Method, device and electronic equipment for determining trademark similarity

Abstract

The application discloses a method for determining trademark similarity, comprising: obtaining first character information corresponding to a first trademark and second character information corresponding to a second trademark; obtaining first feature information of all characters in the first trademark according to the first character information, and obtaining second feature information of all characters in the second trademark according to the second character information; and determining whether the first trademark and the second trademark are similar trademarks according to the first feature information and the second feature information. The method can judge the similarity of the first trademark and the second trademark from multiple dimensions, quickly and accurately obtain the similarity between the two trademarks, and avoid the problems of inaccurate manual design rules or manual design rule calculation, thereby saving manpower and improving user experience.

Description

Technical Field

[0001] This disclosure relates to the field of natural language processing technology, and more specifically, to a method, apparatus, electronic device, and computer-readable storage medium for determining trademark similarity. This application also relates to a method for predicting trademark application conflicts. Background Technology

[0002] A trademark is a mark used to distinguish the brand or services of one business from the goods or services of other businesses. In increasingly fierce market competition, the value of trademarks is becoming more and more important.

[0003] Currently, for trademarks pending application, enterprises or individuals can generally use electronic devices to search the registered trademark database before making a formal application, or when trademark examiners are examining the trademarks pending application, to confirm whether there are any trademarks similar to or conflicting with the trademarks pending application, thereby generating a preliminary judgment or review result for the trademarks pending application.

[0004] However, existing electronic devices generally use simple rules to search for trademarks similar to the trademark to be applied for, and the search results are not accurate, often requiring time-consuming manual verification. Summary of the Invention

[0005] One objective of this disclosure is to provide a new technical solution for quickly and accurately determining the similarity between trademarks.

[0006] A first aspect of this disclosure provides a method for determining trademark similarity, the method comprising:

[0007] Obtain the first character information corresponding to the first trademark and the second character information corresponding to the second trademark;

[0008] Based on the first character information, first feature information of all characters in the first trademark is obtained, and based on the second character information, second feature information of all characters in the second trademark is obtained;

[0009] Based on the first feature information and the second feature information, determine whether the first trademark and the second trademark are similar trademarks.

[0010] Optionally, determining whether the first trademark and the second trademark are similar trademarks based on the first feature information and the second feature information includes:

[0011] Based on the first feature information of the i-th character in the first trademark and the second feature information of the j-th character in the second trademark, the ij-th similarity value in the similarity matrix is ​​obtained; i is a positive integer, and is less than or equal to the total number of characters in the first trademark; j is a positive integer, and is less than or equal to the total number of characters in the second trademark;

[0012] Based on the similarity matrix, a target similarity vector is obtained, wherein the target similarity vector is composed of the maximum value of each row in the similarity matrix;

[0013] Based on the target similarity vector, the first similarity between the first trademark and the second trademark is determined.

[0014] Optionally, the first character information includes the Chinese character information, pinyin information, and Wubi input method root information of the corresponding character in the first trademark;

[0015] The step of obtaining the first feature information of all characters in the first trademark based on the first character information includes:

[0016] Based on the first character information, obtain the first Chinese character information, the first Pinyin information, and the first Wubi input root information corresponding to all characters in the first trademark;

[0017] Based on the first Chinese character information and the first Pinyin information, obtain the first Chinese character embedding vector and the first Pinyin embedding vector; and,

[0018] Based on the first Wubi character root information, a first character shape vector is obtained, wherein the first character shape vector represents the character shape;

[0019] The first Chinese character embedding vector, the first Pinyin embedding vector, and the first character shape vector are subjected to a first unified connection process to obtain the first feature information.

[0020] Optionally, obtaining the first character shape vector based on the first Wubi input method root information includes:

[0021] Based on the first Wubi character root information, obtain the first Wubi character root embedding vector that corresponds one-to-one with the Wubi character root of the character;

[0022] The first Wubi character root embedding vector is subjected to a second unified connection process to obtain the first character shape vector.

[0023] Optionally, the method further includes the step of preprocessing known trademarks to establish a trademark database, specifically including:

[0024] For a known trademark, obtain the pinyin and / or Wubi code corresponding to the Chinese characters of the known trademark;

[0025] A trademark database is established based on the Chinese characters, pinyin, and Wubi codes of the known trademarks, as well as the classification information and registrant information of the known trademarks.

[0026] Optionally, the second trademark is obtained by means of:

[0027] Based on the first trademark, obtain the top n trademarks with the second similarity in the trademark database, where n is a positive integer;

[0028] Choose one trademark from the first n trademarks as the second trademark.

[0029] A second aspect of this disclosure also provides a method for predicting trademark application conflicts, including:

[0030] Obtain the target trademark to be applied for;

[0031] Obtain a set of similar trademarks corresponding to the target trademark;

[0032] A similarity set is obtained based on the target trademark and the set of similar trademarks, wherein a first similarity in the similarity set represents the degree of similarity between the target trademark and the trademarks in the set of similar trademarks, and the first similarity is obtained according to the method described in the first aspect of this disclosure;

[0033] Based on the similarity set, the preliminary application result of the target trademark is obtained.

[0034] Optionally, the method further includes:

[0035] A list of similar trademarks is generated based on the set of similar trademarks and the set of similarities, wherein the list of similar trademarks includes multiple data pairs, each data pair consisting of a trademark in the set of similar trademarks and a first similarity score corresponding to the trademark in the set of similarities.

[0036] Optionally, the method is applied to a server, and the method further includes:

[0037] The list of similar trademarks and the predicted application results are provided to the terminal device.

[0038] Optionally, the method is applied to a terminal device, and the method further includes:

[0039] The list of similar trademarks and the predicted application results are displayed.

[0040] According to a third aspect of this disclosure, an apparatus for determining trademark similarity is also provided, comprising:

[0041] The character information acquisition module is used to acquire the first character information corresponding to the first trademark and the second character information corresponding to the second trademark.

[0042] The feature information acquisition module is used to obtain first feature information of all characters in the first trademark based on the first character information, and to obtain second feature information of all characters in the second trademark based on the second character information;

[0043] The similar trademark determination module is used to determine whether the first trademark and the second trademark are similar trademarks based on the first feature information and the second feature information.

[0044] According to a fourth aspect of this disclosure, an electronic device is also provided, comprising the means described in a third aspect of this disclosure; or, comprising:

[0045] Memory is used to store executable instructions;

[0046] A processor, configured to, under the control of the executable instructions, operate the electronic device to perform the method according to the first or second aspect of this disclosure.

[0047] According to a fifth aspect of this disclosure, a computer-readable storage medium is also provided, the computer-readable storage medium storing a computer program that can be read and executed by a computer, the computer program being used, when read and executed by the computer, to perform the method described according to the first or second aspect of this disclosure.

[0048] According to embodiments of this disclosure, when an electronic device needs to determine whether two trademarks are similar, it can first obtain the first character information corresponding to the first trademark and the second character information corresponding to the second trademark. Then, by obtaining the first feature information of all characters in the first trademark and the second feature information of all characters in the second trademark, the device can determine whether the first and second trademarks are similar based on the first and second feature information. In this method, since the first and second feature information can respectively characterize the meaning, pronunciation, and glyph information of the characters in the first and second trademarks, this method allows the electronic device to evaluate the similarity between the first and second trademarks from multiple dimensions, quickly and accurately obtaining the similarity between the two trademarks, thereby confirming whether the two trademarks are similar. Furthermore, the method provided in this embodiment can avoid the problems of manually designing rules or inaccurate calculations of manually designed rules, thereby saving manpower and improving user experience.

[0049] Other features and advantages of this disclosure will become clear from the following detailed description of exemplary embodiments with reference to the accompanying drawings. Attached Figure Description

[0050] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments of the present disclosure and, together with their description, serve to explain the principles of the present disclosure.

[0051] Figure 1 This is a schematic diagram of a scenario illustrating the method for determining trademark similarity provided in this embodiment.

[0052] Figure 2 This is a hardware configuration diagram of an electronic device that can be used to implement the method for determining trademark similarity in the embodiments of this disclosure.

[0053] Figure 3 This is a flowchart illustrating the method for determining trademark similarity provided in this embodiment.

[0054] Figure 4 This is a schematic diagram of the feature information acquisition process provided in the embodiments of this disclosure.

[0055] Figure 5 This is a schematic diagram illustrating the process of obtaining the similarity matrix provided in the embodiments of this disclosure.

[0056] Figure 6 This is a flowchart illustrating the trademark application conflict prediction method provided in this embodiment.

[0057] Figure 7 This is a schematic block diagram of the apparatus for determining trademark similarity provided in the embodiments of this disclosure.

[0058] Figure 8a This is a schematic block diagram of an electronic device according to an embodiment of the present disclosure.

[0059] Figure 8b This is a schematic block diagram of an electronic device according to another embodiment of the present disclosure. Detailed Implementation

[0060] Various exemplary embodiments of the present disclosure will now be described in detail with reference to the accompanying drawings. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments do not limit the scope of the present disclosure.

[0061] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit this disclosure or its application or use.

[0062] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.

[0063] In all the examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0064] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be discussed further in subsequent figures.

[0065] To address the low accuracy and time-consuming manual verification issues inherent in existing trademark examination technologies when electronic devices retrieve similar trademarks, a machine learning approach was discovered to quickly and accurately determine the similarity between two trademarks. For example, feature information of the two trademarks can be manually extracted and input into a pre-trained neural network model to obtain their similarity. However, this method requires manually designed feature information for the model, which has limitations. Furthermore, when dealing with similar-looking characters, dictionaries and manually designed rules are needed for auxiliary judgment. Therefore, this method is insufficient in terms of accuracy and convenience.

[0066] In this embodiment, to conveniently and accurately obtain the similarity between two trademarks, thereby improving the accuracy of trademark search results, this disclosure provides a method for determining trademark similarity. Please refer to [link / reference]. Figure 1 This is a schematic diagram illustrating an application scenario of the method for determining trademark similarity provided in this embodiment.

[0067] like Figure 1As shown, for a target trademark to be applied for, such as the trademark "Coolbar Shopping Website", when a user searches for registered trademarks similar to the target trademark through electronic device 1100, the user can send the target trademark to electronic device 1100 through terminal device 1200, for example, to a server. After obtaining the target trademark, electronic device 1100 can first search for the top n trademarks with the second similarity to the target trademark according to simple rules, such as matching Chinese characters, matching pinyin, and / or matching Wubi input method radicals under the trademark classification corresponding to the target trademark in the trademark database, as a set of similar trademarks. Then, in order to improve accuracy and reduce the time-consuming problem caused by manual secondary confirmation, electronic device 1100 can regard the target trademark as the first trademark and one of the n trademarks as the second trademark. By obtaining the first character information corresponding to the first trademark and the second character information corresponding to the second trademark respectively, and based on the first character information and the second character information, the first characteristic of all characters in the first trademark is obtained. The electronic device 1100 obtains the first feature information and the second feature information of all characters in the first and second trademarks, and then determines whether the first and second trademarks are similar trademarks based on the first feature information and the second feature information. Through the first feature information and the second feature information, the electronic device 1100 can obtain the first similarity between the target trademark and each of the n similar trademarks, and then obtain a similarity set. Then, based on the similarity set, the electronic device 1100 can quickly and accurately obtain the preliminary application result of the target trademark. For example, it can output a preliminary application result that directly represents "success" or "failure" based on whether the statistical value such as the maximum value or average value in the similarity set is greater than a preset similarity threshold. In addition, the electronic device 1100 can also generate a list of similar trademarks based on the similar trademark set and the similarity set. In order to facilitate user viewing, the electronic device 1100 can provide the list of similar trademarks and the preliminary application result to the terminal device 1200 for the terminal device 1200 to display the list of similar trademarks and the preliminary application result for user viewing.

[0068] <Hardware Configuration>

[0069] Figure 2 This is a hardware configuration diagram of an alternative electronic device that can be used to implement the method for determining trademark similarity in embodiments of this disclosure.

[0070] like Figure 2 As shown, the method for determining trademark similarity provided in this embodiment can be applied to electronic device 1100. In specific implementation, electronic device 1100 can be a server or a terminal device, without any special limitation here.

[0071] When the electronic device 1100 is a server, the server can be a blade server, a rack server, or a server cluster deployed in the cloud; there are no restrictions on this.

[0072] like Figure 2 As shown, the electronic device 1100 may include a processor 1110, a memory 1120, an interface device 1130, a communication device 1140, a display device 1150, and an input device 1160. The processor 1110 may be, for example, a central processing unit (CPU). The memory 1120 may include, for example, ROM (Read-Only Memory), RAM (Random Access Memory), or non-volatile memory such as a hard disk. The interface device 1130 may include, for example, a USB interface or a serial interface. The communication device 1140 may be capable of wired or wireless communication. The display device 1150 may be, for example, a liquid crystal display (LCD). The input device 1160 may include, for example, a touchscreen or a keyboard.

[0073] In this embodiment, the electronic device 1100 can be used to participate in the method for determining trademark similarity according to any embodiment of this disclosure.

[0074] In embodiments of this disclosure, the memory 1120 of the electronic device 1100 stores instructions for controlling the processor 1110 to operate in order to support the implementation of a method for determining trademark similarity according to any embodiment of this disclosure. Those skilled in the art can design instructions based on the schemes disclosed herein. How the instructions control the processor to operate is well known in the art and will not be described in detail here.

[0075] Those skilled in the art should understand that, although in Figure 2 The present disclosure illustrates multiple devices of electronic device 1100; however, electronic device 1100 of embodiments of the present disclosure may refer to only some of these devices, for example, only processor 1110 and memory 1120.

[0076] In addition, when the electronic device 1100 is a terminal device, the terminal device can be a smartphone, a portable computer, a desktop computer, a tablet computer, etc., without any special restrictions.

[0077] It should be noted that the method provided in this embodiment can be applied to the electronic device 1100 alone, that is, to a server or terminal device, or it can be applied to the interaction scenario between the terminal device and the server as needed, without any special limitation here.

[0078] <Method Example 1>

[0079] Figure 3 This is a flowchart illustrating a method for determining trademark similarity according to an embodiment of this disclosure. The method provided in this embodiment can be applied to electronic devices; for example, it can be applied to… Figure 2 In the electronic device 1100 shown.

[0080] As Figure 3 shown, the method for determining the trademark similarity in this embodiment may include the following steps S3100 - S3300, which will be described in detail below.

[0081] Step S3100, obtain the first character information corresponding to the first trademark and the second character information corresponding to the second trademark.

[0082] A trademark can include words, graphics, letters, numbers, three - dimensional marks, sounds, color combinations, or combinations of the above elements; in this embodiment, unless otherwise specified, the trademark is a character - type trademark, that is, a trademark composed of characters such as Chinese characters, numbers, letters, etc. is taken as an example for illustration; of course, in specific implementation, the method described in this embodiment can also be used to process other types of trademarks. For example, for trademarks containing graphics or sounds, the characters in the graphics or the words in the sounds can be recognized to convert such trademarks into character - type trademarks, or the similarity between such trademarks can also be obtained by combining methods such as image similarity acquisition methods and audio similarity acquisition methods, which is not specifically limited here.

[0083] The first trademark and the second trademark can be any two trademarks for which trademark similarity calculation is to be performed. For example, the first trademark can be "Cool Bar Shopping Website" and the second trademark can be "Ku Ba", which is not specifically restricted here.

[0084] In this embodiment, the first character information may include Chinese character information, pinyin information, and Wubi root information of the corresponding characters in the first trademark; correspondingly, the second character information may include Chinese character information, pinyin information, and Wubi root information of the corresponding characters in the second trademark.

[0085] For example, for the trademark "Ku Ba", its corresponding character information may include {("裤", "ku", "衤", "冫", "广", "车"), ("吧", "ba", "口", "巴")}.

[0086] Specifically, in this embodiment, in order to accurately determine whether two trademarks are similar trademarks, in addition to constructing feature information using the Chinese characters and pinyin information of the characters in the trademark, in order to conveniently process similar - shaped characters, the Wubi root information of the characters in the trademark is also used to construct feature information representing the character shape, so as to construct feature information for determining the trademark similarity from multiple aspects such as the meaning, pinyin, and shape of the characters.

[0087] In specific implementation, the pinyin information of the characters in the trademark can be obtained by converting the Chinese characters in the trademark into pinyin, while the English characters remain unchanged; the Wubi root information of the characters in the trademark can be obtained by disassembling the characters, and the detailed processing is not elaborated here.

[0088] Step S3200: Obtain the first feature information of all characters in the first trademark according to the first character information, and obtain the second feature information of all characters in the second trademark according to the second character information.

[0089] After obtaining the two trademarks to be determined whether they are similar trademarks through step S3100, and obtaining the first character information and the second character information according to the two trademarks, the electronic device can construct the feature information for similarity calculation from multiple aspects according to the first character information and the second character information, which is described in detail below.

[0090] In specific implementation, the obtaining of the first feature information of all characters in the first trademark according to the first character information includes the following steps S3201 - S3204.

[0091] Step S3201: Obtain the first Chinese character information, the first pinyin information and the first Wubi root information of all characters in the first trademark according to the first character information.

[0092] The first Chinese character information, for example, for the trademark "Cool Bar Shopping Website", the Chinese character information of the character "酷" is "酷".

[0093] The first pinyin information is the information of the pinyin corresponding to the character. For example, the pinyin information of the character "酷" can be "ku".

[0094] The first Wubi root information is the information of all Wubi roots corresponding to the character. For example, the Wubi root information of the character "酷" is respectively "西", "一", "丿" and "口".

[0095] It should be noted that when the character in the trademark is not a Chinese character, the corresponding Chinese character information and Wubi root information can be empty, or can also be filled with other information, and no special limitation is made here.

[0096] Step S3202: Obtain the first Chinese character embedding vector and the first pinyin embedding vector according to the first Chinese character information and the first pinyin information.

[0097] The first Chinese character embedding vector is a vector obtained by word embedding processing of the Chinese character represented by the first Chinese character information; the first Pinyin embedding vector is a vector obtained by word embedding processing of the Pinyin represented by the first Pinyin information; correspondingly, word embedding processing can also be performed on at least one Wubi character root represented in the first Wubi character root information to obtain the first Wubi character root embedding vector that corresponds one-to-one with the Wubi character root in the first Wubi character root information; wherein, word embedding processing is a process of converting words into numerical vectors for analysis using machine learning techniques.

[0098] In specific implementation, when performing word embedding processing on Chinese characters, pinyin, and Wubi radicals, registered trademarks can be obtained as training data. By indexing and numbering the Chinese characters, pinyin, and Wubi radicals corresponding to the characters in the training data, a Chinese character dictionary, a pinyin dictionary, and a Wubi radical dictionary can be constructed for word embedding processing. In this embodiment, the indexing number can be a unique numerical identifier assigned to the Chinese characters, pinyin, and Wubi radicals corresponding to the characters in the training data.

[0099] In this embodiment, when performing word embedding processing on the characters in the first trademark, the Chinese character dictionary, Pinyin dictionary, and Wubi character root dictionary constructed above can be used to perform embedding processing to obtain their corresponding embedding vectors.

[0100] Step S3203: Obtain a first character shape vector based on the first Wubi character root information, wherein the first character shape vector represents the character shape.

[0101] In a specific implementation, obtaining the first character shape vector based on the first Wubi character root information includes: obtaining a first Wubi character root embedding vector that corresponds one-to-one with the Wubi character root of the character based on the first Wubi character root information; and performing a second unified connection processing on the first Wubi character root embedding vector to obtain the first character shape vector.

[0102] The process of obtaining the embedding vector of the first Wubi character root will not be elaborated here.

[0103] In this embodiment, the second unified connection processing can specifically be: inputting all Wubi character root embedding vectors of a character into the gated recurrent unit (GRU) of the second neural network model, so as to obtain the glyph vector representing the character shape based on the context information between the character roots. The second neural network model is used to perform unified connection processing on the multiple Wubi character root embedding vectors based on the context information between the multiple input Wubi character root embedding vectors, so as to obtain the glyph vector representing the character shape by the multiple Wubi character root embedding vectors.

[0104] Step S3204: Perform a first unified connection process on the first Chinese character embedding vector, the first pinyin embedding vector, and the first character shape vector to obtain the first feature information.

[0105] In this embodiment, the first unified connection processing can specifically be: inputting the first Chinese character embedding vector, the first Pinyin embedding vector, and the first glyph vector of the character into the gated recurrent unit (GRU) of the first neural network model to obtain unified feature information that represents the character from three aspects: meaning, Pinyin, and glyph.

[0106] Please refer to Figure 4 This is a schematic diagram illustrating the process of obtaining feature information provided in an embodiment of this disclosure. Figure 4 In the design, the trademark consists of five characters: w0, w1, w2, w3, and w4. Each character contains five Wubi input method radicals. For example, for the character w0, its Wubi input method radical embedding vector can be represented by w... 0，0 ,w 0，1 ,w 0，2 ,w 0，3 ,w 0，4 This represents the word embedding vector for each Chinese character, denoted as `we`. For example, the word embedding vectors for the five Chinese characters of each trademark can be represented as `we0`, `we1`, `we2`, `we3`, and `we4`, respectively. The pinyin embedding vector for each Chinese character is represented as `pe`. For example, the pinyin embedding vectors for the five Chinese characters of each trademark can be represented as `pe0`, `pe1`, `pe2`, `pe3`, and `pe4`, respectively. Then, as shown... Figure 4 As shown, by inputting the Wubi character root embedding vector of each character into the second neural network model, a character shape vector representing the shape of each character can be obtained, i.e., h0 wb h1 wb h2 wb h3 wb h4 wb Then, the Chinese character embedding vector, the pinyin embedding vector, and the character shape vector of each character are input into the first neural network model, and the feature information corresponding to each character can be obtained, namely, h0, h1, h2, h3, h4.

[0107] It should be noted that the first and second neural network models can be pre-trained. In specific implementation, the first and second neural network models can be models based on gated recurrent unit network structures, models based on long short-term memory (LSTM) network structures, or other structures, such as bidirectional recurrent neural network (BiRNN) structures. No special limitation is made here.

[0108] The above provides a detailed explanation of how to obtain the first feature information of all characters in the first trademark based on the first character information. In specific implementation, the second feature information of all characters in the second trademark can also be obtained based on the second character information. The processing procedure is similar to that for obtaining the first feature information, and will not be elaborated here.

[0109] After step S3200, step S3300 is executed to determine whether the first trademark and the second trademark are similar trademarks based on the first feature information and the second feature information.

[0110] In specific implementation, determining whether the first trademark and the second trademark are similar trademarks based on the first feature information and the second feature information specifically includes the following steps S3301-3303, which will be described in detail below.

[0111] Step S3301: Based on the first feature information of the i-th character in the first trademark and the second feature information of the j-th character in the second trademark, obtain the ij-th similarity value in the similarity matrix; i is a positive integer and is less than or equal to the total number of characters in the first trademark; j is a positive integer and is less than or equal to the total number of characters in the second trademark.

[0112] In this embodiment, the similarity value between two feature information can be obtained by calculating the cosine similarity. Cosine similarity is also known as cosine similarity, which evaluates the similarity between two vectors by calculating the cosine value of the angle between them. In specific implementation, other algorithms can also be used to calculate the similarity between two feature information, and no special limitation is made here.

[0113] In practical implementation, if the first characteristic information of the first trademark is set as h0, h1, h2, h3, h4, and the second characteristic information of the second trademark is set as h'0, h'1, h'2, h'3, h'4, then it can be done as follows: Figure 5 The similarity matrix M described in this embodiment is obtained as shown. In this similarity matrix, each element M... i,jThe cosine similarity between the i-th character in the first trademark and the j-th character in the second trademark is represented.

[0114] Step S3302: Obtain the target similarity vector based on the similarity matrix, wherein the target similarity vector is composed of the maximum value of each row in the similarity matrix.

[0115] After obtaining the similarity matrix through the above processing, by obtaining the maximum value of each row in the similarity matrix, for example, taking the maximum value of the i-th row, we can obtain the value with the highest similarity between the i-th character in the first trademark and all characters in the second trademark. Then, the similarity matrix can be simplified into an n-dimensional vector, where the value of n can be determined according to the characters contained in the first trademark. For example, when the first trademark contains five characters, the target similarity vector can be a five-dimensional vector.

[0116] Step S3303: Determine the second similarity between the first trademark and the second trademark based on the target similarity vector.

[0117] Specifically, after obtaining the target similarity vector, the vector can be input into the fully connected layer and logistic regression layer (softmax layer) of the neural network model used for similarity calculation. The second similarity between the two trademarks can be obtained by the numerical value representing the degree of similarity between the two trademarks output by the logistic regression layer, and the two trademarks can be determined as similar based on the second similarity.

[0118] It should be noted that, in this embodiment, for the purpose of distinction, the numerical value representing the degree of similarity between the two trademarks, calculated using the first feature information of all characters in the first trademark and the second feature information of all characters in the second trademark, is called the second similarity; while the first similarity can be a rough numerical value representing the degree of similarity between trademarks obtained by the user using an electronic device after obtaining the first trademark, for example, the trademark to be applied for by the user, through a simple matching based on rules such as Chinese character matching, Pinyin matching, and / or Wubi character root matching from a trademark database containing known trademarks.

[0119] In one embodiment, a trademark database containing information about known trademarks can be established by preprocessing the known trademarks. The establishment steps may include: for a known trademark, obtaining the pinyin and / or Wubi code corresponding to the Chinese characters of the known trademark; and establishing a trademark database based on the Chinese characters, pinyin, Wubi code, classification information, and registrant information of the known trademark.

[0120] After establishing this trademark database, when it is necessary to quickly and accurately determine whether two trademarks are similar, a matching process can be performed on the first trademark to be determined in the trademark database to obtain the top n trademarks with the second similarity through simple matching rules, where n is a positive integer; then, a trademark can be selected from these n trademarks as the second trademark, and a second matching process can be performed to accurately determine whether the two trademarks are similar trademarks by using the first feature information of all characters in the first trademark and the second feature information of all characters in the second trademark.

[0121] In one embodiment, the logistic regression layer can simultaneously output a first value representing the degree of similarity between trademarks and a second value representing the degree of dissimilarity between trademarks. The model can also output a similarity determination result while comparing the two values, in addition to outputting the second similarity.

[0122] For example, if the first value representing the degree of similarity is 0.8 and the second value representing the degree of dissimilarity is 0.2, then the second similarity between the two trademarks is 0.8, and the similarity judgment result is a value representing "similar"; while when the first value is 0.3 and the second value is 0.7, then the second similarity can be output as 0.3, and the similarity judgment result is a value representing "dissimilar".

[0123] In summary, the method for determining trademark similarity provided in this embodiment can quickly and accurately determine whether two trademarks are similar because the first feature information and the second feature information used for similarity calculation can characterize the meaning, pinyin, and glyph information of the first and second trademarks.

[0124] It should be noted that, in specific implementation, the method described in this embodiment can be used in trademark application prediction scenarios. Specifically, it can be used when an agent in a trademark agency receives a client's commission and needs to determine whether the trademark to be applied for by the client conflicts with a known trademark, i.e., a registered trademark. Alternatively, it can be used when the Trademark Office examines whether a trademark submitted by a user for examination conflicts with a known trademark. The trademark to be applied for or the trademark to be examined can be regarded as the target trademark, and the method provided in this embodiment can be used to search for known trademarks similar to the target trademark in order to quickly and accurately obtain the prediction result.

[0125] For example, regarding the target trademark "Coolbar Shopping Website", users, such as agents or examiners, can send the target trademark to a server used to obtain preliminary application results via a terminal device. This could be done by searching for the target trademark in a trademark search engine and sending it to the corresponding server. This search engine server can be connected to a trademark database established after preprocessing known trademarks. After obtaining the target trademark, the server can first perform a coarse recall process, or a first-stage matching process. This involves using simple rules, such as Chinese character matching, pinyin matching, and Wubi input method root matching, to search for second-stage matches with the target trademark based on pre-built information about known trademarks. A similar trademark set is constructed by ranking the top n trademarks by similarity. This similar trademark set can be, for example, {“Cube”, “Kooba”, “Cool”}. Then, for each trademark in the similar trademark set and the target trademark, the first similarity between the target trademark and each similar trademark can be determined using the method provided in this embodiment to obtain a similarity set, for example, {3%, 13%, 20%}. Then, based on the similarity set, the server can return a preliminary application result to the terminal device. For example, if a preset similarity threshold is set to 80%, then if the maximum value in the similarity set is not greater than the similarity threshold, the server can determine that there are no similar trademarks among the registered trademarks, and can return a “successful” preliminary application result to the terminal device.

[0126] Of course, to facilitate user viewing and secondary confirmation, after obtaining the set of similar trademarks and the set of similarity, the server can also generate a list of similar trademarks based on the set of similar trademarks and the set of similarity. For example, it can generate a list of similar trademarks consisting of data pairs {(“Cube”, 3%), (“Kooba”, 13%), (“Cooler”, 20%)} and provide the list of similar trademarks and the above-mentioned preliminary application results to the terminal device so that the terminal device can display the list of similar trademarks and the preliminary application results for the user to view.

[0127] It should be noted that the above is only one embodiment of the method provided in this application. In specific implementation, the method can also be applied to other scenarios as needed, which will not be elaborated here.

[0128] In summary, the method for determining trademark similarity provided in this embodiment uses the high-dimensional representation of the trademark's meaning, pinyin, and glyphs—that is, the embedded vector—as input. This allows for simultaneous consideration of the similarity between the three aspects of meaning, pinyin, and glyphs during similarity calculation, thereby improving the similarity determination results. Furthermore, this method avoids the problems of manually designed rules or inaccurate calculations by manually designed rules, thus saving manpower and improving user experience.

[0129] <Method Example 2>

[0130] Corresponding to the first embodiment of the method described above, this embodiment also provides a method for predicting trademark application conflicts. Please refer to [link / reference]. Figure 6 This is a flowchart illustrating the trademark application conflict prediction method provided in this disclosure. This method can be applied to electronic devices, for example, it can be applied to... Figure 2 The electronic device 1100 shown here is not specifically limited.

[0131] like Figure 6 As shown, the method provided in this embodiment may include steps S6100-S6400.

[0132] Step S6100: Obtain the target trademark to be applied for.

[0133] Step S6200: Obtain a set of similar trademarks corresponding to the target trademark.

[0134] Step S6300: Based on the target trademark and the set of similar trademarks, obtain a similarity set, wherein the first similarity in the similarity set represents the degree of similarity between the target trademark and the trademarks in the set of similar trademarks, and the first similarity is obtained according to the method described in Method Embodiment 1.

[0135] Step S6400: Based on the similarity set, obtain the preliminary application result of the target trademark.

[0136] In one embodiment, the method further includes: generating a list of similar trademarks based on the set of similar trademarks and the set of similarities, wherein the list of similar trademarks includes multiple data pairs, each data pair consisting of a trademark in the set of similar trademarks and a first similarity score corresponding to the trademark in the set of similarities.

[0137] In one embodiment, the method can be applied to a server, in which case the method further includes providing the list of similar trademarks and the predicted application results to a terminal device.

[0138] In one embodiment, the method can be applied to a terminal device, in which case the method further includes: displaying the list of similar trademarks and the predicted application results.

[0139] The method provided in this embodiment, for the target trademark to be applied for, can first use simple rules, such as Chinese character matching, Pinyin matching, and Wubi input method matching, to perform a matching to obtain a set of similar trademarks corresponding to the target trademark. The set of similar trademarks can be a set of the top n trademarks with the second similarity ranking calculated according to the above simple rules. Then, in order to avoid the time-consuming problem caused by the user's second confirmation, after obtaining the set of similar trademarks through the first matching, this embodiment uses the method for determining trademark similarity described in the first embodiment to obtain the similarity between the target trademark and each trademark in the set of similar trademarks, thereby obtaining a similarity set. Based on this similarity set, the preliminary application result for the target trademark can be accurately obtained.

[0140] <Device Embodiment>

[0141] Corresponding to the above embodiments, this embodiment also provides a device for determining trademark similarity, such as... Figure 7 The diagram shown is a schematic block diagram of an apparatus for determining trademark similarity provided in an embodiment of this disclosure.

[0142] according to Figure 7 As shown, the apparatus 7000 for determining trademark similarity in this embodiment includes a character information acquisition module 7100, a feature information construction module 7200, and a similar trademark determination module 7300.

[0143] The character information acquisition module 7100 is used to acquire the first character information corresponding to the first trademark and the second character information corresponding to the second trademark.

[0144] The feature information construction module 7200 is used to obtain first feature information of all characters in the first trademark based on the first character information, and to obtain second feature information of all characters in the second trademark based on the second character information.

[0145] In one embodiment, the first character information includes the Chinese character information, pinyin information, and Wubi input method radical information of the corresponding character in the first trademark. When the feature information obtaining module 7200 obtains the first feature information of all characters in the first trademark based on the first character information, it can be used to: obtain the first Chinese character information, the first pinyin information, and the first Wubi input method radical information corresponding to all characters in the first trademark based on the first character information; obtain the first Chinese character embedding vector and the first pinyin embedding vector based on the first Chinese character information and the first pinyin information; and obtain the first glyph vector based on the first Wubi input method radical information, wherein the first glyph vector represents the glyph of the character; and perform a first unified connection processing on the first Chinese character embedding vector, the first pinyin embedding vector, and the first glyph vector to obtain the first feature information.

[0146] In this embodiment, when the feature information acquisition module 7200 obtains the first character shape vector based on the first Wubi character root information, it can be used to: obtain the first Wubi character root embedding vector corresponding one-to-one with the Wubi character root of the character based on the first Wubi character root information; and perform a second unified connection processing on the first Wubi character root embedding vector to obtain the first character shape vector.

[0147] The similar trademark determination module 7300 is used to determine whether the first trademark and the second trademark are similar based on the first feature information and the second feature information.

[0148] In one embodiment, when determining whether the first trademark and the second trademark are similar trademarks based on the first feature information and the second feature information, the similarity acquisition module 7300 can be used to: obtain the ij-th similarity value in the similarity matrix based on the first feature information of the i-th character in the first trademark and the second feature information of the j-th character in the second trademark; i is a positive integer and is less than or equal to the total number of characters in the first trademark; j is a positive integer and is less than or equal to the total number of characters in the second trademark; obtain a target similarity vector based on the similarity matrix, wherein the target similarity vector is composed of the maximum value of each row in the similarity matrix; and determine the second similarity between the first trademark and the second trademark based on the target similarity vector.

[0149] <Equipment Example>

[0150] Corresponding to the above embodiments, this embodiment provides an electronic device, such as... Figure 8a As shown, the electronic device 100 includes a device 7000 for determining trademark similarity according to any embodiment of the present disclosure.

[0151] In another embodiment, such as Figure 8b As shown, the electronic device 100 may include a memory 110 and a processor 120. The memory 110 is used to store executable instructions, and the processor 120 is used to execute methods as described in any embodiment of the present disclosure, under the control of the executable instructions.

[0152] <Media Example>

[0153] Corresponding to the above method embodiments, this embodiment also provides a computer-readable storage medium storing a computer program that can be read and executed by a computer. The computer program is used to execute the method described in any of the above embodiments of this disclosure when read and executed by the computer.

[0154] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.

[0155] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination thereof. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0156] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0157] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0158] Various aspects of this disclosure are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0159] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processor of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner; thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0160] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions that execute on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0161] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions. It will be known to those skilled in the art that implementation in hardware, implementation in software, and implementation in a combination of software and hardware are equivalent.

[0162] The various embodiments of this disclosure have been described above. These descriptions are exemplary and not exhaustive, and are not limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or technical improvements to the embodiments in the market, or to enable others skilled in the art to understand the embodiments disclosed herein. The scope of this disclosure is defined by the appended claims.

Claims

1. A method for determining trademark similarity, comprising: Obtain the first character information corresponding to the first trademark and the second character information corresponding to the second trademark; Based on the first character information, first feature information of all characters in the first trademark is obtained, and based on the second character information, second feature information of all characters in the second trademark is obtained; Based on the first feature information and the second feature information, determine whether the first trademark and the second trademark are similar trademarks; Wherein, the first feature information is obtained by the gated recurrent unit in the first neural network model by processing the first Chinese character embedding vector, the first Pinyin embedding vector and the first glyph vector corresponding to the characters in the first trademark, and the first Chinese character embedding vector, the first Pinyin embedding vector and the first glyph vector are determined by the first character information; The step of determining whether the first trademark and the second trademark are similar trademarks based on the first feature information and the second feature information includes: obtaining the ij-th similarity value in the similarity matrix based on the first feature information of the i-th character in the first trademark and the second feature information of the j-th character in the second trademark, wherein i is a positive integer and is less than or equal to the total number of characters in the first trademark, and j is a positive integer and is less than or equal to the total number of characters in the second trademark; and determining whether the first trademark and the second trademark are similar trademarks based on the similarity matrix.

2. The method according to claim 1, wherein determining whether the first trademark and the second trademark are similar trademarks based on the similarity matrix comprises: Based on the similarity matrix, a target similarity vector is obtained, wherein the target similarity vector is composed of the maximum value of each row in the similarity matrix; Based on the target similarity vector, the first similarity between the first trademark and the second trademark is determined.

3. The method according to claim 1, wherein the first character information includes the Chinese character information, pinyin information, and Wubi input method root information of the corresponding character in the first trademark; The step of obtaining the first feature information of all characters in the first trademark based on the first character information includes: Based on the first character information, obtain the first Chinese character information, the first Pinyin information, and the first Wubi input root information corresponding to all characters in the first trademark; Based on the first Chinese character information and the first Pinyin information, obtain the first Chinese character embedding vector and the first Pinyin embedding vector; and, Based on the first Wubi character root information, a first character shape vector is obtained, wherein the first character shape vector represents the character shape; The first Chinese character embedding vector, the first Pinyin embedding vector, and the first character shape vector are subjected to a first unified connection process to obtain the first feature information.

4. The method according to claim 3, wherein obtaining the first character shape vector based on the first Wubi character root information includes: Based on the first Wubi character root information, obtain the first Wubi character root embedding vector that corresponds one-to-one with the Wubi character root of the character; The first Wubi character root embedding vector is subjected to a second unified connection process to obtain the first character shape vector.

5. The method according to claim 1, further comprising: The steps for establishing a trademark database after preprocessing known trademarks specifically include: For a known trademark, obtain the pinyin and / or Wubi code corresponding to the Chinese characters of the known trademark; A trademark database is established based on the Chinese characters, pinyin, and Wubi codes of the known trademarks, as well as the classification information and registrant information of the known trademarks.

6. The method according to claim 5, wherein the second trademark is obtained by means of: Based on the first trademark, obtain the top n trademarks with the second similarity ranking from the trademark database, wherein... n is a positive integer; Choose one trademark from the first n trademarks as the second trademark.

7. A method for predicting trademark application conflicts, comprising: Obtain the target trademark to be applied for; Obtain a set of similar trademarks corresponding to the target trademark; A similarity set is obtained based on the target trademark and the set of similar trademarks, wherein a first similarity in the similarity set represents the degree of similarity between the target trademark and the trademarks in the set of similar trademarks, and the first similarity is obtained by the method according to any one of claims 1-6; Based on the similarity set, the preliminary application result of the target trademark is obtained.

8. The method according to claim 7, further comprising: A list of similar trademarks is generated based on the set of similar trademarks and the set of similarities, wherein the list of similar trademarks includes multiple data pairs, each data pair consisting of a trademark in the set of similar trademarks and a first similarity score corresponding to the trademark in the set of similarities.

9. The method according to claim 8, wherein the method is applied to a terminal device, and the method further comprises: The list of similar trademarks and the predicted application results are displayed.

10. An apparatus for determining trademark similarity, comprising: The character information acquisition module is used to acquire the first character information corresponding to the first trademark and the second character information corresponding to the second trademark. The feature information acquisition module is used to obtain first feature information of all characters in the first trademark based on the first character information, and to obtain second feature information of all characters in the second trademark based on the second character information; A similar trademark determination module is used to determine whether the first trademark and the second trademark are similar trademarks based on the first feature information and the second feature information; Wherein, the first feature information is obtained by the gated recurrent unit in the first neural network model by processing the first Chinese character embedding vector, the first Pinyin embedding vector and the first glyph vector corresponding to the characters in the first trademark, and the first Chinese character embedding vector, the first Pinyin embedding vector and the first glyph vector are determined by the first character information; The similar trademark determination module is used to determine whether the first trademark and the second trademark are similar trademarks by means of the following steps based on the first feature information and the second feature information: obtaining the ij-th similarity value in the similarity matrix based on the first feature information of the i-th character in the first trademark and the second feature information of the j-th character in the second trademark, wherein i is a positive integer and is less than or equal to the total number of characters in the first trademark, and j is a positive integer and is less than or equal to the total number of characters in the second trademark; and determining whether the first trademark and the second trademark are similar trademarks based on the similarity matrix.

11. An electronic device comprising the apparatus of claim 10; or comprising: Memory is used to store executable instructions; A processor, configured to, under the control of the executable instructions, run the electronic device to perform the method according to any one of claims 1-9.

12. A computer-readable storage medium, wherein, The computer-readable storage medium stores a computer program that can be read and executed by a computer, the computer program being used, when read and executed by the computer, to perform the method according to any one of claims 1-9.

Images