claim part and/or the specification part of the target patent document;
  • the candidate technical solution sentence determining unit is configured to determine, for each candidate technical solution sentence in the at least one candidate technical solution sentence, the dominant feature group of the candidate technical solution sentence, and to compare the candidate technology according to the dominant feature group. Scoring or categorizing the plan sentences; according to the results of the scoring or classification, determine whether the above candidate technical plan sentences are technical plan sentences;
  • the solution information generating unit is configured to generate solution information based on the determined technical solution sentence.
  • a device for generating component content information includes:
  • the claim sentence extraction unit is configured to extract the claim sentence from the target patent document
  • the claim determination unit is configured to determine whether the above-mentioned claim sentence corresponds to a product claim or a method claim;
  • the component information extraction unit is configured to extract component information from the claim sentence when it is determined that the claim sentence corresponds to the product claim;
  • the component relationship information extraction unit is configured to extract component relationship information between components represented by the component information from the claim sentence;
  • the component content information generating unit is configured to generate component content information based on the component information and the component relationship information.
  • An electronic device including:
  • One or more processors are One or more processors;
  • a storage device on which one or more programs are stored
  • the one or more processors implement the method described above.
  • a computer-readable medium is also provided, on which a computer program is stored, wherein the program is executed by a processor to realize the method as described above.
  • FIG. 1 is a schematic diagram of an application scenario of a method for generating patent summary information provided by an embodiment of the present disclosure
  • FIG. 2 is a flowchart of a method for generating patent summary information provided by an embodiment of the present disclosure
  • FIG. 3 is a flowchart of a method for generating technical problem information according to an embodiment of the present disclosure
  • FIG. 5 is a flowchart of a method for generating content information of parts provided by an embodiment of the present disclosure
  • Fig. 6 is a flowchart of a method for generating beneficial effect information provided by an embodiment of the present disclosure
  • FIG. 7 is a flowchart of a method for generating technical field information provided by an embodiment of the present disclosure.
  • FIG. 8 is a schematic structural diagram of a device for generating patent summary information provided by an embodiment of the present disclosure.
  • FIG. 9 is a schematic structural diagram of an apparatus for generating technical problem information according to an embodiment of the present disclosure.
  • FIG. 10 is a schematic structural diagram of an apparatus for generating solution information provided by an embodiment of the present disclosure.
  • FIG. 11 is a schematic structural diagram of a device for generating content information of parts and components provided by an embodiment of the present disclosure
  • FIG. 12 is a schematic structural diagram of an apparatus for generating beneficial effect information according to an embodiment of the present disclosure.
  • FIG. 13 is a schematic structural diagram of an apparatus for generating technical field information according to an embodiment of the present disclosure
  • FIG. 14 is a schematic structural diagram of an electronic device provided by an embodiment of the present disclosure.
  • first and second mentioned in this disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependence. relationship.
  • Fig. 1 is a schematic diagram of an application scenario of a method for generating patent summary information provided by an embodiment of the present disclosure.
  • the user first selects a patent document as the target patent document.
  • the electronic device 101 shown as a server in the figure
  • analyzes the target patent document and extracts technical field information, technical problem information, solution information, and beneficial effect information.
  • patent summary information is generated.
  • the above-mentioned electronic device 101 may be hardware or software.
  • the electronic device When the electronic device is hardware, it can be implemented as a distributed cluster composed of multiple servers or terminal devices, or as a single server or a single terminal device.
  • the electronic device When the electronic device is embodied as software, it can be implemented as multiple software or software modules for providing distributed services, or as a single software or software module. There is no limitation here.
  • FIG. 2 shows a flow 200 of a method for generating patent summary information provided by an embodiment of the present disclosure.
  • the method of generating patent summary information includes the following steps:
  • Step 201 Extract technical problem information from the target patent document.
  • the execution body of the method for generating patent summary information may use multiple syntax analysis methods to extract at least one candidate sentence related to the technical problem from the target patent document.
  • the aforementioned grammatical analysis methods include but are not limited to at least one of the following: syntactic analysis, part-of-speech analysis, and reference resolution.
  • the execution subject may input each candidate sentence of the at least one candidate sentence into a pre-trained extraction model, and then determine whether the candidate sentence is a sentence of a predefined category.
  • the predefined category may be a category that is strongly related to technical issues.
  • the extraction model may be a machine learning model (for example, a classification task model).
  • the above extraction model can be trained through the following steps:
  • the training positive sample can be the accurate and strongly related sentences obtained by syntactic analysis, part-of-speech analysis, referential resolution and other grammatical analysis methods in the sample patent document.
  • the training negative samples can be sentences that are weakly related or irrelevant to the technical problem in the above sample patent documents.
  • the initial machine learning model is trained to obtain the extraction model.
  • the obtained extraction model can learn different characteristics of negative emotion expression and different forms of problem expression from training samples, so as to achieve the ability of generalization.
  • the above-mentioned execution subject may combine multiple candidate sentences of predefined categories to generate technical problem information.
  • Step 202 Extract solution information from the above-mentioned target patent document.
  • the above-mentioned executive body may extract at least one candidate technical solution sentence from the claim part and/or the specification part of the target patent document.
  • the execution subject may determine the dominant feature group of the candidate technical solution statement for each candidate technical solution statement in the at least one candidate technical solution statement, and compare the candidate technology according to the dominant feature set.
  • Scheme sentences are scored or classified.
  • the execution subject may determine whether the candidate technical solution sentence is a technical solution sentence based on the score; and generate solution information based on the determined technical solution sentence.
  • Step 203 Generate patent summary information based on the above-mentioned technical problem information and solution information.
  • the above-mentioned executive body may combine the above-obtained technical problem information and solution information to generate patent abstract information.
  • the above-mentioned extracting technical problem information from the target patent document includes: extracting at least one candidate sentence related to the technical problem from the above-mentioned target patent document; and from the above-mentioned at least one candidate sentence
  • the target candidate sentence is determined in the target candidate sentence, wherein the target candidate sentence includes at least one of the following: a sentence that is successfully compared with the subject of the target patent document, a sentence of a predefined category; and based on the determined target candidate sentence, technical problem information is generated.
  • the above-mentioned executive body may use multiple grammatical and syntactic analysis methods to extract at least one candidate sentence related to the technical problem from the target patent document. Then, for each candidate sentence in the at least one candidate sentence, a pre-trained extraction model is used to determine whether the candidate sentence is a sentence of a predefined category. Finally, multiple candidate sentences of predefined categories are combined to generate technical problem information.
  • the above extracting solution information from the target patent document includes: extracting at least one candidate technical solution sentence from the claim part and/or specification part of the target patent document; For each candidate technical solution sentence in the at least one candidate technical solution sentence, determine the dominant feature group of the candidate technical solution sentence, and score or classify the candidate technical solution sentence according to the dominant feature group; according to the score, It is determined whether the above candidate technical solution sentence is a technical solution sentence; based on the determined technical solution sentence, solution information is generated.
  • the above-mentioned executive body may perform keyword detection on the specification part of the above-mentioned target patent document, and then extract the sentence where the detected keyword is located as a candidate technical solution sentence. Then, according to the score, it is determined whether the candidate technical solution sentence is a technical solution sentence. Finally, based on the determined technical solution statement, solution information is generated.
  • the above method further includes: in response to the at least one candidate technical solution statement including a claim statement, determining whether the claim statement corresponds to a product claim or a method claim; in response to determining The claim sentence corresponds to the product claim, and the component information is extracted from the claim sentence; the component relationship information between the components represented by the component information is extracted from the claim sentence; based on the component information and The above-mentioned part relationship information generates part content information.
  • the above-mentioned execution subject can retrieve the claim sentence from the target patent document through keyword search. Then, the executive body can determine whether the above-mentioned claim sentence corresponds to the product claim. Finally, in response to determining that the above-mentioned claim sentence is a corresponding product claim, the executive body may extract the component information from the above-mentioned claim sentence.
  • the above method further includes: in response to determining that the claim statement corresponds to a method claim, extracting logical information from the claim statement to generate logical content information.
  • the above-mentioned logical information may be information used to characterize the logical relationship of each step in the method.
  • the above-mentioned execution subject can extract logical information from the above-mentioned claim sentence by searching for logical keywords.
  • the logical keywords can be artificially set according to the writing specifications of the claims. For example, the logical keyword can be "according to", "thereby", or "and".
  • the claim sentence includes "according to the obtained target word, get the word text.” Then, since it includes “according to” and “get”, the claim sentence in which the above "according to” and “get” are located can be regarded as Logical information is extracted. As another example, the claim sentence includes “sending information to the terminal device and displaying it on the display of the terminal device”, then, since it includes “and”, the claim sentence where the “and” is Can be extracted as logical information.
  • the above-mentioned execution subject may combine the above-mentioned extracted claim sentences to generate logical content information.
  • the foregoing generating solution information based on the determined technical solution sentence includes: generating solution information based on at least one of the following: the foregoing determined technical solution sentence, component content information , Logical content information.
  • the foregoing method further includes: extracting beneficial effect information from the foregoing target patent document; and/or extracting technical field information from the foregoing target patent document; the foregoing is based on the foregoing technical problem information, Solution information, generating patent overview information, including: generating patent overview information based on the above technical problem information, solution information, and the above beneficial effect information; and/or generating patent overview information based on the above technical problem information, solution information and the above technical field information Information; and/or generate patent summary information based on the above-mentioned technical problem information, solution information, beneficial effect information and the above-mentioned technical field information.
  • the above-mentioned extracting technical field information from the above-mentioned target patent document includes: determining the section in which the technical field information is located from the above-mentioned target patent document; and extracting the technical field from the above-mentioned section information.
  • the method for generating patent summary information may generate patent summary information in the following manner. First, extract the technical field information from the target patent document. Secondly, extract technical problem information from the above-mentioned target patent documents. Third, extract the solution information from the above-mentioned target patent document. From the second, the beneficial effect information is extracted from the above-mentioned target patent document. Finally, based on the above-mentioned technical field information, technical problem information, solution information, and beneficial effect information, patent summary information is generated. Through the extraction and integration of the above-mentioned information, the original patent document is streamlined, while the useful information is preserved, thereby saving the time for reading the patent document and improving the reading efficiency.
  • patents can be classified accurately (for example, patent classification based on the extracted technical fields, technical problems, solutions, beneficial effects, etc.).
  • the extraction and integration based on the above information will also help improve the accuracy of retrieval.
  • FIG. 3 shows a flow 300 of a method for generating technical problem information according to an embodiment of the present disclosure.
  • the method of generating technical problem information includes the following steps:
  • Step 301 Extract at least one candidate sentence related to the technical problem from the target patent document.
  • the above-mentioned executive body may extract the target patent document to extract at least one candidate sentence related to the technical problem.
  • Keywords can be words used to characterize technical problems, for example, "problems", “insufficiency”, “technical problems”, “technical points”.
  • Step 302 For each candidate sentence in the at least one candidate sentence, it is determined whether the candidate sentence is successfully compared with the subject of the target patent document.
  • the above-mentioned executive body may first determine the subject of the above-mentioned target patent document. After that, compare each candidate sentence with the above topics.
  • the subject of the above-mentioned target patent document can be the title of the above-mentioned target patent document, or the subject name in the claims.
  • the comparison can be a semantic comparison between the candidate sentence and the above topic. If the semantics match, the comparison is determined to be successful. For example, if the candidate sentence is "there is a problem of high traffic consumption" and the subject of the patent document is "methods to reduce traffic consumption", then it can be determined that the comparison is successful.
  • Step 303 Generate technical problem information based on candidate sentences that are successfully compared among the at least one candidate sentence.
  • the above-mentioned execution subject may combine multiple candidate sentences that have been successfully compared to generate technical problem information.
  • the method of generating technical problem information includes the following steps:
  • the extraction model may be a machine learning model (for example, a classification task model).
  • the above extraction model can be trained through the following steps:
  • the training positive sample can be the accurate and strongly related sentences obtained by syntactic analysis, part-of-speech analysis, referential resolution and other grammatical analysis methods in the sample patent document.
  • the training negative samples can be sentences that are weakly related or irrelevant to the technical problem in the above sample patent documents.
  • the initial machine learning model is trained to obtain the extraction model.
  • the obtained extraction model can learn different characteristics of negative emotion expression and different forms of problem expression from training samples, so as to achieve the ability of generalization.
  • multiple syntax analysis methods may be used to extract at least one candidate sentence related to the technical problem from the target patent document, and then input the at least one candidate sentence To the pre-trained extraction model.
  • the aforementioned grammatical analysis methods include but are not limited to at least one of the following: syntactic analysis, part-of-speech analysis, and reference resolution.
  • two extraction methods can be used to extract technical problems at the same time, and the final technical problem information is determined through arbitration of the two output results, so as to improve the accuracy of extraction.
  • One of the extraction methods may be the above-mentioned pre-trained extraction model, and the other extraction method may be the syntax analysis method.
  • FIG. 4 shows a flow 400 of a method for generating solution information provided by an embodiment of the present disclosure.
  • the method of generating solution information includes the following steps:
  • Step 401 Extract at least one candidate technical solution sentence from the claim part and/or the specification part of the target patent document.
  • the execution subject performs keyword detection on the claim part and/or specification part of the target patent document, and then extracts the sentence where the detected keyword is located as a candidate technical solution sentence.
  • the candidate technical solution sentence is a sentence related to the technical solution in the specification of the target patent document.
  • any sentence extracted from the specification part of the target patent document may be used as a candidate technical solution sentence.
  • Step 402 For each candidate technical solution sentence in at least one candidate technical solution sentence, perform the following sub-steps:
  • Sub-step 4021 for each candidate technical solution sentence in the at least one candidate technical solution sentence, determine the dominant feature group of the candidate technical solution sentence.
  • the above-mentioned executive body may determine the explicit feature group of the above-mentioned candidate technical solution sentence.
  • the dominant feature usually refers to the feature that can help distinguish the target object from the non-target object, such as the attribute or phenomenon of the object.
  • the dominant feature group may include at least one of the following dominant features: the degree of similarity between the candidate technical solution statement and the independent claim statement; the position information of the candidate technical solution statement in the aforementioned part of the specification; the candidate technical solution statement includes The number of keywords.
  • the similarity between the candidate technical solution sentence and the independent claim sentence can be determined by the Recall-Oriented Understudy for Gisting Evaluation (ROUGE) evaluation method.
  • the ROUGE evaluation index is one of the evaluation indexes in the field of automatic text summarization. It is used to indicate the degree of similarity between the automatic text summary generator and the summary written by experts, and mainly refers to the similarity in the use of words.
  • the ROUGE evaluation method evaluates abstracts based on the co-occurrence information of n-grams in the abstracts. It is an evaluation method oriented to the recall rate of n-grams.
  • the similarity between the candidate technical solution sentence and the independent claim sentence can also be determined by the following method: determining whether the vocabulary combination use of any few words in the candidate technical solution sentence appears in the independent claim.
  • This lexical collocation is a collocation of any two adjacent words or three or more words adjacent to each other.
  • the position information of the candidate technical solution sentence in the description part can be obtained by the following formula: the position of the candidate technical solution sentence in the sentence sequence in the specification/the number of sentences in the entire specification.
  • the above sentence sequence is a sequence composed of the sentences in the entire specification arranged in the order of appearance.
  • the keywords included in the candidate technical solution sentence may be the vocabulary in the independent claim (for example, independent claim 1).
  • Such vocabulary may include nominal phrases, verb phrases, adjectives and adverbs mentioned in the independent claims.
  • the candidate technical solution sentences are scored or classified according to the dominant feature group.
  • the above-mentioned execution subject may use each dominant feature in the above-mentioned dominant feature group to score the above-mentioned candidate technical solution sentences respectively. Based on the obtained multiple scores, a summary score of the above candidate technical solution sentences is obtained.
  • a classifier is used to classify the aforementioned candidate technical solution sentences according to the aforementioned dominant feature group.
  • the classifier is pre-trained by using annotated sample data to help identify the explicit features of the target sentence and a specific algorithm.
  • sub-step 4023 it is determined whether the candidate technical solution sentence is a technical solution sentence.
  • the above-mentioned execution subject may determine the candidate technical solution sentence with the highest score as the technical solution sentence according to the score obtained in the above step 4022, or determine the candidate technical solution sentence with a score higher than a preset threshold as the technical solution Statement.
  • the preset threshold may be preset according to actual experience.
  • the above-mentioned execution subject may determine whether the above-mentioned candidate technical solution sentence is a technical solution sentence according to the classification result obtained in the above-mentioned step 4022.
  • Step 403 Generate solution information based on the determined technical solution sentence.
  • the above-mentioned execution subject may combine the determined technical solution statements to generate solution information.
  • FIG. 5 shows a process 500 of a method for generating content information of a component provided by an embodiment of the present disclosure.
  • the method for generating component content information includes the following steps:
  • step 501 the claim sentence is extracted from the target patent document.
  • the executor of the method for generating component content information may first extract the claims from the target patent document through keyword search or structured text extraction. Then, the execution subject can recognize specific characters and extract claim sentences from the claims.
  • the keywords can be artificially set according to the writing specifications of the claims.
  • the keyword can be "claims.”
  • the specific character can be the digital number of each claim in the claims, or it can be the period at the end of each claim in the claims. Through such a number or period, the claim sentence corresponding to each claim can be extracted from the above-mentioned claims.
  • Step 502 Determine whether the above-mentioned claim sentence corresponds to a product claim or a method claim.
  • the above-mentioned executive body may determine whether the above-mentioned claim statement corresponds to a product claim in a variety of ways.
  • the execution subject may perform type keyword detection on the claim sentence to determine whether the claim sentence contains a method type keyword. In response to determining that the claim sentence does not contain the method type keyword, it is determined that the claim sentence corresponds to the product claim.
  • Type keyword detection is to detect the above-mentioned claim sentence to determine whether the type keyword can be found.
  • type keywords may include method type keywords and product type keywords.
  • Method type keywords can be words used to prompt methods, processes, and usage.
  • the product type keywords can be words used to indicate the relationship between parts and parts.
  • the method type keyword can be "method", "process”, or "step”.
  • the claim sentence includes "a method for peeling apples”. Then, since it includes "method”, the above-mentioned claim sentence generally corresponds to the method claim sentence.
  • the claim sentence includes "a pot for boiling soup", then, since it does not include the method type keyword, the above claim sentence generally corresponds to the product claim. Because method claims often include method type keywords, and such claims often include descriptions of multiple implementation steps. This feature can be used to determine whether the claim is a method claim. If the claim does not belong to a method claim, it belongs to a product claim.
  • the above-mentioned executive body may use machine learning to train a classifier to complete the judgment of the type of claim sentence.
  • the training data labeled with the claim type can be used to train the initial Convolutional Neural Networks (CNN), and the trained convolutional neural network can be used as a classifier.
  • CNN Convolutional Neural Networks
  • the generalization ability of the classifier obtained in this way is better than the recognition method based on type keywords.
  • the training data contains special case data, the classifier can also deal with this kind of special case well.
  • the trained classifier can more easily help classify more refined claim types (such as chemistry, computer software, etc.).
  • targeted processing methods such as domain-related dictionaries or word segmentation tools
  • Step 503 in response to determining that the claim sentence corresponds to the product claim, extract component information from the claim sentence.
  • the execution subject may extract component information from the claim statement.
  • the above-mentioned executive body may use a trained neural network to extract component information.
  • the above neural network can be trained using labeled data.
  • the marked content is the claim sentence and the parts or other components directly included in the claim sentence. Parts usually refer to parts of a product.
  • the component information may include but is not limited to at least one of the following: the name of the component, and the specification information of the component (for example, length, width, and height).
  • the above-mentioned execution subject may extract component information from the above-mentioned claim sentence based on a predefined claim writing rule.
  • a claim writing rule that is, "write the part information of important parts (such as part names) at the beginning of the sentence after splitting with a semicolon or a newline character".
  • the part information of important parts appears at the beginning of the clause separated by a comma or all appear only in a preceding clause. Therefore, it is necessary to consider these situations and use different rules for extraction in different situations.
  • Step 504 extracting component relationship information between components represented by the component information from the statement of the above claims.
  • the above-mentioned executive body may use some common relationship expressions to extract the component relationship information between the components represented by the above-mentioned component information.
  • the component relationship information can describe the relationship between one component and another component in the product. Taking the component relationship such as "include relationship” as an example, the relationship expression can usually be expressed as follows: “A has/include/comprise B", or the corresponding Chinese form "A includes B”.
  • the parallel structure for example, "the present invention includes a case, battery and engine”.
  • articles such as “a/the” in English) or context (such as “a”, “said” in Chinese) to distinguish and use similarity (such as edit distance) to share Refers to digestion.
  • "a pen” corresponds to "the pen”.
  • Such problems can also be solved in other ways.
  • a regular approach can be used.
  • the structure of "xx, xxx, and xx" is used to express the parallel relationship.
  • Syntactic analysis can also be used to assist in solving this type of problem, such as using dependency analysis to determine whether multiple components depend on the same verb (for example, "has” or “has”).
  • Step 505 Based on the above-mentioned component information and the above-mentioned component relationship information, component content information is generated.
  • the above-mentioned execution subject may generate component content information based on the component information obtained in the above step 503 and the component relationship information obtained in the above step 504.
  • component information and component relationship information can be directly combined together as component content information.
  • FIG. 6 shows a flow 600 of a method for generating beneficial effect information provided by an embodiment of the present disclosure.
  • the method for generating beneficial effect information includes the following steps:
  • Step 601 Extract at least one candidate sentence from the specification part of the target patent document according to the predefined beneficial effect sentence pattern.
  • the predefined beneficial effect sentence pattern may be a sentence pattern preset according to writing rules or writing habits and related to the beneficial effects of the patent document.
  • the predefined beneficial effect sentence pattern may be the following sentence pattern: "predicate verb + noun".
  • the "predicate verb” here can be, for example, a verb such as "improve” or “enhance”.
  • the "noun” here may be, for example, a noun such as "performance” or "effect”.
  • Step 602 Screen the at least one candidate sentence mentioned above according to the first predefined screening rule.
  • the matching rules may be set first.
  • some specific "predicate verbs” can only be used with specific “nouns”.
  • provide verb group 1 “reduce”, “avoid”, “eliminate”, “decrease”; noun group 1: “loss”, “failure ( Failure)”, “pollution”; Verb 2: “improve”, “enhance”, “increase”; Noun 2: “quality”, “comfort (ease)”, “efficiency (efficiency)”.
  • the collocation rule can be: verbs in verb group 1 can only be collocated with nouns in noun group 1, and verbs in verb group 2 can only be collocated with nouns in noun group 2.
  • the first predefined screening rule may be to remove candidate sentences that do not meet the collocation rule.
  • Step 603 Generate beneficial effect information based on the remaining candidate sentences in the at least one candidate sentence after the screening.
  • the above-mentioned execution subject may combine the remaining candidate sentences to generate beneficial effect information.
  • generating beneficial effect information based on the candidate sentences remaining in the at least one candidate sentence after screening includes: determining the context of the remaining candidate sentences in the target patent document; The above-mentioned remaining candidate sentences and the above-mentioned context generate beneficial effect information.
  • the foregoing method further includes: determining the morphological characteristics of the remaining candidate sentences in the target patent document; At least one of a predefined screening rule is adjusted.
  • FIG. 7 shows a flow 700 of a method for generating technical field information according to an embodiment of the present disclosure.
  • the method for generating technical field information includes the following steps:
  • step 701 at least one candidate sentence is selected from the target patent document according to the sentence pattern of the predefined technical field.
  • the predefined technical field sentence pattern may be a pre-set sentence pattern related to the technical field of the patent document according to writing rules or writing habits.
  • the predefined technical field sentence pattern may be the following sentence pattern: "'invention/publication/application' +'involved/belonging to/about'".
  • the predefined technical field sentence pattern can also be the following sentence pattern: "‘invention/disclosure/application’+‘be verb/preposition’+‘relates/pertain/concern’”.
  • it can be used as a candidate sentence.
  • Step 702 Screen the above-mentioned at least one candidate sentence according to the second predefined screening rule.
  • the second predefined screening rule may be set first.
  • the second pre-defined screening rule may be: if a specific word appears in the candidate sentence, remove the above candidate sentence.
  • the above-mentioned specific words may include but are not limited to at least one of the following: claims, not, person skilled in the art.
  • the second predefined screening rule may also be: if the candidate sentence appears in the description of the content of the drawing, the above candidate sentence is removed.
  • Step 703 Generate technical field information based on the remaining candidate sentences in the at least one candidate sentence after the screening.
  • the above-mentioned execution subject may combine the remaining candidate sentences to generate technical field information.
  • FIG. 8 is a schematic structural diagram of a device for generating patent summary information provided by an embodiment of the present disclosure.
  • the device embodiment corresponds to the method embodiment shown in FIG. 2, and the device can be applied to a variety of In electronic equipment.
  • the apparatus 800 for generating patent summary information in some embodiments includes: a technical problem extracting unit 801, a solution extracting unit 802, and a patent summary information generating unit 803.
  • the technical problem extraction unit 801 is configured to extract technical problem information from the target patent document
  • the solution extraction unit 802 is configured to extract solution information from the target patent document
  • the patent summary information generating unit 803 is configured To generate patent summary information based on the above-mentioned technical problem information and solution information.
  • FIG. 9 is a schematic structural diagram of a device for generating technical problem information provided by an embodiment of the present disclosure.
  • the device embodiment corresponds to the method embodiment shown in FIG. 3, and the device can be applied to a variety of In electronic equipment.
  • the apparatus 900 for generating technical problem information in some embodiments includes: a candidate sentence extraction unit 901, a candidate sentence determination unit 902 and a technical problem information generation unit 903.
  • the candidate sentence extraction unit 901 is configured to extract at least one candidate sentence related to the technical problem from the target patent document
  • the candidate sentence determination unit 902 is configured to determine the target candidate sentence from the above at least one candidate sentence, wherein The target candidate sentence includes at least one of the following: a sentence that is successfully compared with the subject of the target patent document, and a sentence of a predefined category
  • the technical problem information generating unit 903 is configured to generate a technology based on the determined target candidate sentence Problem information.
  • FIG. 10 is a schematic structural diagram of a device for generating solution information provided by an embodiment of the present disclosure.
  • This device embodiment corresponds to the method embodiment shown in FIG. 4, and the device can be applied to a variety of In electronic equipment.
  • the apparatus 1000 for generating solution information in some embodiments includes: a candidate technical solution sentence extracting unit 1001, a candidate technical solution sentence determining unit 1002, and a solution information generating unit 1003.
  • the candidate technical solution sentence extracting unit 1001 is configured to extract at least one candidate technical solution sentence from the claim part and/or specification part of the target patent document;
  • the candidate technical solution sentence determining unit 1002 is configured to extract at least one candidate technical solution sentence.
  • the solution information generating unit 1003 For each candidate technical solution sentence in a candidate technical solution sentence, determine the dominant feature group of the candidate technical solution sentence, and score or classify the candidate technical solution sentence according to the dominant feature group; determine the candidate according to the score Whether the technical solution sentence is a technical solution sentence; the solution information generating unit 1003 is configured to generate solution information based on the determined technical solution sentence.
  • FIG. 11 is a schematic structural diagram of a device for generating content information of parts and components provided by an embodiment of the present disclosure. This device embodiment corresponds to the method embodiment shown in FIG. Kind of electronic equipment.
  • the device 1100 for generating component content information in some embodiments includes: a claim sentence extraction unit 1101, a claim determination unit 1102, a component information extraction unit 1103, a component relationship information extraction unit 1104, and a component Content information generating unit 1105.
  • the claim sentence extraction unit 1101 is configured to extract the claim sentence from the target patent document; the claim determination unit 1102 is configured to determine whether the claim sentence corresponds to a product claim or a method claim; parts information
  • the extraction unit 1103 is configured to extract component information from the claim sentence in response to determining the corresponding product claim; the component relationship information extraction unit 1104 is configured to extract the component information representation from the claim sentence
  • the component relationship information between the components; the component content information generating unit 1105 is configured to generate component content information based on the above-mentioned component information and the above-mentioned component relationship information.
  • FIG. 12 is a schematic structural diagram of a device for generating beneficial effect information provided by an embodiment of the present disclosure.
  • the device embodiment corresponds to the method embodiment shown in FIG. 6, and the device can be applied to a variety of In electronic equipment.
  • the apparatus 1200 for generating beneficial effect information in some embodiments includes: a beneficial effect candidate sentence extraction unit 1201, a beneficial effect candidate sentence screening unit 1202, and a beneficial effect information generating unit 1203.
  • the beneficial effect candidate sentence extraction unit 1201 is configured to extract at least one candidate sentence from the specification part of the target patent document according to a predefined beneficial effect sentence pattern
  • the beneficial effect candidate sentence selection unit 1202 is configured to extract at least one candidate sentence according to the first
  • a predefined screening rule is used to screen the at least one candidate sentence
  • the beneficial effect information generating unit 1203 is configured to generate beneficial effect information based on the remaining candidate sentences in the at least one candidate sentence after screening.
  • FIG. 13 is a schematic structural diagram of a device for generating technical field information provided by an embodiment of the present disclosure.
  • the device embodiment corresponds to the method embodiment shown in FIG. 7, and the device can be applied to a variety of In electronic equipment.
  • the apparatus 1300 for generating technical field information includes: a technical field candidate sentence extraction unit 1301, a technical field candidate sentence screening unit 1302, and a technical field information generating unit 1303.
  • the technical field candidate sentence extraction unit 1301 is configured to select at least one candidate sentence from the target patent document according to a predefined technical field sentence pattern
  • the technical field candidate sentence screening unit 1302 is configured to filter according to a second predefined According to the rules, the above-mentioned at least one candidate sentence is screened
  • the technical field information generating unit 1303 is configured to generate technical field information based on the remaining candidate sentences in the above-mentioned at least one candidate sentence after screening.
  • FIG. 14 is a schematic structural diagram of an electronic device (such as the electronic device in FIG. 1) 1400 provided by an embodiment of the present disclosure.
  • the electronic device shown in FIG. 14 is only an example, and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.
  • the electronic device 1400 may include a processing device (such as a central processing unit, a graphics processor, etc.) 1401, which may be based on a program stored in a read-only memory (Read-Only Memory, ROM) 1402 or from a storage device 1408 is loaded into the program in random access memory (Random Access Memory, RAM) 1403 to execute various appropriate actions and processes.
  • a processing device such as a central processing unit, a graphics processor, etc.
  • RAM random access memory
  • various programs and data required for the operation of the electronic device 1400 are also stored.
  • the processing device 1401, ROM 1402, and RAM 1403 are connected to each other through a bus 1404.
  • An Input/Output (I/O) interface 1405 is also connected to the bus 1404.
  • the following devices can be connected to the I/O interface 1405: including input devices 1406 such as touch screens, touch pads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; including, for example, liquid crystal displays (LCD) Output devices 1407 such as speakers, vibrators, etc.; storage devices 1408 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1409.
  • the communication device 1409 may allow the electronic device 1400 to perform wireless or wired communication with other devices to exchange data.
  • FIG. 14 shows an electronic device 1400 with multiple devices, it is not required to implement or have all the devices shown. It may be implemented alternatively or provided with more or fewer devices. Each block shown in FIG. 14 may represent one device, or may represent multiple devices as needed.
  • the process described above with reference to the flowchart may be implemented as a computer software program.
  • some embodiments of the present disclosure include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart.
  • the computer program may be downloaded and installed from the network through the communication device 1409, or installed from the storage device 1408, or installed from the ROM 1402.
  • the processing device 1401 the above-mentioned functions defined in the methods of some embodiments of the present disclosure are executed.
  • the computer-readable medium described in some embodiments of the present disclosure may be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two.
  • the computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or a combination of any of the above.
  • Examples of computer-readable storage media may include, but are not limited to: electrical connections with one or more wires, portable computer disks, hard disks, RAM, ROM, Erasable Programmable Read-Only Memory (EPROM) Or flash memory), optical fiber, portable compact disc read-only memory (Compact Disc-ROM, CD-ROM), optical storage device, magnetic storage device, or any suitable combination of the above.
  • the computer-readable storage medium may be any tangible medium that contains or stores a program, and the program may be used by or in combination with an instruction execution system, apparatus, or device.
  • a computer-readable signal medium may include a data signal propagated in a baseband or as a part of a carrier wave, and a computer-readable program code is carried therein. This propagated data signal can take many forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the foregoing.
  • the computer-readable signal medium may also be any computer-readable medium other than the computer-readable storage medium.
  • the computer-readable signal medium may send, propagate or transmit the program for use by or in combination with the instruction execution system, apparatus, or device .
  • the program code contained on the computer-readable medium can be transmitted by any suitable medium, including but not limited to: wire, optical cable, radio frequency (RF), etc., or any suitable combination of the foregoing.
  • the client and server can communicate with any known or future-developed network protocol such as HyperText Transfer Protocol (HTTP), and can communicate with any form or medium of digital data (
  • HTTP HyperText Transfer Protocol
  • communication networks are interconnected.
  • Examples of communication networks include Local Area Network (LAN), Wide Area Network (WAN), the Internet (for example, the Internet), and end-to-end networks (for example, ad hoc end-to-end networks), and any known Or a network developed in the future.
  • the above-mentioned computer-readable medium may be included in the above-mentioned electronic device; or it may exist alone without being assembled into the electronic device.
  • the aforementioned computer-readable medium carries one or more programs. When the aforementioned one or more programs are executed by the electronic device, the electronic device: extracts technical problem information from the target patent document; extracts solutions from the aforementioned target patent document Program information; based on the above-mentioned technical problem information and solution information, generate patent summary information.
  • the computer program code used to perform the operations of some embodiments of the present disclosure can be written in one or more programming languages or a combination thereof.
  • the above-mentioned programming languages include object-oriented programming languages such as Java, Smalltalk, C++, Also includes conventional procedural programming languages-such as "C" language or similar programming languages.
  • the program code can be executed entirely on the user's computer, partly on the user's computer, executed as an independent software package, partly on the user's computer and partly executed on a remote computer, or entirely executed on the remote computer or server.
  • the remote computer may be connected to the user's computer through any kind of network including LAN or WAN, or may be connected to an external computer (for example, using an Internet service provider to connect through the Internet).
  • each block in the flowchart or block diagram can represent a module, program segment, or part of code, and the module, program segment, or part of code contains one or more for realizing the specified logic function.
  • Executable instructions can be included in the blocks in the flowchart or block diagram.
  • the functions marked in the block may also occur in a different order from the order marked in the drawings. For example, two blocks shown one after another can actually be executed substantially in parallel, and they can sometimes be executed in the reverse order, depending on the functions involved.
  • each block in the block diagram and/or flowchart, and the combination of the blocks in the block diagram and/or flowchart can be implemented by a dedicated hardware-based system that performs the specified functions or operations Or it can be realized by a combination of dedicated hardware and computer instructions.
  • the units described in some embodiments of the present disclosure may be implemented in software or hardware.
  • the described unit can also be provided in the processor, for example, it can be described as: a processor includes a technical problem extraction unit, a solution extraction unit, and a patent summary information generation unit. Among them, the names of these units do not constitute a limitation on the unit itself in one case.
  • the technical problem extraction unit can also be described as "a unit for extracting technical problem information from the target patent document".
  • exemplary types of hardware logic components include: Field Programmable Gate Array (FPGA), Application Specific Integrated Circuit (ASIC), and application specific standard products (Application Specific Standard Parts, ASSP), System-on-a-Chip (SOC), Complex Programmable Logic Device (CPLD), etc.
  • FPGA Field Programmable Gate Array
  • ASIC Application Specific Integrated Circuit
  • ASSP Application Specific Standard Parts
  • SOC System-on-a-Chip
  • CPLD Complex Programmable Logic Device

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • Tourism & Hospitality (AREA)
  • Technology Law (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Library & Information Science (AREA)
  • Strategic Management (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Operations Research (AREA)
  • Economics (AREA)
  • Human Resources & Organizations (AREA)
  • Marketing (AREA)
  • Primary Health Care (AREA)
  • General Business, Economics & Management (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Artificial Intelligence (AREA)
  • Computational Linguistics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Machine Translation (AREA)

Abstract

生成专利概述信息的方法、装置、电子设备和计算机可读介质。生成专利概述信息的方法包括：从上述目标专利文档中提取技术问题信息（201）；从上述目标专利文档中提取解决方案信息（202）；基于上述技术问题信息和上述解决方案信息，生成专利摘要信息（203）。

Description

生成专利概述信息的方法、装置、电子设备和介质
本申请要求在2019年12月19日提交中国专利局、申请号为201911319575.9的中国专利申请的优先权，该申请的全部内容通过引用结合在本申请中。
技术领域
本公开涉及计算机技术领域，例如涉及生成专利概述信息的方法、装置、电子设备和介质。
背景技术
随着互联网技术的发展，互联网用户从网络中获取到的信息量越来越大，如何从海量信息中快速获取到有效信息一直是信息检索领域的研究热点。文本摘要技术是一种常见的提取有效信息的技术，其通常是利用计算机来处理自然语言文本，自动从上述自然语言文本中提取能准确地反映出文本的中心内容的部分内容。这样的提取有效信息的技术有助于降低互联网用户所面临的信息过载问题，帮助用户更快、更有效地从互联网定位到需要的信息。
发明内容
本公开提出了生成专利概述信息的方法、装置、电子设备和介质，以从海量专利信息中获得有效专利信息。
提供了一种生成专利概述信息的方法，该方法包括：
从目标专利文档中提取技术问题信息；
从上述目标专利文档中提取解决方案信息；
基于上述技术问题信息和上述解决方案信息，生成专利概述信息。
还提供了一种生成技术问题信息的方法，该方法包括：
从目标专利文档中提取出与技术问题相关的至少一个候选语句；
从上述至少一个候选语句中确定目标候选语句，其中，上述目标候选语句包括以下至少一项：与上述目标专利文档的主题比对成功的语句，预定义类别的语句；
基于所确定的目标候选语句，生成技术问题信息。
还提供了一种生成解决方案信息的方法，该方法包括：
从目标专利文档的权利要求部分和/或说明书部分中提取出至少一个候选技术方案语句；
对于上述至少一个候选技术方案语句中的每一个候选技术方案语句，确定上述候选技术方案语句的显性特征组，以及根据上述显性特征组对上述候选技术方案语句进行评分或分类；
根据评分或分类的结果，确定上述候选技术方案语句是否为技术方案语句；
基于确定的技术方案语句，生成解决方案信息。
还提供了一种生成零部件内容信息的方法，该方法包括：
从目标专利文档中提取出权利要求语句；
确定上述权利要求语句是对应产品权利要求还是方法权利要求；
响应于确定上述权利要求语句对应产品权利要求，从上述权利要求语句中提取零部件信息；
从上述权利要求语句中提取出上述零部件信息表征的零部件之间的零部件关系信息；
基于上述零部件信息和上述零部件关系信息，生成零部件内容信息。
还提供了一种生成专利概述信息的装置，装置包括：
技术问题提取单元，被配置成从目标专利文档中提取技术问题信息；
解决方案提取单元，被配置成从上述目标专利文档中提取解决方案信息；专利概述信息生成单元，被配置成基于上述技术问题信息和上述解决方案信息，生成专利概述信息。
还提供了一种生成技术问题信息的装置，装置包括：
候选语句提取单元，被配置成从目标专利文档中提取出与技术问题相关的至少一个候选语句；
候选语句确定单元，被配置成从上述至少一个候选语句中确定目标候选语句，其中，上述目标候选语句包括以下至少一项：与上述目标专利文档的主题比对成功的语句，预定义类别的语句；
技术问题信息生成单元，被配置成基于所确定的目标候选语句，生成技术问题信息。
还提供了一种生成解决方案信息的装置，装置包括：
候选技术方案语句提取单元，被配置成从目标专利文档的权利要求部分和/ 或说明书部分中提取出至少一个候选技术方案语句；
候选技术方案语句确定单元，被配置成对于上述至少一个候选技术方案语句中的每一个候选技术方案语句，确定上述候选技术方案语句的显性特征组，以及根据上述显性特征组对上述候选技术方案语句进行评分或分类；根据评分或分类的结果，确定上述候选技术方案语句是否为技术方案语句；
解决方案信息生成单元，被配置成基于确定的技术方案语句，生成解决方案信息。
还提供了一种生成零部件内容信息的装置，装置包括：
权利要求语句提取单元，被配置成从目标专利文档中提取出权利要求语句；
权利要求确定单元，被配置成确定上述权利要求语句是对应产品权利要求还是方法权利要求；
零部件信息提取单元，被配置成在确定上述权利要求语句对应产品权利要求的情况下，从上述权利要求语句中提取零部件信息；
零部件关系信息提取单元，被配置成从上述权利要求语句中提取出上述零部件信息表征的零部件之间的零部件关系信息；
零部件内容信息生成单元，被配置成基于上述零部件信息和上述零部件关系信息，生成零部件内容信息。
还提供了一种电子设备，包括：
一个或多个处理器；
存储装置，其上存储有一个或多个程序；
当一个或多个程序被一个或多个处理器执行，使得一个或多个处理器实现如上所述的方法。
还提供了一种计算机可读介质，其上存储有计算机程序，其中，程序被处理器执行时实现如上所述的方法。
附图说明
图1是本公开的一实施例提供的一种生成专利概述信息的方法的应用场景的示意图；
图2是本公开的一实施例提供的一种生成专利概述信息的方法的流程图；
图3是本公开的一实施例提供的一种生成技术问题信息的方法的流程图；
图4是本公开的一实施例提供的一种生成解决方案信息的方法的流程图；
图5是本公开的一实施例提供的一种生成零部件内容信息的方法的流程图；
图6是本公开的一实施例提供的一种生成有益效果信息的方法的流程图；
图7是本公开的一实施例提供的一种生成技术领域信息的方法的流程图；
图8是本公开的一实施例提供的一种生成专利概述信息的装置的结构示意图；
图9是本公开的一实施例提供的一种生成技术问题信息的装置的结构示意图；
图10是本公开的一实施例提供的一种生成解决方案信息的装置的结构示意图；
图11是本公开的一实施例提供的一种生成零部件内容信息的装置的结构示意图；
图12是本公开的一实施例提供的一种生成有益效果信息的装置的结构示意图；
图13是本公开的一实施例提供的一种生成技术领域信息的装置的结构示意图；
图14是本公开的一实施例提供的一种电子设备的结构示意图。
具体实施方式
下面将参照附图描述本公开的实施例。虽然附图中显示了本公开的一些实施例，但本公开可以通过多种形式来实现。
为了便于描述，附图中仅示出了与有关公开相关的部分。
本公开中提及的“第一”、“第二”等概念仅用于对不同的装置、模块或单元进行区分，并非用于限定这些装置、模块或单元所执行的功能的顺序或者相互依存关系。
本公开中提及的“一个”、“多个”的修饰是示意性而非限制性的，除非在上下文另有明确指出，否则应该理解为“一个或多个”。
本公开实施方式中的多个装置之间所交互的消息或者信息的名称仅用于说明性的目的，而并不是用于对这些消息或信息的范围进行限制。
下面将参考附图并结合实施例来说明本公开。
图1是本公开的一实施例提供的一种生成专利概述信息的方法的应用场景的示意图。在图1的应用场景100中，用户首先选取一个专利文档作为目标专 利文档。然后，电子设备101(图中示出为服务器)对目标专利文档进行分析，提取技术领域信息、技术问题信息、解决方案信息和有益效果信息。最后，基于上述技术领域信息、技术问题信息、解决方案信息和有益效果信息，生成专利概述信息。
上述电子设备101可以是硬件，也可以是软件。当电子设备为硬件时，可以实现成多个服务器或终端设备组成的分布式集群，也可以实现成单个服务器或单个终端设备。当电子设备体现为软件时，可以实现成例如用来提供分布式服务的多个软件或软件模块，也可以实现成单个软件或软件模块。在此不做限定。
参考图2，图2示出了本公开的一实施例提供的一种生成专利概述信息的方法的流程200。该生成专利概述信息的方法，包括以下步骤：
步骤201，从目标专利文档中提取技术问题信息。
在一些实施例中，生成专利概述信息的方法的执行主体(例如图1所示的电子设备101)可以使用多种语法分析方式从目标专利文档中提取出与技术问题相关的至少一个候选语句。上述语法分析方式包括但不限于以下至少一项：句法分析，词性分析，指代消解。
在一些实施例中，上述执行主体可以将上述至少一个候选语句中的每个候选语句输入预先训练的抽取模型，然后确定上述候选语句是否是预定义类别的语句。其中，预定义类别可以是与技术问题强相关的类别。
抽取模型可以是一种机器学习模型(例如分类任务模型)。上述抽取模型可以通过如下步骤训练得到：
第一，获取训练正样本的集合和训练负样本的集合。
训练正样本可以是样本专利文档中之前通过句法分析，词性分析，指代消解等语法分析方式获取的准确的、与技术问题强相关的句子。训练负样本可以是上述样本专利文档中与技术问题弱相关或不相关的句子。
第二，基于上述训练正样本的集合和训练负样本的集合，对初始的机器学习模型进行训练，得到抽取模型。
得到的抽取模型可以从训练样本中学习负面情感表达的不同特征，以及学习问题表达的不同形式，从而达到泛化的能力。
在一些实施例中，作为示例，上述执行主体可以将预定义类别的多条候选语句进行组合，生成技术问题信息。
步骤202，从上述目标专利文档中提取解决方案信息。
在一些实施例中，上述执行主体可以从目标专利文档的权利要求部分和/或说明书部分中提取出至少一个候选技术方案语句。
在一些实施例中，上述执行主体可以对于上述至少一个候选技术方案语句中的每一个候选技术方案语句，确定上述候选技术方案语句的显性特征组，以及根据上述显性特征组对上述候选技术方案语句进行评分或分类。
在一些实施例中，上述执行主体可以根据评分，确定上述候选技术方案语句是否为技术方案语句；基于确定的技术方案语句，生成解决方案信息。
步骤203，基于上述技术问题信息、解决方案信息，生成专利概述信息。
在一些实施例中，上述执行主体可以将上述得到的技术问题信息、和解决方案信息组合，生成专利摘要信息。
在一些实施例的一些可选的实现方式中，上述从目标专利文档中提取技术问题信息，包括：从上述目标专利文档中提取出与技术问题相关的至少一个候选语句；从上述至少一个候选语句中确定目标候选语句，其中，上述目标候选语句包括以下至少一项：与上述目标专利文档的主题比对成功的语句，预定义类别的语句；基于所确定的目标候选语句，生成技术问题信息。
作为示例，首先，上述执行主体可以使用多种语法句法分析方式从目标专利文档中提取出与技术问题相关的至少一个候选语句。然后，对于上述至少一个候选语句中的每个候选语句，利用预先训练的抽取模型，确定上述候选语句是否是预定义类别的语句。最后，将预定义类别的多条候选语句进行组合，生成技术问题信息。
在一些实施例的一些可选的实现方式中，上述从上述目标专利文档中提取解决方案信息，包括：从目标专利文档的权利要求部分和/或说明书部分中提取出至少一个候选技术方案语句；对于上述至少一个候选技术方案语句中的每一个候选技术方案语句，确定上述候选技术方案语句的显性特征组，以及根据上述显性特征组对上述候选技术方案语句进行评分或分类；根据评分，确定上述候选技术方案语句是否为技术方案语句；基于确定的技术方案语句，生成解决方案信息。
作为示例，首先，上述执行主体可以对上述目标专利文档的说明书部分进行关键词检测，然后将检测到的关键词所在的语句作为候选技术方案语句提取出来。然后，根据评分，确定上述候选技术方案语句是否为技术方案语句。最后，基于确定的技术方案语句，生成解决方案信息。
在一些实施例的一些可选的实现方式中，上述方法还包括：响应于上述至少一个候选技术方案语句包括权利要求语句，确定上述权利要求语句是对应产 品权利要求还是方法权利要求；响应于确定上述权利要求语句对应产品权利要求，从上述权利要求语句中提取零部件信息；从上述权利要求语句中提取出上述零部件信息表征的零部件之间的零部件关系信息；基于上述零部件信息和上述零部件关系信息，生成零部件内容信息。
作为示例，首先，上述执行主体可以通过关键词检索，从目标专利文档中提取出权利要求语句。然后，执行主体可以确定上述权利要求语句是否对应产品权利要求。最后，响应于确定上述权利要求语句是对应产品权利要求，执行主体可以从上述权利要求语句中提取零部件信息。
在一些实施例的一些可选的实现方式中，上述方法还包括：响应于确定上述权利要求语句对应方法权利要求，从上述权利要求语句中提取逻辑信息以生成逻辑内容信息。其中，上述逻辑信息可以是用于表征方法中每个步骤的逻辑关系的信息。作为示例，上述执行主体可以通过逻辑关键词的检索，从上述权利要求语句中提取逻辑信息。其中，逻辑关键词可以是根据权利要求的撰写规范而人为设定的。例如，逻辑关键词可以是“根据”，也可以是“从而”，还可以是“以及”。作为示例，权利要求语句中包括“根据所得到的目标单词，得到单词文本”，那么，由于其包括“根据”和“得到”，则上述“根据”和“得到”所在的权利要求语句可以作为逻辑信息被提取。作为另一种示例，权利要求语句中包括“将信息发送给终端设备，以及在上述终端设备的显示器上进行显示”，那么，由于其包括“以及”，则上述“以及”所在的权利要求语句可以作为逻辑信息被提取。上述执行主体可以将上述提取到的权利要求语句进行组合，生成逻辑内容信息。
在一些实施例的一些可选的实现方式中，上述基于确定的技术方案语句，生成解决方案信息，包括：基于以下至少一项，生成解决方案信息：上述确定的技术方案语句，零部件内容信息，逻辑内容信息。
在一些实施例的一些可选的实现方式中，上述方法还包括：从上述目标专利文档中提取有益效果信息；和/或从上述目标专利文档中提取技术领域信息；上述基于上述技术问题信息、解决方案信息，生成专利概述信息，包括：基于上述技术问题信息、解决方案信息和上述有益效果信息生成专利概述信息；和/或基于上述技术问题信息、解决方案信息和上述技术领域信息生成专利概述信息；和/或基于上述技术问题信息、解决方案信息、有益效果信息和上述技术领域信息生成专利概述信息。
在一些实施例的一些可选的实现方式中，上述从上述目标专利文档中提取技术领域信息，包括：从上述目标专利文档中确定出技术领域信息所处的章节；从上述章节中提取技术领域信息。
本公开的一些实施例提供的生成专利概述信息的方法，可以按如下方式生成专利概述信息。首先，从目标专利文档中提取技术领域信息。其次，从上述目标专利文档中提取技术问题信息。再次，从上述目标专利文档中提取解决方案信息。从次，从上述目标专利文档中提取有益效果信息。最后，基于上述技术领域信息、技术问题信息、解决方案信息和有益效果信息，生成专利概述信息。通过上述信息的提取和整合，实现了对原专利文档的精简，同时还保留了有用的信息，由此可以节省对专利文档的阅读时间，提高了阅读效率。同时，基于上述信息的提取和整合，能够准确地将专利进行分类(例如，基于提取的技术领域、技术问题、解决方案、有益效果等进行专利分类)。另外，基于上述信息的提取和整合，还有助于提高检索的精准性。
参考图3，图3示出了本公开的一实施例提供的一种生成技术问题信息的方法的流程300。该生成技术问题信息的方法，包括以下步骤：
步骤301，从目标专利文档中提取出与技术问题相关的至少一个候选语句。
在一些实施例中，上述执行主体可以对目标专利文档进行提取，提取出与技术问题相关的至少一个候选语句。其中，对目标专利文档进行提取可以首先对目标专利文档进行关键词检索。然后，提取关键词所在的语句作为候选语句。关键词可以是用于表征技术问题的词语，例如，“问题”，“不足”，“技术问题”，“技术要点”。
步骤302，对于上述至少一个候选语句中的每个候选语句，确定上述候选语句与上述目标专利文档的主题是否比对成功。
在一些实施例中，上述执行主体可以首先确定上述目标专利文档的主题。之后，将每个候选语句与上述主题进行比对。上述目标专利文档的主题可以是上述目标专利文档的标题，也可以是权利要求中的主题名称。在这里，比对可以是将候选语句与上述主题进行语义比对，如果语义相符，则确定比对成功。例如，候选语句是“存在流量消耗大的问题”，专利文档的主题是“减少流量消耗的方法”，那么，可以确定比对成功。
步骤303，基于上述至少一个候选语句中比对成功的候选语句，生成技术问题信息。
在一些实施例中，作为示例，上述执行主体可以将比对成功的多条候选语句进行组合，生成技术问题信息。
在一些实施例中，生成技术问题信息的方法，包括以下步骤：
将目标专利文档输入预先训练的抽取模型，然后确定技术问题信息。
抽取模型可以是一种机器学习模型(例如分类任务模型)。上述抽取模型可以 通过如下步骤训练得到：
第一，获取训练正样本的集合和训练负样本的集合。
训练正样本可以是样本专利文档中之前通过句法分析，词性分析，指代消解等语法分析方式获取的准确的、与技术问题强相关的句子。训练负样本可以是上述样本专利文档中与技术问题弱相关或不相关的句子。
第二，基于上述训练正样本的集合和训练负样本的集合，对初始的机器学习模型进行训练，得到抽取模型。
得到的抽取模型可以从训练样本中学习负面情感表达的不同特征，以及学习问题表达的不同形式，从而达到泛化的能力。
在一些实施例中，将目标专利文档输入预先训练的抽取模型之前，可以使用多种语法分析方式从目标专利文档中提取出与技术问题相关的至少一个候选语句，然后将上述至少一个候选语句输入到预先训练的抽取模型。上述语法分析方式包括但不限于以下至少一项：句法分析，词性分析，指代消解。
在一些实施例中，可以采用两种抽取方式同时进行技术问题的提取，通过两个输出结果的仲裁确定最终的技术问题信息，以提高提取的准确性。其中一种抽取方式可以是上述预训练的抽取模型，另一种抽取方式可以是语法分析方式。
参考图4，图4示出了本公开的一实施例提供的一种生成解决方案信息的方法的流程400。该生成解决方案信息的方法，包括以下步骤：
步骤401，从目标专利文档的权利要求部分和/或说明书部分中提取出至少一个候选技术方案语句。
在一些实施例中，上述执行主体对上述目标专利文档的权利要求部分和/或说明书部分进行关键词检测，然后将检测到的关键词所在的语句作为候选技术方案语句提取出来。候选技术方案语句是目标专利文档的说明书中与技术方案相关的语句。
在一些实施例中，可以将从目标专利文档的说明书部分中提取的任一语句作为候选技术方案语句。
步骤402，对于至少一个候选技术方案语句中的每一个候选技术方案语句，执行如下子步骤：
子步骤4021，对于上述至少一个候选技术方案语句中的每一个候选技术方案语句，确定上述候选技术方案语句的显性特征组。
在一些实施例中，上述执行主体可以确定上述候选技术方案语句的显性特 征组。其中，显性特征通常是指能帮助区分目标物体和非目标物体的特征，例如物体的属性或现象。
可选地，显性特征组可以包括如下显性特征中的至少一项：候选技术方案语句与独立权利要求语句的相似度；候选技术方案语句在上述说明书部分的位置信息；候选技术方案语句包括的关键词的数量。
作为示例，候选技术方案语句与独立权利要求语句的相似度可以通过面向召回率摘要评估研究(Recall-Oriented Understudy for Gisting Evaluation，ROUGE)评价方法确定。ROUGE评估指标是自动文本摘要领域的评估指标之一，用于表示自动文本摘要生成器和专家撰写摘要之间的相似程度，主要是指字词使用上的相似度。ROUGE评价方法基于摘要中n元词(n-gram)的共现信息来评价摘要，是一种面向n元词召回率的评价方法。
可选地，候选技术方案语句与独立权利要求语句的相似度还可以通过如下方式确定：确定候选技术方案语句中的任意几个词的词汇搭配使用是否在独立权利要求中出现过。这种词汇搭配使用是任意两个相邻词或者相邻三个词或者更多词的搭配。
在专利文档的说明书中，会有部分段落涉及到本专利的重要技术方案点。这样的部分段落一般会出现在说明书中的发明内容的前几段，或者是具体实施方式中的一些实施例中。所以，技术方案的相关语句具有一定的位置特征，例如在特定的说明书版块中以及版块中靠前的段落中。专利文档的说明书中一般包括多个版块，例如技术领域，背景技术，发明内容和具体实施方式。利用这些版块信息可以帮助掌握该版块下的语句段落主要的核心主题是什么。又由于每个专利文档的长度都不一样，所以通过记录技术方案在一篇专利文档中的相对位置来作为技术方案的特征更为合理。作为示例，候选技术方案语句在上述说明书部分的位置信息可以通过如下公式得到：上述候选技术方案语句在说明书中的语句序列中的位置/整篇说明书的语句数量。其中，上述语句序列是整篇说明书的语句按照出现顺序排列组成的序列。
作为示例，候选技术方案语句包括的关键词可以是独立权利要求(例如独立权利要求1)中的词汇。这样的词汇可以包括独立权利要求里面提到的名词性短语，动词短语，形容词和副词等。
子步骤4022，根据上述显性特征组对上述候选技术方案语句进行评分或分类。
在一些实施例中，上述执行主体可以利用上述显性特征组中的每个显性特征对上述候选技术方案语句分别进行评分。再基于得到的多个评分，得到对上 述候选技术方案语句的汇总的评分。
在一些实施例中，利用分类器根据上述显性特征组对上述候选技术方案语句进行分类。该分类器是预先通过标注好的样本数据，帮助识别目标语句的显性特征和特定算法去训练得到的。
子步骤4023，根据评分，确定上述候选技术方案语句是否为技术方案语句。
在一些实施例中，上述执行主体可以根据上述步骤4022中得到的评分，将评分最高的候选技术方案语句确定为技术方案语句，或者将评分高于预设阈值的候选技术方案语句确定为技术方案语句。作为示例，预设阈值可以是根据实际经验预先设定的。
在一些实施例中，上述执行主体可以根据上述步骤4022中得到的分类结果确定上述候选技术方案语句是否为技术方案语句。
步骤403，基于确定的技术方案语句，生成解决方案信息。
在一些实施例中，作为示例，上述执行主体可以将确定的技术方案语句进行组合，生成解决方案信息。
参考图5，图5示出了本公开的一实施例提供的一种生成零部件内容信息的方法的流程500。该生成零部件内容信息的方法，包括以下步骤：
步骤501，从目标专利文档中提取出权利要求语句。
在一些实施例中，生成零部件内容信息的方法的执行主体首先可以通过关键词检索或结构化文本提取，从目标专利文档中提取出权利要求书。接着，上述执行主体可以识别特定字符，从上述权利要求书中提取权利要求语句。关键词可以是根据权利要求的撰写规范而人为设定的。例如，关键词可以是“权利要求”。特定字符可以是权利要求书中每项权利要求的数字编号，也可以是权利要求书中每项权利要求结尾的句号。通过这样的数字编号或者句号，能够将对应每项权利要求的权利要求语句从上述权利要求书中提取出来。
步骤502，确定上述权利要求语句是对应产品权利要求还是方法权利要求。
在一些实施例中，上述执行主体可以通过多种方式确定上述权利要求语句是否对应产品权利要求。作为示例，上述执行主体可以对上述权利要求语句进行类型关键词检测，确定上述权利要求语句中是否含有方法类型关键词。响应于确定上述权利要求语句中不含有方法类型关键词，确定上述权利要求语句对应产品权利要求。
类型关键词检测是对上述权利要求语句进行检测，确定能否找到类型关键词。作为示例，类型关键词可以包括方法类型关键词和产品类型关键词。方法 类型关键词可以是用于提示方法、进程、用法的词语。而产品类型关键词可以是用于提示零部件、零部件关系的词语。例如，方法类型关键词可以是“方法(method)”，也可以是“进程(process)”，还可以是“步骤(step)”。作为示例，权利要求语句中包括“一种用于削苹果的方法”，那么，由于其包括“方法”，则上述权利要求语句一般对应方法权利要求语句。又例如，权利要求语句中包括“一种用于煲汤的锅”，那么，由于其不包括方法类型关键词，则上述权利要求语句一般对应产品权利要求。因为方法权利要求常常包含方法类型关键词，同时这类权利要求中，常常包含对多个实现步骤的描述。可以通过这样的特点来判断权利要求是否为方法权利要求，如果权利要求不属于方法权利要求，则属于产品权利要求。
在一些实施例的一些可选的实现方式中，上述执行主体可以使用机器学习的方式训练出一个分类器完成权利要求语句所属类型的判断。例如，可以使用标注好权利要求类型的训练数据对初始的卷积神经网络(Convolutional Neural Networks，CNN)进行训练，将训练后的卷积神经网络作为分类器。这种方式得到的分类器的泛化能力相比较于基于类型关键词的识别方式更好。如果训练数据中包含特殊情况的数据，则分类器也可以很好地应对该类特殊情况。同时，训练出的分类器能够更容易的帮助分类出更精细的权利要求类型(如化学类，计算机软件类等等)。从而，有助于在之后的工作中使用针对性的处理方式(比如领域相关的字典或者分词工具)。
步骤503，响应于确定上述权利要求语句对应产品权利要求，从上述权利要求语句中提取零部件信息。
在一些实施例中，上述执行主体响应于确定上述权利要求语句对应产品权利要求，可以从上述权利要求语句中提取零部件信息。作为示例，上述执行主体可以使用训练后的神经网络进行零部件信息提取。上述神经网络可以使用标注好的数据训练得到。其中，标注的内容是权利要求语句以及其中直接被包含于权利要求语句的零部件或其他组成成分。零部件通常是指产品的部件。而零部件信息可以包括但不限于以下至少一项：零部件的名称，零部件的规格信息(例如，长、宽、高)。
在一些实施例的一些可选的实现方式中，上述执行主体可以基于预定义的权利要求书写规则，从上述权利要求语句中提取零部件信息。作为示例，可以有这样的权利要求书写规则，即“将重要零部件的零部件信息(例如零部件名)写在用分号或者换行符拆分后的子句中的句首”。但有些权利要求语句中，重要零部件的零部件信息出现在用逗号拆分的子句的句首或者全部只出现在一条靠前的子句中。因此，需要考虑这些情况，在不同情况下使用不同的规则进行 抽取。
步骤504，从上述权利要求语句中提取出上述零部件信息表征的零部件之间的零部件关系信息。
在一些实施例中，上述执行主体可以使用一些常见的关系表达方式提取出上述零部件信息表征的零部件之间的零部件关系信息。零部件关系信息可以描述产品中的一个零部件和另一个零部件之间的关系。以“包含关系”这样的零部件关系为例，关系表达方式通常可以如下体现：“A has/include/comprise B”，或者对应的中文形式“A包括B”。
有时也需要考虑到如下问题：第一，共同指代问题，例如“a pen,…,the pen has a hat”；第二，并列结构，例如，“本发明包括一个壳，电池和发动机”。对于共同指代问题，可以使用冠词(比如英文中的“a/the”)或者上下文(比如中文中的“一个”，“所述”)进行区分并且使用相似度(如编辑距离)进行共指消解。例如“a pen”对应上“the pen”。这类问题也可以使用其他的方式进行解决。对于并列结构，可以使用规则的方式。如利用“xx,xxx,and xx”结构来表达并列关系。也可以使用句法分析的方式辅助解决该类问题，比如使用依存分析(dependency parsing)判断多个零部件的是否依赖于同一个表示包含的动词(例如，“has”或“具有”)。
步骤505，基于上述零部件信息和上述零部件关系信息，生成零部件内容信息。
在一些实施例中，上述执行主体可以基于上述步骤503得到的零部件信息和上述步骤504得到的零部件关系信息，生成零部件内容信息。作为示例，可以将零部件信息和零部件关系信息直接组合在一起，作为零部件内容信息。
参考图6，图6示出了本公开的一实施例提供的一种生成有益效果信息的方法的流程600。该生成有益效果信息的方法，包括以下步骤：
步骤601，根据预定义有益效果句式，从目标专利文档的说明书部分中提取出至少一个候选语句。
在一些实施例中，预定义有益效果句式可以是根据书写规则或者书写习惯预先设定的、与专利文档的有益效果相关的句式。作为示例，预定义有益效果句式可以是如下句式：“谓语动词+名词”。这里的“谓语动词”例如可以是“改进”或者“提高”之类的动词。这里的“名词”例如可以是“性能”或者“效果”之类的名词。
对于目标专利文档的说明书部分中的任一语句，如果符合上述句式，则可以作为候选语句。
步骤602，根据第一预定义筛选规则，对上述至少一个候选语句进行筛选。
在一些实施例中，可以首先设定搭配规则。例如，一些特定的“谓语动词”只能搭配特定的“名词”。作为示例，提供动词组1：“减少(reduce)”，“避免(avoid)”，“消除(eliminate)”，“降低(decrease)”；名词组1：“损失(loss)”，“失败(failure)”，“污染(pollution)”；动词组2：“改进(improve)”，“增强(enhance)”，“提高(increase)”；名词组2：“质量(quality)”，“舒适性(ease)”，“效率(efficiency)”。则搭配规则可以为：动词组1里的动词只能与名词组1里面的名词进行搭配，动词组2里的动词只能与名词组2里面的名词进行搭配。
作为示例，第一预定义筛选规则可以是将不符合搭配规则的候选语句去除。
步骤603，基于筛选后上述至少一个候选语句中剩余的候选语句，生成有益效果信息。
在一些实施例中，作为示例，上述执行主体可以将剩余的候选语句进行组合，生成有益效果信息。
在一些实施例的一些可选的实现方式中，上述基于筛选后上述至少一个候选语句中剩余的候选语句，生成有益效果信息，包括：确定上述目标专利文档中上述剩余的候选语句的上下文；基于上述剩余的候选语句和上述上下文，生成有益效果信息。
在一些实施例的一些可选的实现方式中，上述方法还包括：确定上述目标专利文档中所述剩余的候选语句的形态特征；根据上述形态特征，对上述预定义有益效果句式和上述第一预定义筛选规则中的至少一项进行调整。
参考图7，图7示出了本公开的一实施例提供的一种生成技术领域信息的方法的流程700。该生成技术领域信息的方法，包括以下步骤：
步骤701，根据预定义技术领域句式，从目标专利文档中选取出至少一个候选语句。
在一些实施例中，预定义技术领域句式可以是根据书写规则或者书写习惯预先设定的、与专利文档的技术领域相关的句式。作为示例，预定义技术领域句式可以是如下句式：“‘发明/公开/申请’+‘涉及/属于/关于’”。或者，对于英文，预定义技术领域句式还可以是如下句式：“‘invention/disclosure/application’+‘be动词/介词’+‘relates/pertain/concern’”。对于目标专利文档的说明书部分中的任一语句，如果符合上述句式，则可以作为候选语句。
步骤702，根据第二预定义筛选规则，对上述至少一个候选语句进行筛选。
在一些实施例中，可以首先设定第二预定义筛选规则。作为示例，第二预定义筛选规则可以是：如果候选语句中出现了特定词，则将上述候选语句去除。例如，上述特定词可以包括但不限于以下至少一项：权利要求(claims)，不(not)，本领域技术人员(person skilled in the art)。又例如，第二预定义筛选规则还可以是：如果候选语句出现在对附图内容的描述中，则将上述候选语句去除。
步骤703，基于筛选后上述至少一个候选语句中剩余的候选语句，生成技术领域信息。
在一些实施例中，作为示例，上述执行主体可以将剩余的候选语句进行组合，生成技术领域信息。
参考图8，图8是本公开的一实施例提供的一种生成专利概述信息的装置的结构示意图，本装置实施例与图2所示的方法实施例相对应，该装置可以应用于多种电子设备中。
如图8所示，一些实施例的生成专利概述信息的装置800包括：技术问题提取单元801、解决方案提取单元802和专利概述信息生成单元803。其中，技术问题提取单元801，被配置成从目标专利文档中提取技术问题信息；解决方案提取单元802，被配置成从上述目标专利文档中提取解决方案信息；专利概述信息生成单元803，被配置成基于上述技术问题信息、解决方案信息，生成专利概述信息。
参考图9，图9是本公开的一实施例提供的一种生成技术问题信息的装置的结构示意图，本装置实施例与图3所示的方法实施例相对应，该装置可以应用于多种电子设备中。
如图9所示，一些实施例的生成技术问题信息的装置900包括：候选语句提取单元901、候选语句确定单元902和技术问题信息生成单元903。其中，候选语句提取单元901，被配置成从目标专利文档中提取出与技术问题相关的至少一个候选语句；候选语句确定单元902，被配置成从上述至少一个候选语句中确定目标候选语句，其中，上述目标候选语句包括以下至少一项：与上述目标专利文档的主题比对成功的语句，预定义类别的语句；技术问题信息生成单元903，被配置成基于所确定的目标候选语句，生成技术问题信息。
参考图10，图10是本公开的一实施例提供的一种生成解决方案信息的装置的结构示意图，本装置实施例与图4所示的方法实施例相对应，该装置可以应用于多种电子设备中。
如图10所示，一些实施例的生成解决方案信息的装置1000包括：候选技术方案语句提取单元1001、候选技术方案语句确定单元1002和解决方案信息生 成单元1003。其中，候选技术方案语句提取单元1001，被配置成从目标专利文档的权利要求部分和/或说明书部分中提取出至少一个候选技术方案语句；候选技术方案语句确定单元1002，被配置成对于上述至少一个候选技术方案语句中的每一个候选技术方案语句，确定上述候选技术方案语句的显性特征组，以及根据上述显性特征组对上述候选技术方案语句进行评分或分类；根据评分，确定上述候选技术方案语句是否为技术方案语句；解决方案信息生成单元1003，被配置成基于确定的技术方案语句，生成解决方案信息。
参考图11，图11是本公开的一实施例提供的一种生成零部件内容信息的装置的结构示意图，本装置实施例与图5所示的方法实施例相对应，该装置可以应用于多种电子设备中。
如图11所示，一些实施例的生成零部件内容信息的装置1100包括：权利要求语句提取单元1101、权利要求确定单元1102、零部件信息提取单元1103、零部件关系信息提取单元1104和零部件内容信息生成单元1105。其中，权利要求语句提取单元1101，被配置成从目标专利文档中提取出权利要求语句；权利要求确定单元1102，被配置成确定上述权利要求语句是对应产品权利要求还是方法权利要求；零部件信息提取单元1103，被配置成响应于确定对应产品权利要求，从上述权利要求语句中提取零部件信息；零部件关系信息提取单元1104，被配置成从上述权利要求语句中提取出上述零部件信息表征的零部件之间的零部件关系信息；零部件内容信息生成单元1105，被配置成基于上述零部件信息和上述零部件关系信息，生成零部件内容信息。
参考图12，图12是本公开的一实施例提供的一种生成有益效果信息的装置的结构示意图，本装置实施例与图6所示的方法实施例相对应，该装置可以应用于多种电子设备中。
如图12所示，一些实施例的生成有益效果信息的装置1200包括：有益效果候选语句提取单元1201、有益效果候选语句筛选单元1202和有益效果信息生成单元1203。其中，有益效果候选语句提取单元1201，被配置成根据预定义有益效果句式，从目标专利文档的说明书部分中提取出至少一个候选语句；有益效果候选语句筛选单元1202，被配置成根据第一预定义筛选规则，对上述至少一个候选语句进行筛选；有益效果信息生成单元1203，被配置成基于筛选后上述至少一个候选语句中剩余的候选语句，生成有益效果信息。
参考图13，图13是本公开的一实施例提供的一种生成技术领域信息的装置的结构示意图，本装置实施例与图7所示的方法实施例相对应，该装置可以应用于多种电子设备中。
如图13所示，一些实施例的生成技术领域信息的装置1300包括：技术领 域候选语句提取单元1301、技术领域候选语句筛选单元1302和技术领域信息生成单元1303。其中，技术领域候选语句提取单元1301，被配置成根据预定义技术领域句式，从目标专利文档中选取出至少一个候选语句；技术领域候选语句筛选单元1302，被配置成根据第二预定义筛选规则，对上述至少一个候选语句进行筛选；技术领域信息生成单元1303，被配置成基于筛选后上述至少一个候选语句中剩余的候选语句，生成技术领域信息。
参考图14，图14是本公开的一实施例提供的一种电子设备(例如图1中的电子设备)1400的结构示意图。图14示出的电子设备仅仅是一个示例，不应对本公开的实施例的功能和使用范围带来任何限制。
如图14所示，电子设备1400可以包括处理装置(例如中央处理器、图形处理器等)1401，其可以根据存储在只读存储器(Read-Only Memory，ROM)1402中的程序或者从存储装置1408加载到随机访问存储器(Random Access Memory，RAM)1403中的程序而执行多种适当的动作和处理。在RAM 1403中，还存储有电子设备1400操作所需的多种程序和数据。处理装置1401、ROM 1402以及RAM 1403通过总线1404彼此相连。输入/输出(Input/Output，I/O)接口1405也连接至总线1404。
通常，以下装置可以连接至I/O接口1405：包括例如触摸屏、触摸板、键盘、鼠标、摄像头、麦克风、加速度计、陀螺仪等的输入装置1406；包括例如液晶显示器(Liquid Crystal Display，LCD)、扬声器、振动器等的输出装置1407；包括例如磁带、硬盘等的存储装置1408；以及通信装置1409。通信装置1409可以允许电子设备1400与其他设备进行无线或有线通信以交换数据。虽然图14示出了具有多种装置的电子设备1400，并不要求实施或具备所有示出的装置。可以替代地实施或具备更多或更少的装置。图14中示出的每个方框可以代表一个装置，也可以根据需要代表多个装置。
根据本公开的一些实施例，上文参考流程图描述的过程可以被实现为计算机软件程序。例如，本公开的一些实施例包括一种计算机程序产品，其包括承载在计算机可读介质上的计算机程序，该计算机程序包含用于执行流程图所示的方法的程序代码。在这样的一些实施例中，该计算机程序可以通过通信装置1409从网络上被下载和安装，或者从存储装置1408被安装，或者从ROM 1402被安装。在该计算机程序被处理装置1401执行时，执行本公开的一些实施例的方法中限定的上述功能。
本公开的一些实施例所描述的计算机可读介质可以是计算机可读信号介质或者计算机可读存储介质或者是上述两者的任意组合。计算机可读存储介质例如可以是——但不限于——电、磁、光、电磁、红外线、或半导体的系统、装 置或器件，或者任意以上的组合。计算机可读存储介质的例子可以包括但不限于：具有一个或多个导线的电连接、便携式计算机磁盘、硬盘、RAM、ROM、可擦式可编程只读存储器(Erasable Programmable Read-Only Memory，EPROM或闪存)、光纤、便携式紧凑磁盘只读存储器(Compact Disc-ROM，CD-ROM)、光存储器件、磁存储器件、或者上述的任意合适的组合。在本公开的一些实施例中，计算机可读存储介质可以是任何包含或存储程序的有形介质，该程序可以被指令执行系统、装置或者器件使用或者与其结合使用。而在本公开的一些实施例中，计算机可读信号介质可以包括在基带中或者作为载波一部分传播的数据信号，其中承载了计算机可读的程序代码。这种传播的数据信号可以采用多种形式，包括但不限于电磁信号、光信号或上述的任意合适的组合。计算机可读信号介质还可以是计算机可读存储介质以外的任何计算机可读介质，该计算机可读信号介质可以发送、传播或者传输用于由指令执行系统、装置或者器件使用或者与其结合使用的程序。计算机可读介质上包含的程序代码可以用任何适当的介质传输，包括但不限于：电线、光缆、射频(Radio Frequency，RF)等等，或者上述的任意合适的组合。
在一些实施方式中，客户端、服务器可以利用诸如超文本传输协议(HyperText TransferProtocol，HTTP)之类的任何已知或未来研发的网络协议进行通信，并且可以与任意形式或介质的数字数据通信(例如，通信网络)互连。通信网络的示例包括局域网(Local Area Network，LAN)，广域网(Wide Area Network，WAN)，网际网(例如，互联网)以及端对端网络(例如，ad hoc端对端网络)，以及任何已知或未来研发的网络。
上述计算机可读介质可以是上述电子设备中所包含的；也可以是单独存在，而未装配入该电子设备中。上述计算机可读介质承载有一个或者多个程序，当上述一个或者多个程序被该电子设备执行时，使得该电子设备：从目标专利文档中提取技术问题信息；从上述目标专利文档中提取解决方案信息；基于上述技术问题信息、解决方案信息，生成专利概述信息。
可以以一种或多种程序设计语言或其组合来编写用于执行本公开的一些实施例的操作的计算机程序代码，上述程序设计语言包括面向对象的程序设计语言—诸如Java、Smalltalk、C++，还包括常规的过程式程序设计语言—诸如“C”语言或类似的程序设计语言。程序代码可以完全地在用户计算机上执行、部分地在用户计算机上执行、作为一个独立的软件包执行、部分在用户计算机上部分在远程计算机上执行、或者完全在远程计算机或服务器上执行。在涉及远程计算机的情形中，远程计算机可以通过任意种类的网络——包括LAN或WAN——连接到用户计算机，或者，可以连接到外部计算机(例如利用因特网服务提供商来通过因特网连接)。
附图中的流程图和框图，图示了按照本公开多种实施例的系统、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上，流程图或框图中的每个方框可以代表一个模块、程序段、或代码的一部分，该模块、程序段、或代码的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。在有些作为替换的实现中，方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如，两个接连地表示的方框实际上可以基本并行地执行，它们有时也可以按相反的顺序执行，这依所涉及的功能而定。也要注意的是，框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合，可以用执行规定的功能或操作的专用的基于硬件的系统来实现，或者可以用专用硬件与计算机指令的组合来实现。
描述于本公开的一些实施例中的单元可以通过软件的方式实现，也可以通过硬件的方式来实现。所描述的单元也可以设置在处理器中，例如，可以描述为：一种处理器包括技术问题提取单元、解决方案提取单元和专利概述信息生成单元。其中，这些单元的名称在一种情况下并不构成对该单元本身的限定，例如，技术问题提取单元还可以被描述为“从目标专利文档中提取技术问题信息的单元”。
本文中以上描述的功能可以至少部分地由一个或多个硬件逻辑部件来执行。例如，非限制性地，可以使用的示范类型的硬件逻辑部件包括：现场可编程门阵列(Field Programmable Gate Array，FPGA)、专用集成电路(Application Specific Integrated Circuit，ASIC)、专用标准产品(Application Specific Standard Parts，ASSP)、片上系统(System-on-a-Chip，SOC)、复杂可编程逻辑设备(ComplexProgrammable Logic Device，CPLD)等等。

Claims (22)

  1. 一种生成专利概述信息的方法，包括：
    从目标专利文档中提取技术问题信息；
    从所述目标专利文档中提取解决方案信息；
    基于所述技术问题信息和所述解决方案信息，生成专利概述信息。
  2. 根据权利要求1所述的方法，其中，所述从目标专利文档中提取技术问题信息，包括：
    将所述目标专利文档输入至预先训练的抽取模型，得到所述技术问题信息。
  3. 根据权利要求1所述的方法，其中，所述从目标专利文档中提取技术问题信息，包括：
    从所述目标专利文档中提取出与技术问题相关的至少一个候选语句；
    从所述至少一个候选语句中确定目标候选语句，其中，所述目标候选语句包括以下至少一项：与所述目标专利文档的主题比对成功的语句，预定义类别的语句；
    基于所确定的目标候选语句，生成所述技术问题信息。
  4. 根据权利要求1所述的方法，其中，所述从所述目标专利文档中提取解决方案信息，包括：
    从所述目标专利文档的权利要求部分和说明书部分中的至少之一中提取出至少一个候选技术方案语句；
    对于所述至少一个候选技术方案语句中的每一个候选技术方案语句，确定所述候选技术方案语句的显性特征组，以及根据所述显性特征组对所述候选技术方案语句进行评分或分类；根据评分或分类的结果，确定所述候选技术方案语句是否为技术方案语句；
    基于确定的技术方案语句，生成所述解决方案信息。
  5. 根据权利要求4所述的方法，还包括：
    响应于所述至少一个候选技术方案语句包括权利要求语句，确定所述权利要求语句是对应产品权利要求还是方法权利要求；
    响应于确定所述权利要求语句对应产品权利要求，从所述权利要求语句中提取零部件信息；
    从所述权利要求语句中提取出所述零部件信息表征的零部件之间的零部件关系信息；
    基于所述零部件信息和所述零部件关系信息，生成零部件内容信息。
  6. 根据权利要求5所述的方法，还包括：
    响应于确定所述权利要求语句对应方法权利要求，从所述权利要求语句中提取逻辑信息以生成逻辑内容信息。
  7. 根据权利要求6所述的方法，其中，所述基于确定的技术方案语句，生成所述解决方案信息，包括：
    基于以下至少一项，生成所述解决方案信息：所述确定的技术方案语句，所述零部件内容信息，所述逻辑内容信息。
  8. 根据权利要求1-7任一项所述的方法，还包括以下至少之一：
    从所述目标专利文档中提取有益效果信息；
    从所述目标专利文档中提取技术领域信息；
    所述基于所述技术问题信息和所述解决方案信息，生成专利概述信息，包括：
    基于所述技术问题信息、所述解决方案信息和所述有益效果信息生成所述专利概述信息；或，
    基于所述技术问题信息、所述解决方案信息和所述技术领域信息生成所述专利概述信息；或，
    基于所述技术问题信息、所述解决方案信息、所述有益效果信息和所述技术领域信息生成所述专利概述信息。
  9. 一种生成技术问题信息的方法，包括：
    从目标专利文档中提取出与技术问题相关的至少一个候选语句；
    从所述至少一个候选语句中确定目标候选语句，其中，所述目标候选语句包括以下至少一项：与所述目标专利文档的主题比对成功的语句，预定义类别的语句；
    基于所确定的目标候选语句，生成技术问题信息。
  10. 根据权利要求9所述的方法，其中，所述从目标专利文档中提取出与技术问题相关的至少一个候选语句，包括：
    对所述目标专利文档进行关键词检索；
    将关键词所在的语句确定为所述至少一个候选语句。
  11. 根据权利要求9所述的方法，其中，所述预定义类别的语句是通过如 下方式确定的：
    将所述至少一个候选语句中的每个候选语句输入至预先训练的抽取模型，确定所述候选语句是否是所述预定义类别的语句。
  12. 一种生成解决方案信息的方法，包括：
    从目标专利文档的权利要求部分和说明书部分中的至少之一中提取出至少一个候选技术方案语句；
    对于所述至少一个候选技术方案语句中的每一个候选技术方案语句，确定所述候选技术方案语句的显性特征组，以及根据所述显性特征组对所述候选技术方案语句进行评分或分类；根据评分或分类的结果，确定所述候选技术方案语句是否为技术方案语句；
    基于确定的技术方案语句，生成解决方案信息。
  13. 根据权利要求12所述的方法，其中，所述显性特征组包括以下至少一项：候选技术方案语句与独立权利要求语句的相似度；候选技术方案语句在所述说明书部分的位置信息；候选技术方案语句包括的关键词的数量。
  14. 一种生成零部件内容信息的方法，包括：
    从目标专利文档中提取出权利要求语句；
    确定所述权利要求语句是对应产品权利要求还是方法权利要求；
    响应于确定所述权利要求语句对应产品权利要求，从所述权利要求语句中提取零部件信息；
    从所述权利要求语句中提取出所述零部件信息表征的零部件之间的零部件关系信息；
    基于所述零部件信息和所述零部件关系信息，生成零部件内容信息。
  15. 根据权利要求14所述的方法，其中，所述从所述权利要求语句中提取零部件信息，包括：
    基于预定义的权利要求书写规则，从所述权利要求语句中提取所述零部件信息。
  16. 根据权利要求14所述的方法，其中，所述确定所述权利要求语句是对应产品权利要求还是方法权利要求，包括：
    对所述权利要求语句进行类型关键词检测，确定所述权利要求语句中是否含有方法类型关键词；
    响应于确定所述权利要求语句中含有方法类型关键词，确定所述权利要求 语句对应方法权利要求；
    响应于确定所述权利要求语句中不含方法类型关键词，确定所述权利要求语句对应产品权利要求。
  17. 一种生成专利概述信息的装置，包括：
    技术问题提取单元，被配置成从目标专利文档中提取技术问题信息；
    解决方案提取单元，被配置成从所述目标专利文档中提取解决方案信息；
    专利概述信息生成单元，被配置成基于所述技术问题信息和所述解决方案信息，生成专利概述信息。
  18. 一种生成技术问题信息的装置，包括：
    候选语句提取单元，被配置成从目标专利文档中提取出与技术问题相关的至少一个候选语句；
    候选语句确定单元，被配置成从所述至少一个候选语句中确定目标候选语句，其中，所述目标候选语句包括以下至少一项：与所述目标专利文档的主题比对成功的语句，预定义类别的语句；
    技术问题信息生成单元，被配置成基于所确定的目标候选语句，生成技术问题信息。
  19. 一种生成解决方案信息的装置，包括：
    候选技术方案语句提取单元，被配置成从目标专利文档的权利要求部分和说明书部分中的至少之一中提取出至少一个候选技术方案语句；
    候选技术方案语句确定单元，被配置成对于所述至少一个候选技术方案语句中的每一个候选技术方案语句，确定所述候选技术方案语句的显性特征组，以及根据所述显性特征组对所述候选技术方案语句进行评分或分类；根据评分或分类的结果，确定所述候选技术方案语句是否为技术方案语句；
    解决方案信息生成单元，被配置成基于确定的技术方案语句，生成解决方案信息。
  20. 一种生成零部件内容信息的装置，包括：
    权利要求语句提取单元，被配置成从目标专利文档中提取出权利要求语句；
    权利要求确定单元，被配置成确定所述权利要求语句是对应产品权利要求还是方法权利要求；
    零部件信息提取单元，被配置成响应于确定所述权利要求语句对应产品权利要求，从所述权利要求语句中提取零部件信息；
    零部件关系信息提取单元，被配置成从所述权利要求语句中提取出所述零部件信息表征的零部件之间的零部件关系信息；
    零部件内容信息生成单元，被配置成基于所述零部件信息和所述零部件关系信息，生成零部件内容信息。
  21. 一种电子设备，包括：
    至少一个处理器；
    存储装置，存储有至少一个程序；
    当所述至少一个程序被所述至少一个处理器执行，使得所述至少一个处理器实现如权利要求1-16中任一项所述的方法。
  22. 一种计算机可读介质，存储有计算机程序，其中，所述程序被处理器执行时实现如权利要求1-16中任一项所述的方法。
PCT/CN2020/130300 2019-12-19 2020-11-20 生成专利概述信息的方法、装置、电子设备和介质 WO2021120979A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP20901335.8A EP4080381A4 (en) 2019-12-19 2020-11-20 METHOD AND DEVICE FOR GENERATION OF SUMMARY INFORMATION ON PATENTS, AS WELL AS ELECTRONIC DEVICE AND MEDIA
US17/844,822 US20220365956A1 (en) 2019-12-19 2022-06-21 Method and apparatus for generating patent summary information, and electronic device and medium

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201911319575.9A CN111046168B (zh) 2019-12-19 2019-12-19 用于生成专利概述信息的方法、装置、电子设备和介质
CN201911319575.9 2019-12-19

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US17/844,822 Continuation US20220365956A1 (en) 2019-12-19 2022-06-21 Method and apparatus for generating patent summary information, and electronic device and medium

Publications (1)

Publication Number Publication Date
WO2021120979A1 true WO2021120979A1 (zh) 2021-06-24

Family

ID=70238012

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/130300 WO2021120979A1 (zh) 2019-12-19 2020-11-20 生成专利概述信息的方法、装置、电子设备和介质

Country Status (4)

Country Link
US (1) US20220365956A1 (zh)
EP (1) EP4080381A4 (zh)
CN (1) CN111046168B (zh)
WO (1) WO2021120979A1 (zh)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111046168B (zh) * 2019-12-19 2023-12-01 智慧芽信息科技(苏州)有限公司 用于生成专利概述信息的方法、装置、电子设备和介质
CN115640375A (zh) * 2022-10-11 2023-01-24 智慧芽信息科技(苏州)有限公司 专利文献中的技术问题抽取方法及相关设备

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109033074A (zh) * 2018-06-29 2018-12-18 北京百度网讯科技有限公司 新闻摘要生成方法、装置、设备及计算机可读介质
US20190057276A1 (en) * 2017-08-16 2019-02-21 Fuji Xerox Co., Ltd. Image analysis apparatus, image analysis method, and non-transitory computer readable medium
CN109509556A (zh) * 2018-11-09 2019-03-22 天津开心生活科技有限公司 知识图谱生成方法、装置、电子设备及计算机可读介质
CN110555440A (zh) * 2019-09-10 2019-12-10 杭州橙鹰数据技术有限公司 一种事件抽取方法及装置
CN111046168A (zh) * 2019-12-19 2020-04-21 智慧芽信息科技(苏州)有限公司 用于生成专利概述信息的方法、装置、电子设备和介质

Family Cites Families (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040059566A1 (en) * 2002-09-23 2004-03-25 Ronny Fogel Method of checking patent claims
TWI340329B (en) * 2004-10-01 2011-04-11 Inst Information Industry Patent summarization system, method and machine-readable storage medium
US20060074854A1 (en) * 2004-10-04 2006-04-06 Gosakan Aravamudan Innovative product design using application trees
US9336304B2 (en) * 2008-11-10 2016-05-10 Gary J. Speier Patent analytics system
US9892193B2 (en) * 2013-03-22 2018-02-13 International Business Machines Corporation Using content found in online discussion sources to detect problems and corresponding solutions
CN103885934B (zh) * 2014-02-19 2017-05-03 中国专利信息中心 一种专利文献关键短语自动提取方法
US9317498B2 (en) * 2014-05-23 2016-04-19 Codeq Llc Systems and methods for generating summaries of documents
US20170161372A1 (en) * 2015-12-04 2017-06-08 Codeq Llc Method and system for summarizing emails and extracting tasks
JP6842167B2 (ja) * 2017-05-08 2021-03-17 国立研究開発法人情報通信研究機構 要約生成装置、要約生成方法及びコンピュータプログラム
CN107622122A (zh) * 2017-09-27 2018-01-23 合肥博力生产力促进中心有限公司 一种专利关键词自动提取方法
CN110633264B (zh) * 2018-06-22 2023-08-11 蔡政育 应用专利数据库的研发辅助系统及其方法
US11348195B2 (en) * 2019-07-03 2022-05-31 Aon Risk Services, Inc. Of Maryland Analysis of intellectual-property data in relation to products and services

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190057276A1 (en) * 2017-08-16 2019-02-21 Fuji Xerox Co., Ltd. Image analysis apparatus, image analysis method, and non-transitory computer readable medium
CN109033074A (zh) * 2018-06-29 2018-12-18 北京百度网讯科技有限公司 新闻摘要生成方法、装置、设备及计算机可读介质
CN109509556A (zh) * 2018-11-09 2019-03-22 天津开心生活科技有限公司 知识图谱生成方法、装置、电子设备及计算机可读介质
CN110555440A (zh) * 2019-09-10 2019-12-10 杭州橙鹰数据技术有限公司 一种事件抽取方法及装置
CN111046168A (zh) * 2019-12-19 2020-04-21 智慧芽信息科技(苏州)有限公司 用于生成专利概述信息的方法、装置、电子设备和介质

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP4080381A4

Also Published As

Publication number Publication date
EP4080381A4 (en) 2022-12-14
EP4080381A1 (en) 2022-10-26
US20220365956A1 (en) 2022-11-17
CN111046168A (zh) 2020-04-21
CN111046168B (zh) 2023-12-01

Similar Documents

Publication Publication Date Title
US10586155B2 (en) Clarification of submitted questions in a question and answer system
US10713571B2 (en) Displaying quality of question being asked a question answering system
US10339453B2 (en) Automatically generating test/training questions and answers through pattern based analysis and natural language processing techniques on the given corpus for quick domain adaptation
US9798818B2 (en) Analyzing concepts over time
US8060357B2 (en) Linguistic user interface
US20180341871A1 (en) Utilizing deep learning with an information retrieval mechanism to provide question answering in restricted domains
US9373075B2 (en) Applying a genetic algorithm to compositional semantics sentiment analysis to improve performance and accelerate domain adaptation
US10671929B2 (en) Question correction and evaluation mechanism for a question answering system
US20160132572A1 (en) Collecting, organizing, and searching knowledge about a dataset
JP2015109068A (ja) 回答候補間の関係を識別および表示する機構、方法、コンピュータ・プログラム、ならびに装置
US9720962B2 (en) Answering superlative questions with a question and answer system
US20110078192A1 (en) Inferring lexical answer types of questions from context
WO2013088287A1 (en) Generation of natural language processing model for information domain
KR20160060253A (ko) 자연어 질의 응답 시스템 및 방법
US20220365956A1 (en) Method and apparatus for generating patent summary information, and electronic device and medium
JP2011118689A (ja) 検索方法及びシステム
JP2023507286A (ja) 自然言語クエリを構造化クエリ言語に変換するためのスキーマ注釈ファイルの自動作成
Hu et al. Text sentiment analysis: A review
Zhang et al. Event-based summarization method for scientific literature
Saneifar et al. Enhancing passage retrieval in log files by query expansion based on explicit and pseudo relevance feedback
Fauzi et al. Image understanding and the web: a state-of-the-art review
JP2011065380A (ja) 意見分類装置およびプログラム
CN115146025A (zh) 一种问答语句分类方法、终端设备及存储介质
Tu English complex verb constructions: identification and inference
Wang et al. Sweet Apple, company? or food? Adjective-centric commonsense knowledge acquisition with taxonomy-guided induction

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20901335

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2020901335

Country of ref document: EP

Effective date: 20220719