CN117350898A - Intelligent early warning system and method for annual patent fee - Google Patents
Intelligent early warning system and method for annual patent fee Download PDFInfo
- Publication number
- CN117350898A CN117350898A CN202311435774.2A CN202311435774A CN117350898A CN 117350898 A CN117350898 A CN 117350898A CN 202311435774 A CN202311435774 A CN 202311435774A CN 117350898 A CN117350898 A CN 117350898A
- Authority
- CN
- China
- Prior art keywords
- semantic
- value
- training
- related information
- text description
- Prior art date
- Legal status (The legal status 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 status listed.)
- Withdrawn
Links
- 238000000034 method Methods 0.000 title claims abstract description 21
- 238000012098 association analyses Methods 0.000 claims abstract description 20
- 239000013598 vector Substances 0.000 claims description 130
- 238000012549 training Methods 0.000 claims description 75
- 239000011159 matrix material Substances 0.000 claims description 11
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000001514 detection method Methods 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 3
- 230000011218 segmentation Effects 0.000 claims description 3
- 238000012937 correction Methods 0.000 claims description 2
- 238000013507 mapping Methods 0.000 claims description 2
- 238000005516 engineering process Methods 0.000 abstract description 7
- 238000012423 maintenance Methods 0.000 abstract description 7
- 238000010586 diagram Methods 0.000 description 6
- 238000011156 evaluation Methods 0.000 description 6
- 239000003795 chemical substances by application Substances 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 230000007246 mechanism Effects 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 230000002860 competitive effect Effects 0.000 description 3
- 238000003058 natural language processing Methods 0.000 description 3
- 230000000750 progressive effect Effects 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000013210 evaluation model Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 230000000670 limiting effect Effects 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000015654 memory Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 230000001902 propagating effect Effects 0.000 description 1
- 230000002829 reductive effect Effects 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
Classifications
-
- 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
- G06F40/30—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
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0206—Price or cost determination based on market factors
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Finance (AREA)
- Development Economics (AREA)
- Accounting & Taxation (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Technology Law (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- General Health & Medical Sciences (AREA)
- Economics (AREA)
- Artificial Intelligence (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- General Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Data Mining & Analysis (AREA)
- Primary Health Care (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Machine Translation (AREA)
Abstract
The patent annual fee intelligent early warning system and method are characterized in that related information of patent values, such as technical feature text description, market demand text description, competition situation text description and the like, is collected, semantic understanding technology is introduced into the rear end to conduct semantic association analysis of the related information of the patent values, so that the value score of the patent is judged, and the annual fee early warning level, such as high, medium, low and the like, is generated based on the value score and the annual fee payment state of the patent. Through the mode, different grades of early warning prompts can be intelligently carried out on the annual fee payment of the patent according to the value and the annual fee payment state of the patent so as to improve the management efficiency and the management level of the patent, optimize the maintenance cost and the income of the patent and protect the rights and interests of the patent.
Description
Technical Field
The application relates to the field of intelligent early warning, and more particularly, to an intelligent early warning system and method for annual patent fee.
Background
The patent is a legal tool for protecting innovation results, and a patent holder or an agent needs to pay patent annual fees on time to keep the validity of the patent. The payment of the annual patent fee has important significance for protecting the legal rights and interests of patentees and promoting technological innovation and social progress. However, because the annual fee payment period of the patent is longer, and the annual fee amount and the payment period of different patents are different, the patent holder or the agent often has difficulty in accurately grasping the annual fee payment condition of each patent, and the condition of missing payment, wrong payment or multiple payments easily occurs, so that unnecessary loss or risk is caused. In addition, because the value of the patent changes along with the changes of factors such as time, market, technology, competition and the like, in the existing annual fee payment monitoring and early warning system of the patent, a patent holder or an agent cannot dynamically adjust an annual fee payment strategy according to the value of the patent, so that the maintenance cost of the patent is higher and the benefit is lower.
Therefore, an optimized patent annual fee intelligent early warning system is desired.
Disclosure of Invention
The present application has been made in order to solve the above technical problems. The embodiment of the application provides an intelligent early warning system and method for annual fee of a patent, which are used for judging the value score of the patent by collecting related information of the patent, such as technical feature text description, market demand text description, competition situation text description and the like, introducing a semantic understanding technology at the rear end to carry out semantic association analysis of the related information of the patent, and generating an annual fee early warning grade, such as high, medium and low, based on the value score and the annual fee payment state of the patent. Through the mode, different grades of early warning prompts can be intelligently carried out on the annual fee payment of the patent according to the value and the annual fee payment state of the patent so as to improve the management efficiency and the management level of the patent, optimize the maintenance cost and the income of the patent and protect the rights and interests of the patent.
According to one aspect of the present application, there is provided an intelligent early warning system for annual fee, comprising:
the annual fee payment state detection module is used for acquiring the annual fee payment state of the patent to be evaluated;
the patent value information acquisition module is used for collecting value related information of the patent to be evaluated, wherein the value related information comprises technical feature text description, market demand text description and competition situation text description;
the patent value information semantic understanding module is used for respectively carrying out semantic coding on the technical feature text description, the market demand text description and the competition situation text description in the value related information so as to obtain technical feature semantic coding feature vectors, market demand semantic coding feature vectors and competition situation semantic coding feature vectors;
the patent value semantic feature association analysis module is used for carrying out semantic association analysis on the technical feature semantic coding feature vector, the market demand semantic coding feature vector and the competition situation semantic coding feature vector to obtain value related information global semantic coding features;
the patent value score calculation module is used for determining the value score of the patent to be evaluated based on the global semantic coding features of the value related information;
and the patent annual fee early warning module is used for determining the early warning level of the patent to be evaluated based on the decoding value and the annual fee payment state of the patent to be evaluated.
According to another aspect of the present application, there is provided an intelligent early warning method for annual fee, including:
acquiring annual fee payment states of the to-be-evaluated patent;
collecting value related information of the patent to be evaluated, wherein the value related information comprises technical feature text description, market demand text description and competition situation text description;
carrying out semantic coding on technical feature text description, market demand text description and competition situation text description in the value related information respectively to obtain technical feature semantic coding feature vectors, market demand semantic coding feature vectors and competition situation semantic coding feature vectors;
carrying out semantic association analysis on the technical feature semantic coding feature vector, the market demand semantic coding feature vector and the competition situation semantic coding feature vector to obtain global semantic coding features of the value related information;
determining the value score of the patent to be evaluated based on the global semantic coding features of the value related information;
and determining the early warning level of the to-be-evaluated patent based on the decoding value and the annual fee payment state of the to-be-evaluated patent.
Compared with the prior art, the intelligent early warning system and method for the annual fee of the patent provided by the application are used for judging the value score of the patent by collecting the related information of the value of the patent, such as technical feature text description, market demand text description, competition situation text description and the like, introducing a semantic understanding technology at the rear end to carry out semantic association analysis on the related information of the value of the patent, and generating the early warning grade of the annual fee, such as high, medium and low, based on the value score and the annual fee payment state of the patent. Through the mode, different grades of early warning prompts can be intelligently carried out on the annual fee payment of the patent according to the value and the annual fee payment state of the patent so as to improve the management efficiency and the management level of the patent, optimize the maintenance cost and the income of the patent and protect the rights and interests of the patent.
Drawings
The foregoing and other objects, features and advantages of the present application will become more apparent from the following more particular description of embodiments of the present application, as illustrated in the accompanying drawings. The accompanying drawings are included to provide a further understanding of embodiments of the application and are incorporated in and constitute a part of this specification, illustrate the application and not constitute a limitation to the application. In the drawings, like reference numerals generally refer to like parts or steps.
FIG. 1 is a block diagram of an intelligent early warning system for annual fee patent according to an embodiment of the present application;
FIG. 2 is a system architecture diagram of an intelligent early warning system for annual fee patent according to an embodiment of the present application;
FIG. 3 is a block diagram of a training phase of an intelligent early warning system for annual fee patent according to an embodiment of the present application;
fig. 4 is a flowchart of an intelligent early warning method for annual fee in patent according to an embodiment of the application.
Detailed Description
Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application and not all of the embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described herein.
As used in this application and in the claims, the terms "a," "an," "the," and/or "the" are not specific to the singular, but may include the plural, unless the context clearly dictates otherwise. In general, the terms "comprises" and "comprising" merely indicate that the steps and elements are explicitly identified, and they do not constitute an exclusive list, as other steps or elements may be included in a method or apparatus.
Although the present application makes various references to certain modules in a system according to embodiments of the present application, any number of different modules may be used and run on a user terminal and/or server. The modules are merely illustrative, and different aspects of the systems and methods may use different modules.
Flowcharts are used in this application to describe the operations performed by systems according to embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in order precisely. Rather, the various steps may be processed in reverse order or simultaneously, as desired. Also, other operations may be added to or removed from these processes.
Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application and not all of the embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described herein.
Because the annual fee payment period of the patent is longer, and the annual fee amount and the payment period of different patents are different, the annual fee payment condition of each patent is difficult to be accurately mastered by a patent holder or an agent, and the condition of missing payment, wrong payment or multiple payments is easy to occur, so that unnecessary loss or risk is caused. In addition, because the value of the patent changes along with the changes of factors such as time, market, technology, competition and the like, in the existing annual fee payment monitoring and early warning system of the patent, a patent holder or an agent cannot dynamically adjust an annual fee payment strategy according to the value of the patent, so that the maintenance cost of the patent is higher and the benefit is lower. Therefore, an optimized patent annual fee intelligent early warning system is desired.
In the technical scheme of the application, an intelligent early warning system for patent annual fee is provided. Fig. 1 is a block diagram of an intelligent early warning system for annual fee in patent according to an embodiment of the present application. Fig. 2 is a system architecture diagram of an intelligent early warning system for annual fee in patent according to an embodiment of the present application. As shown in fig. 1 and 2, an annual fee intelligent warning system 300 according to an embodiment of the present application includes: the annual fee payment state detection module 310 is configured to obtain an annual fee payment state of the patent to be evaluated; the patent value information collection module 320 is configured to collect value related information of the to-be-evaluated patent, where the value related information includes a technical feature text description, a market demand text description, and a competition situation text description; the patent value information semantic understanding module 330 is configured to perform semantic encoding on the technical feature text description, the market demand text description, and the competition situation text description in the value related information to obtain a technical feature semantic encoding feature vector, a market demand semantic encoding feature vector, and a competition situation semantic encoding feature vector; the patent value semantic feature association analysis module 340 is configured to perform semantic association analysis on the technical feature semantic coding feature vector, the market demand semantic coding feature vector and the competition situation semantic coding feature vector to obtain value related information global semantic coding features; the patent value score calculation module 350 is configured to determine a value score of the patent to be evaluated based on the global semantic coding feature of the value related information; and the patent annual fee early warning module 360 is configured to determine an early warning level of the patent to be evaluated based on the decoded value and an annual fee payment status of the patent to be evaluated.
In particular, the annual fee payment status detection module 310 is configured to obtain an annual fee payment status of the patent to be evaluated. It should be understood that the annual fee payment status of a patent refers to whether the applicant or patentee pays an annual fee of the patent according to a specification. The annual patent fee is a fee that needs to be paid periodically in order to maintain the effectiveness of the patent rights.
In particular, the patent value information collection module 320 is configured to collect value related information of the to-be-evaluated patent, where the value related information includes a technical feature text description, a market demand text description, and a competition situation text description. The value related information comprises technical feature text description, market demand text description and competition situation text description, so that value evaluation and annual fee payment early warning are carried out on the to-be-evaluated patent.
In particular, the patent value information semantic understanding module 330 is configured to perform semantic encoding on the technical feature text description, the market demand text description, and the competition text description in the value related information to obtain a technical feature semantic encoding feature vector, a market demand semantic encoding feature vector, and a competition semantic encoding feature vector. It is considered that the technical feature text description, the market demand text description and the competition situation text description in the value related information are all composed of words, and the words have semantic association feature information with context. Therefore, in the technical solution of the present application, in order to cooperate with semantic understanding features of each text description in the value related information, so as to perform the value evaluation and annual fee payment early warning of the to-be-evaluated patent, it is necessary to further perform semantic encoding on the technical feature text description, the market demand text description and the competition situation text description in the value related information, so as to extract semantic understanding feature information of each text description in the value related information, the market demand text description and the competition situation text description, respectively, thereby obtaining a technical feature semantic encoding feature vector, a market demand semantic encoding feature vector and a competition situation semantic encoding feature vector. Firstly, carrying out word segmentation processing on the technical feature text description, the market demand text description and the competition situation text description in the value related information so as to convert the technical feature text description, the market demand text description and the competition situation text description in the value related information into word sequences composed of a plurality of words; then, mapping each word in the word sequence into a word embedding vector by using an embedding layer of the semantic encoder comprising the word embedding layer to obtain a sequence of word embedding vectors; secondly, using a converter of the semantic encoder comprising a word embedding layer to carry out global context semantic encoding based on a converter thought on the sequence of the word embedding vectors so as to obtain a plurality of global context semantic feature vectors; and finally, cascading the global context semantic feature vectors to obtain the technical feature semantic coding feature vector, the market demand semantic coding feature vector and the competition situation semantic coding feature vector.
Notably, a semantic encoder is a model or algorithm for encoding context information into a semantic vector representation. It plays an important role in natural language processing tasks, and can capture semantic information of context, helping understanding and processing text data with context dependency. In natural language processing tasks, understanding the semantics of text requires consideration of context information, as the same word or phrase may have different meanings in different contexts. The semantic encoder converts text data into a vector representation with contextual semantic information by considering the context of the text. The design of semantic encoders may be based on different models and techniques, some of the common approaches including: cyclic neural network: RNNs are a type of neural network of recursive structure that can process sequence data and capture context information. By taking the text sequence as input, the RNN can interact the current word with the previous context information in each time step, encoding the context Wen Yuyi; long and short term memory network (LSTM): LSTM is a special type of RNN that can better address long-term dependency issues by introducing memory cells and gating mechanisms. LSTM is widely used in context semantic coding, and can effectively capture context information in text; gate cycle unit (GRU): GRU is another modified RNN structure similar to LSTM but with a simpler gating mechanism. The method has good performance in context semantic coding, and can effectively code the context information; transformer model: the transducer is a neural network model based on a self-attention mechanism and is widely applied to natural language processing tasks. It can simultaneously consider global context information in text, and associate vocabularies of different positions through a self-attention mechanism, so as to code context semantics.
In particular, the patent value semantic feature association analysis module 340 is configured to perform semantic association analysis on the technical feature semantic coding feature vector, the market demand semantic coding feature vector, and the competition situation semantic coding feature vector to obtain value related information global semantic coding features. In other words, in the technical scheme of the application, the technical feature semantic coding feature vector, the market demand semantic coding feature vector and the competition situation semantic coding feature vector are subjected to association coding, so that global collaborative association analysis of semantic features described in texts of all aspects of the value related information is performed, and value score calculation and annual fee early warning of the to-be-evaluated patent are performed more accurately. In particular, consider that conventional transducer models typically use overlapping local windows to calculate the attention of each window. However, such a partitioning strategy of overlapping windows may result in a lack of connection between windows, thereby limiting the expressive power and learning effect of the model. The Swin transducer solves the problem of lack of connection between windows by introducing a shift window method. The idea of shifting the windows is that in calculating the attention in each window, not only the pixels in the current window, but also the pixels of the surrounding windows are considered. Thus, cross-connection between windows is realized, so that the model can better capture global information of the text description semantic features of all aspects of the value related information. Based on the above, in the technical solution of the present application, the technical feature semantic coding feature vector, the market demand semantic coding feature vector and the competition situation semantic coding feature vector are further processed by a context encoder based on a Swin transform to obtain a value related information global semantic coding feature vector. In this way, semantic association and context information among technical feature text description, market demand text description and competition situation text description in the value related information can be captured, and the context encoder can consider interaction and dependency relationship among different features, so that global semantic encoding feature vectors of the value related information are generated to be used for representing the overall value related information of the patent to be evaluated. By the mode, the calculation efficiency of the model is improved, the complexity of the model is reduced, and the evaluation of the patent value and the annual fee payment early warning can be accurately carried out later.
It should be noted that, in other examples of the present application, the technical feature semantic coding feature vector, the market demand semantic coding feature vector, and the competition situation semantic coding feature vector may be subjected to semantic association analysis in the following manner to obtain value related information global semantic coding features, for example: converting the technical feature data into a semantic vector representation using an appropriate technical feature encoder (such as a text encoder, an image encoder or an audio encoder); the market demand data is converted to a semantic vector representation using an appropriate market demand data encoder. This may include preprocessing and feature extraction of the market demand text, which is then converted into semantic vectors using a corresponding text encoder; for competition situation data, preprocessing and feature extraction are needed to be carried out, and then the competition situation data are converted into semantic vector representations by using corresponding encoders; and carrying out semantic association analysis on the technical feature semantic coding feature vector, the market demand semantic coding feature vector and the competition situation semantic coding feature vector. This may be achieved by calculating the similarity or distance between the vectors. Common methods include cosine similarity, euclidean distance, manhattan distance, etc.; by integrating and integrating the results of the semantic association analysis, global semantic coding features of the value-related information can be obtained.
In particular, the patent value score calculation module 350 is configured to determine a value score of the patent to be evaluated based on the value-related information global semantic coding feature. In one example, the value-related information global semantically encoded feature vector is passed through a decoder to obtain a decoded value, which is used to represent a value score of the patent under evaluation. By inputting the value-related information global semantic coding feature vector into the decoder, the system can restore the specific value of the patent to be evaluated, and in particular, the decoded value can be regarded as an evaluation score of the patent, reflecting the value of the patent. Specifically, the decoder is used for carrying out decoding regression on the global semantic coding feature vector of the value related information in the following formula to obtain a decoding value for representing the value score of the patent to be evaluated; wherein, the formula is:wherein X represents the global semantic coding feature vector of the value related information, Y is the decoding value, W is a weight matrix,>representing matrix multiplication.
It is worth mentioning that in other examples of the present application, the value score of the patent under evaluation may be determined based on the value-related information global semantic coding features by, for example: collecting relevant information of the patent to be evaluated; converting the related information of the patent to be evaluated into a feature vector representation; carrying out semantic coding on the extracted feature vectors by using corresponding encoders to obtain global semantic coding feature vectors; a value assessment model is trained using the training data set and its corresponding known value scores. This may be a regression model, a classification model, or other suitable machine learning model, with the appropriate model being selected according to the specific task requirements; and inputting the global semantic coding feature vector of the patent to be evaluated into a trained value evaluation model to obtain a corresponding value score. The model can predict according to the patent sample with known value score, so as to evaluate the value of the patent to be evaluated; and according to the output result of the model, interpreting the value score of the patent to be evaluated.
In particular, the patent annual fee early warning module 360 is configured to determine an early warning level of the patent to be evaluated based on the decoding value and an annual fee payment status of the patent to be evaluated. That is, in the technical solution of the present application, the early warning level of the to-be-evaluated patent is determined based on the decoding value and the annual fee payment status of the to-be-evaluated patent. In this way, the value and risk of the patent can be more comprehensively evaluated, so that the early warning level of the patent can be determined. In particular, here, the early warning level is to judge whether the patent needs special attention or takes action according to the value score and annual fee payment status of the patent. Specific judgment standards can be defined according to actual demands, for example, different thresholds can be set, and when the decoding value is higher than a certain threshold and the annual fee payment state is poor, patents are classified as high risk early warning grades; when the decoding value is lower and the annual fee payment state is good, the patent is classified as a low-risk early warning grade. Through the mode, different grades of early warning prompts can be intelligently carried out on the annual fee payment of the patent according to the value and the annual fee payment state of the patent, so that the patent management efficiency and level are improved.
It will be appreciated that training of the semantic encoder containing word embedding layers, the Swin transducer based context encoder and the decoder is required before inference can be made using the neural network model described above. That is, the annual fee intelligent warning system 300 according to the present application further includes a training phase 400 for training the semantic encoder including the word embedding layer, the Swin transform-based context encoder, and the decoder.
Fig. 3 is a block diagram of a training phase of the annual fee intelligent warning system in accordance with an embodiment of the present application. As shown in fig. 3, an intelligent early warning system 300 for annual fee patent according to an embodiment of the present application includes: training phase 400, comprising: the training data acquisition unit 410 is configured to acquire training data, where the training data includes a training annual fee payment status of a to-be-evaluated patent, and training value related information of the to-be-evaluated patent, where the training value related information includes a training technical feature text description, a training market demand text description, and a training competition situation text description, and a true value of a value score of the to-be-evaluated patent; the training value related information semantic feature extraction unit 420 is configured to perform semantic coding on the training technical feature text description, the training market demand text description, and the training competition situation text description in the training value related information to obtain a training technical feature semantic coding feature vector, a training market demand semantic coding feature vector, and a training competition situation semantic coding feature vector; the training value related information semantic association coding unit 430 is configured to pass the training technical feature semantic encoding feature vector, the training market demand semantic encoding feature vector, and the training competition situation semantic encoding feature vector through the Swin transform-based context encoder to obtain a training value related information global semantic encoding feature vector; a decoding loss unit 440, configured to pass the training value related information global semantic coding feature vector through the decoder to obtain a decoding loss function value; and a model training unit 450, configured to train the semantic encoder including the word embedding layer, the context encoder based on Swin transform, and the decoder based on the decoding loss function value and propagating in a gradient descent direction, where, at each weight matrix iteration of the training, feature correction of a weight space is performed on the training value related information global semantic coding feature vector.
Wherein, pass the global semantic coding feature vector of the training value related information through the said decoder in order to get the decoding loss function value, include: performing decoding regression on the training value related information global semantic coding feature vector by using a decoder to obtain a training decoding value; and calculating a mean square error value between the training decoding value and a true value of the early warning level of the patent to be evaluated as the decoding loss function value.
In particular, in the technical solution of the present application, the technical feature semantic coding feature vector, the market demand semantic coding feature vector and the competitive situation semantic coding feature vector express the text semantic coding features of the technical feature text description, the market demand text description and the competitive situation text description in the value related information, respectively, so that after passing through the context encoder based on the Swin Transformer, the context semantic feature coding based on the global text semantic feature context of the technical feature text description, the market demand text description and the competitive situation text description can be further performed, so that the value related information global semantic coding feature vector has a multi-scale text context coding representation, that is, a feature representation with a semantic dense association in a text context semantic feature distribution dimension, which also causes a decrease in training efficiency of the weight of the decoder when the value related information global semantic coding feature vector performs decoding regression training through the decoder. Based on the above, the applicant of the present application performs resource-aware progressive context integration of the weight space for the value-related information global semantic coding feature vector when performing decoding regression training on the value-related information global semantic coding feature vector by a decoder:
wherein M is 1 And M 2 The weight matrix of the previous iteration and the current iteration are respectively adopted, wherein, during the first iteration, M is set by adopting different initialization strategies 1 And M 2 (e.g., M 1 Set as a unitary matrix and M 2 Set as the average diagonal matrix of feature vectors to be decoded), V c Is to be decodedThe value-related information globally semantically encodes feature vectors,and->Respectively represent feature vectors V 1 And V 2 And M is the global mean of b Is a bias matrix, for example initially set as an identity matrix, with the vectors in the form of column vectors. That is, consider that global semantic coding feature vector V based on value-related information to be decoded is in progress c In the dense prediction task of (2), the high resolution representation of the weight matrix needs to be globally semantically encoded with the feature vector V of the value-related information to be decoded c The global context of the model is integrated, so that progressive integration (progressive integrity) is realized based on iterative association representation resource-aware by maximizing the distribution boundary of the weight space in the iterative process, thereby improving the training effect of the weight matrix and improving the training efficiency of the whole model. Therefore, the early warning prompt of different grades can be intelligently carried out on the annual fee payment of the patent based on the value and the annual fee payment state of the patent so as to improve the management efficiency and the management level of the patent, optimize the maintenance cost and the income of the patent and protect the rights and interests of the patent.
As described above, the patent annual fee intelligent warning system 300 according to the embodiment of the present application may be implemented in various wireless terminals, for example, a server or the like having a patent annual fee intelligent warning algorithm. In one possible implementation, the patent annuity intelligent pre-warning system 300 in accordance with embodiments of the present application may be integrated into the wireless terminal as a software module and/or hardware module. For example, the patent annuity intelligent pre-warning system 300 may be a software module in the operating system of the wireless terminal, or may be an application developed for the wireless terminal; of course, the annual fee intelligent warning system 300 could equally be one of many hardware modules of the wireless terminal.
Alternatively, in another example, the patent annuity intelligent pre-warning system 300 and the wireless terminal may be separate devices, and the patent annuity intelligent pre-warning system 300 may be connected to the wireless terminal via a wired and/or wireless network and transmit interactive information in a agreed data format.
Further, an intelligent early warning method for the annual fee of the patent is also provided.
Fig. 4 is a flowchart of an intelligent early warning method for annual fee in patent according to an embodiment of the application. As shown in fig. 4, the patent annual fee intelligent early warning method according to the embodiment of the application includes the steps of: s1, acquiring annual fee payment states of a patent to be evaluated; s2, collecting value related information of the patent to be evaluated, wherein the value related information comprises technical feature text description, market demand text description and competition situation text description; s3, carrying out semantic coding on the technical feature text description, the market demand text description and the competition situation text description in the value related information respectively to obtain technical feature semantic coding feature vectors, market demand semantic coding feature vectors and competition situation semantic coding feature vectors; s4, carrying out semantic association analysis on the technical feature semantic coding feature vector, the market demand semantic coding feature vector and the competition situation semantic coding feature vector to obtain global semantic coding features of the value related information; s5, determining the value score of the to-be-evaluated patent based on the global semantic coding features of the value related information; s6, determining the early warning level of the to-be-evaluated patent based on the decoding value and the annual fee payment state of the to-be-evaluated patent.
In summary, the patent annual fee intelligent early warning method according to the embodiment of the application is explained, which collects the value related information of the patent, such as technical feature text description, market demand text description, competition situation text description and the like, and introduces semantic understanding technology at the rear end to perform semantic association analysis of the value related information of the patent, so as to judge the value score of the patent, and generates the annual fee early warning level, such as high, medium, low and the like, based on the value score and the annual fee payment state of the patent. Through the mode, different grades of early warning prompts can be intelligently carried out on the annual fee payment of the patent according to the value and the annual fee payment state of the patent so as to improve the management efficiency and the management level of the patent, optimize the maintenance cost and the income of the patent and protect the rights and interests of the patent.
The foregoing description of the embodiments of the present disclosure has been presented for purposes of illustration and description, and is not intended to be exhaustive or limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the various embodiments described. The terminology used herein was chosen in order to best explain the principles of the embodiments, the practical application, or the improvement of technology in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein.
Claims (8)
1. An intelligent early warning system for annual fee of patent is characterized by comprising:
the annual fee payment state detection module is used for acquiring the annual fee payment state of the patent to be evaluated;
the patent value information acquisition module is used for collecting value related information of the patent to be evaluated, wherein the value related information comprises technical feature text description, market demand text description and competition situation text description;
the patent value information semantic understanding module is used for respectively carrying out semantic coding on the technical feature text description, the market demand text description and the competition situation text description in the value related information so as to obtain technical feature semantic coding feature vectors, market demand semantic coding feature vectors and competition situation semantic coding feature vectors;
the patent value semantic feature association analysis module is used for carrying out semantic association analysis on the technical feature semantic coding feature vector, the market demand semantic coding feature vector and the competition situation semantic coding feature vector to obtain value related information global semantic coding features;
the patent value score calculation module is used for determining the value score of the patent to be evaluated based on the global semantic coding features of the value related information;
and the patent annual fee early warning module is used for determining the early warning level of the patent to be evaluated based on the decoding value and the annual fee payment state of the patent to be evaluated.
2. The patent annual fee intelligent warning system according to claim 1, wherein said patent value information semantic understanding module comprises:
the word segmentation unit is used for carrying out word segmentation processing on the technical feature text description, the market demand text description and the competition situation text description in the value related information so as to convert the technical feature text description, the market demand text description and the competition situation text description in the value related information into word sequences composed of a plurality of words;
the word embedding unit is used for mapping each word in the word sequence into a word embedding vector by using an embedding layer of the semantic encoder comprising the word embedding layer so as to obtain a sequence of word embedding vectors;
a context coding unit, configured to perform global context semantic coding on the sequence of word embedding vectors using the converter of the semantic encoder including the word embedding layer, where the global context semantic coding is based on a converter thought, so as to obtain a plurality of global context semantic feature vectors; and
and the cascading unit is used for cascading the plurality of global context semantic feature vectors to obtain the technical feature semantic coding feature vector, the market demand semantic coding feature vector and the competition situation semantic coding feature vector.
3. The patent annual fee intelligent warning system according to claim 2, wherein the patent value semantic feature association analysis module is configured to: and the technical feature semantic coding feature vector, the market demand semantic coding feature vector and the competition situation semantic coding feature vector are processed through a context encoder based on a Swin Transformer to obtain value related information global semantic coding feature vector serving as the value related information global semantic coding feature.
4. The patent annual fee intelligent warning system of claim 3, wherein the patent value score calculation module is configured to: and passing the value related information global semantic coding feature vector through a decoder to obtain a decoding value, wherein the decoding value is used for representing the value score of the patent to be evaluated.
5. The patent annual fee intelligent warning system of claim 4, further comprising a training module for training the semantic encoder including a word embedding layer, the Swin fransformer based context encoder, and the decoder.
6. The patent annual fee intelligent warning system of claim 5, wherein the training module includes:
the training data acquisition unit is used for acquiring training data, wherein the training data comprises training annual fee payment states of the to-be-evaluated patents, and training value related information of the to-be-evaluated patents, and the training value related information comprises training technical feature text description, training market demand text description and training competition situation text description, and true values of value scores of the to-be-evaluated patents;
the training value related information semantic feature extraction unit is used for respectively carrying out semantic coding on training technical feature text description, training market demand text description and training competition situation text description in the training value related information to obtain training technical feature semantic coding feature vectors, training market demand semantic coding feature vectors and training competition situation semantic coding feature vectors;
the training value related information semantic association coding unit is used for enabling the training technical feature semantic coding feature vector, the training market demand semantic coding feature vector and the training competition situation semantic coding feature vector to pass through the context encoder based on the Swin transform to obtain a training value related information global semantic coding feature vector;
the decoding loss unit is used for enabling the training value related information global semantic coding feature vector to pass through the decoder so as to obtain a decoding loss function value;
and the model training unit is used for training the semantic encoder comprising the word embedding layer, the context encoder based on the Swin Transformer and the decoder based on the decoding loss function value and through gradient descending direction propagation, wherein the feature correction of the weight space is carried out on the global semantic coding feature vector of the training value related information when the weight matrix of the training is iterated each time.
7. The patent annual fee intelligent warning system according to claim 6, wherein said decoding loss unit is configured to:
performing decoding regression on the training value related information global semantic coding feature vector by using a decoder to obtain a training decoding value; and
and calculating a mean square error value between the training decoding value and a true value of the early warning level of the patent to be evaluated as the decoding loss function value.
8. The patent annual fee intelligent early warning method is characterized by comprising the following steps of:
acquiring annual fee payment states of the to-be-evaluated patent;
collecting value related information of the patent to be evaluated, wherein the value related information comprises technical feature text description, market demand text description and competition situation text description;
carrying out semantic coding on technical feature text description, market demand text description and competition situation text description in the value related information respectively to obtain technical feature semantic coding feature vectors, market demand semantic coding feature vectors and competition situation semantic coding feature vectors;
carrying out semantic association analysis on the technical feature semantic coding feature vector, the market demand semantic coding feature vector and the competition situation semantic coding feature vector to obtain global semantic coding features of the value related information;
determining the value score of the patent to be evaluated based on the global semantic coding features of the value related information;
and determining the early warning level of the to-be-evaluated patent based on the decoding value and the annual fee payment state of the to-be-evaluated patent.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311435774.2A CN117350898A (en) | 2023-10-30 | 2023-10-30 | Intelligent early warning system and method for annual patent fee |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311435774.2A CN117350898A (en) | 2023-10-30 | 2023-10-30 | Intelligent early warning system and method for annual patent fee |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117350898A true CN117350898A (en) | 2024-01-05 |
Family
ID=89369180
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311435774.2A Withdrawn CN117350898A (en) | 2023-10-30 | 2023-10-30 | Intelligent early warning system and method for annual patent fee |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117350898A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117973795A (en) * | 2024-02-26 | 2024-05-03 | 浙江萧东机电工程有限公司 | Electromechanical engineering management system |
| CN118094413A (en) * | 2024-04-01 | 2024-05-28 | 烽台科技（北京）有限公司 | Training method of patent value evaluation model and evaluation method of patent data |
-
2023
- 2023-10-30 CN CN202311435774.2A patent/CN117350898A/en not_active Withdrawn
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117973795A (en) * | 2024-02-26 | 2024-05-03 | 浙江萧东机电工程有限公司 | Electromechanical engineering management system |
| CN118094413A (en) * | 2024-04-01 | 2024-05-28 | 烽台科技（北京）有限公司 | Training method of patent value evaluation model and evaluation method of patent data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110135457B (en) | Event trigger word extraction method and system based on self-encoder fusion document information | |
| CN113434357B (en) | Log anomaly detection method and device based on sequence prediction | |
| CN117350898A (en) | Intelligent early warning system and method for annual patent fee | |
| CN111897908A (en) | Event extraction method and system fusing dependency information and pre-training language model | |
| CN112800782B (en) | Voice translation method, system and equipment integrating text semantic features | |
| CN111382565A (en) | Multi-label-based emotion-reason pair extraction method and system | |
| CN112200664A (en) | Repayment prediction method based on ERNIE model and DCNN model | |
| CN113742733B (en) | Method and device for extracting trigger words of reading and understanding vulnerability event and identifying vulnerability type | |
| CN113657115B (en) | Multi-mode Mongolian emotion analysis method based on ironic recognition and fine granularity feature fusion | |
| CN112767922B (en) | Speech recognition method for contrast predictive coding self-supervision structure joint training | |
| CN108763230B (en) | Neural machine translation method using external information | |
| CN114168754A (en) | Relation extraction method based on syntactic dependency and fusion information | |
| CN116796251A (en) | Poor website classification method, system and equipment based on image-text multi-mode | |
| CN116939320A (en) | Method for generating multimode mutually-friendly enhanced video semantic communication | |
| CN116305119A (en) | APT malicious software classification method and device based on predictive guidance prototype | |
| CN113761875B (en) | Event extraction method and device, electronic equipment and storage medium | |
| CN111597816A (en) | Self-attention named entity recognition method, device, equipment and storage medium | |
| CN113780418A (en) | Data screening method, system, equipment and storage medium | |
| CN114610871B (en) | Information system modeling analysis method based on artificial intelligence algorithm | |
| CN110910905B (en) | Mute point detection method and device, storage medium and electronic equipment | |
| CN116484851A (en) | Pre-training model training method and device based on variant character detection | |
| CN116521863A (en) | Tag anti-noise text classification method based on semi-supervised learning | |
| CN115169285A (en) | Event extraction method and system based on graph analysis | |
| CN115658956B (en) | Hot topic mining method and system based on conference audio data | |
| CN116681428B (en) | Intelligent recycling management system and method for electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20240105 |