Equipment fingerprint retrieving method and device, electronic equipment and storage medium

Abstract

The invention relates to an equipment fingerprint retrieving method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring an equipment fingerprint retrieving request of equipment to be retrieved; acquiring device fingerprint information of the to-be-retrieved device based on the device fingerprint retrieval request; converting the equipment fingerprint information into a target vector based on a pre-trained word vector model; and determining a first vector from a preset vector database, wherein the similarity between the first vector and the target vector meets a preset condition. According to the method, the device fingerprint information of the to-be-retrieved device can be obtained based on the device fingerprint retrieval request of the to-be-retrieved device, and the device fingerprint information is converted into the target vector based on the pre-trained word vector model; therefore, the first vector of which the similarity with the target vector meets the preset condition can be quickly and accurately determined from the preset vector database, and the target fingerprint information corresponding to the first vector is timely issued to the to-be-retrieved equipment, so that the equipment fingerprint retrieval speed and accuracy are improved.

Description

Technical Field

[0001] The present application relates to the field of device fingerprint technology, and in particular to a device fingerprint retrieval method, apparatus, electronic device and storage medium. Background Art

[0002] Device fingerprint is a unique ID assigned to a device within a software application, which helps track, authenticate and manage the device. Every time a device opens the same application App, the App backend needs to determine whether the device has been issued a device fingerprint in the past based on the dimensional information collected on the device. If so, it will find the issued fingerprint. If not, it will issue a new device fingerprint to the device. This process requires high timeliness and accuracy. Normally, for example, when a user opens an App to watch a video on a mobile phone, a large number of programs need to be run in an instant. The order of the device fingerprint retrieval process is very forward, because a large number of device features and tags are based on device fingerprints (for example, the display of splash screen ads and the display of home page recommendation streams all rely on device fingerprints). Once the retrieval fails or times out, it is a brand new device for the post-processing link, and the strategy that can be adopted is completely different from that of the old device. In addition, the accuracy of the retrieval is also critical. As mentioned above, a large number of tags are printed on the device fingerprint. If the retrieval is inaccurate, the tag of device A will be moved to device B, resulting in a series of misjudgments of services. Summary of the invention

[0003] The present application provides a device fingerprint retrieval method, apparatus, electronic device and storage medium to solve the technical problem of how to quickly and accurately realize device fingerprint retrieval.

[0004] In a first aspect, the present application provides a device fingerprint retrieval method, the method comprising:

[0005] Obtain a device fingerprint retrieval request for the device to be retrieved;

[0006] Acquire the device fingerprint information of the device to be retrieved based on the device fingerprint retrieval request;

[0007] Converting the device fingerprint information into a target vector based on a pre-trained word vector model;

[0008] Determine a first vector from a preset vector database, the similarity of which with the target vector satisfies a preset condition;

[0009] The target fingerprint information corresponding to the first vector is sent to the device to be retrieved.

[0010] Optionally, determining a first vector whose similarity with the target vector satisfies a preset condition from a preset vector database includes:

[0011] Calculate K second vectors closest to the target vector from the preset vector database by using a K nearest neighbor algorithm;

[0012] Select the first M third vectors from the K second vectors; wherein M is less than K;

[0013] Scoring the third vector based on different dimensions of the target vector to obtain M candidate fingerprint scores;

[0014] When the target candidate fingerprint score with the highest score among all the candidate fingerprint scores is greater than or equal to a preset threshold, the vector corresponding to the target candidate fingerprint score is used as the first vector.

[0015] Optionally, before calculating K second vectors closest to the target vector from the preset vector database using a K nearest neighbor algorithm, the method further includes:

[0016] Obtain a multidimensional vector database;

[0017] Based on the word vector model, each multidimensional vector in the multidimensional vector database is averaged and pooled to obtain the vector database.

[0018] Optionally, the method further comprises:

[0019] When the score of the target candidate fingerprint with the highest score is less than a preset threshold, a new device fingerprint is issued for the device to be retrieved.

[0020] Optionally, scoring the third vector based on different dimensions of the target vector comprises:

[0021] configuring a scorecard for the third vector;

[0022] Comparing different dimensions of the third vector and the target vector one by one, and updating the scorecard according to whether the values ​​of the third vector and the target vector in each dimension are the same;

[0023] After all dimensions of the third vector are scored, the current score of the score card is used as the candidate fingerprint score of the third vector.

[0024] Optionally, after selecting the first M third vectors from the K second vectors, the method further includes:

[0025] Perform feature encryption based on N feature vectors in the device fingerprint information of the device to be retrieved to obtain encrypted features;

[0026] Determine a feature vector identical to the encryption feature from the vector database; and use the feature vector as the third vector.

[0027] Optionally, converting the device fingerprint information into a target vector based on a pre-trained word vector model includes:

[0028] Get historical device fingerprint data;

[0029] Training an initial word vector model according to the historical device fingerprint data to obtain the word vector model;

[0030] The device fingerprint information is input into the word vector model to obtain a target vector output by the word vector model.

[0031] In a second aspect, the present application provides a device fingerprint retrieval apparatus, the apparatus comprising:

[0032] A first acquisition module is used to obtain a device fingerprint retrieval request for a device to be retrieved;

[0033] A second acquisition module, configured to acquire the device fingerprint information of the device to be retrieved based on the device fingerprint retrieval request;

[0034] A conversion module, used to convert the device fingerprint information into a target vector based on a pre-trained word vector model;

[0035] A determination module, used to determine a first vector whose similarity with the target vector satisfies a preset condition from a preset vector database;

[0036] The retrieval module is used to send the target fingerprint information corresponding to the first vector to the device to be retrieved.

[0037] In a third aspect, the present application provides an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus;

[0038] Memory, used to store computer programs;

[0039] The processor is used to implement the device fingerprint retrieval method described in any embodiment of the first aspect when executing the program stored in the memory.

[0040] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the device fingerprint retrieval method as described in any embodiment of the first aspect is implemented.

[0041] The above technical solution provided by the embodiment of the present application has the following advantages over the prior art: the method provided by the embodiment of the present application obtains a device fingerprint retrieval request of the device to be retrieved; obtains the device fingerprint information of the device to be retrieved based on the device fingerprint retrieval request; converts the device fingerprint information into a target vector based on a pre-trained word vector model; determines a first vector whose similarity with the target vector meets a preset condition from a preset vector database; and sends the target fingerprint information corresponding to the first vector to the device to be retrieved. This method can obtain the device fingerprint information of the device to be retrieved based on the device fingerprint retrieval request of the device to be retrieved, and convert the device fingerprint information into a target vector based on a pre-trained word vector model. Since the comparison between vectors can be performed quickly and accurately, the first vector whose similarity with the target vector meets the preset condition can be determined quickly and accurately from the preset vector database, and the target fingerprint information corresponding to the first vector can be sent to the device to be retrieved in a timely manner, thereby improving the speed and accuracy of device fingerprint retrieval. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0043] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0044] One or more embodiments are exemplarily described by pictures in the corresponding drawings, and these exemplified descriptions do not constitute limitations on the embodiments. Elements with the same reference numerals in the drawings represent similar elements, and unless otherwise stated, the figures in the drawings do not constitute proportional limitations.

[0045] Figure 1 A system architecture diagram of a device fingerprint retrieval method provided for one embodiment of the present application;

[0046] Figure 2 A schematic diagram of a process for retrieving a device fingerprint provided by an embodiment of the present application;

[0047] Figure 3 A schematic diagram of the structure of a device fingerprint retrieval apparatus provided in one embodiment of the present application;

[0048] Figure 4 A schematic diagram of the structure of an electronic device provided in accordance with an embodiment of the present application. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0050] The disclosure below provides many different embodiments or examples to realize the different structures of the present application. In order to simplify the disclosure of the present application, the parts and settings of specific examples are described below. Of course, they are only examples, and the purpose is not to limit the present application. In addition, the present application can repeat reference numbers and / or letters in different examples. This repetition is for the purpose of simplification and clarity, and does not itself indicate the relationship between the various embodiments and / or settings discussed.

[0051] In order to solve the technical problem of how to quickly and accurately realize device fingerprint retrieval in the prior art, the present application provides a device fingerprint retrieval method, apparatus, electronic device and storage medium, which can convert device fingerprint information into a target vector based on a pre-trained word vector model. Since the comparison between vectors can be performed quickly and accurately, the first vector whose similarity with the target vector meets the preset conditions can be quickly and accurately determined from the preset vector database, and the target fingerprint information corresponding to the first vector can be promptly sent to the device to be retrieved, thereby improving the speed and accuracy of device fingerprint retrieval.

[0052] The first embodiment of the present application provides a device fingerprint retrieval method, which can be applied to Figure 1 The system architecture shown in the figure includes at least a terminal 101 and a server 102, and the terminal 101 and the server 102 establish a communication connection. The terminal 101 can be a device to be retrieved, for example, a desktop computer, a tablet computer, a notebook, a mobile terminal, etc. The server 102 can be a local server, a cloud server, or a server cluster.

[0053] The method can be applied to the server 102 in the system architecture. Next, based on the system architecture, the device fingerprint retrieval method is described in detail. Figure 2 , the device fingerprint retrieval method includes:

[0054] Step 201: Obtain a device fingerprint retrieval request for a device to be retrieved.

[0055] The device to be retrieved may be a terminal device that opens an application app. When the user operates the terminal device to open the app, a device fingerprint retrieval request of the terminal device is triggered.

[0056] Step 202: Obtain device fingerprint information of the device to be retrieved based on the device fingerprint retrieval request.

[0057] The server side can obtain device fingerprint information from the device to be retrieved based on the device fingerprint retrieval request of the terminal device, for example, obtaining the device fingerprint information of the device to be retrieved based on a series of information such as the hardware information and software information of the device to be retrieved. There are many parameters of the device to be retrieved, such as model parameters. For example, the model parameter corresponding to a certain product tablet Mi Pad 5 is 21051182C. The model parameters of the previous generation of the product may be slightly different from this character, but they are similar in parameters such as screen size, chip type, memory, foundry code, etc. Through vectorization technology, we can realize that although the two model parameters are not so similar in text, the co-occurrence of other parameters can also be used to calculate that the two parameters are close in vector. Among them, co-occurrence refers to the probability of any two device parameters appearing together. For example, if a device has 100 parameters, and the model parameters of two devices are inconsistent in text, but the values ​​of the remaining 90 parameters are the same, then the co-occurrence of the two model parameters will be similar. By accumulating the collinear distribution of multiple remaining parameters, it can be determined that although the model parameters of the two generations of tablet products have different values, their co-occurrence with most of the other parameters is very close.

[0058] Step 203: Convert the device fingerprint information into a target vector based on the pre-trained word vector model.

[0059] The word vector model can be a word vector model based on a neural network, such as the Word2Vec model. The Word2Vec model is a model proposed by Google for natural language processing. It can generate word vectors using a large amount of text data through unsupervised learning methods, so that words can be mapped to a high-dimensional vector space, so that words with similar semantics are close to each other in the vector space. The core idea is to capture the semantic information of words through the frequency relationship of the occurrence of context words. There are two main training methods for the Word2Vec model:

[0060] 1. Continuous Bag of Words (CBOW):

[0061] Goal: Predict the central word given known context words.

[0062] Method: Given a window size, the surrounding context words are input, and the model predicts the central word by learning the probability of these context words co-occurring. The CBOW model is essentially a classification task that attempts to classify each central word in a large corpus.

[0063] 2. Skip-Gram:

[0064] Goal: Predict context words based on the current word.

[0065] Method: Given a central word, the model tries to predict the words that may appear around it. The Skip-Gram model is trained by maximizing the co-occurrence probability of the current word and its context words.

[0066] By training the word2vec model, a vector of a preset length for each dimension can be obtained. Taking the preset length of 128 as an example, a vector of 128 dimensions can be used to represent each device fingerprint dimension, that is, the target vector can be a 128-dimensional vector.

[0067] In one embodiment, device fingerprint information is converted into a target vector based on a pre-trained word vector model, including: obtaining historical device fingerprint data; training an initial word vector model based on the historical device fingerprint data to obtain a word vector model; inputting the device fingerprint information into the word vector model to obtain a target vector output by the word vector model.

[0068] In this embodiment, the initial word2vec model can be trained by historical device fingerprint data to obtain a trained word2vec model. For example, each piece of historical device fingerprint data is trained to obtain a 128-dimensional vector corresponding to each device fingerprint dimension. This vector can be 128 decimals between -1 and 1. In order to simplify the subsequent vector comparison process, the vector corresponding to the device fingerprint information can be averaged and pooled. For example, if a device fingerprint has 100 dimensions, then there are 100 128-dimensional vectors. These 100 vectors can be averaged bit by bit (for example, the first digit of each vector is taken, and the 100 first digits are averaged to obtain the first digit of the final vector), which can be simplified into a final 128-dimensional vector to represent the device fingerprint information. In actual use, the device fingerprint information is input into the trained word vector model, and a 128-dimensional target vector output by the word vector model can be obtained.

[0069] Step 204: determine, from a preset vector database, a first vector whose similarity to the target vector satisfies a preset condition.

[0070] In one embodiment, determining a first vector whose similarity to a target vector satisfies preset conditions from a preset vector database includes: calculating K second vectors closest to the target vector from a preset vector database through a K nearest neighbor algorithm; selecting the first M third vectors from the K second vectors; wherein M is less than K; scoring the third vectors based on different dimensions of the target vector to obtain M candidate fingerprint scores; and when the target candidate fingerprint score with the highest score among all candidate fingerprint scores is greater than or equal to a preset threshold, using the vector corresponding to the target candidate fingerprint score as the first vector.

[0071] In this embodiment, the k-nearest neighbor algorithm can be used to calculate the cosine distance / Euclidean distance between the target vector converted from the current device fingerprint information and all other vectors in the vector database. The closer the distance, the more similar the vectors are. The k other vectors (i.e., second vectors) most similar to the target vector can be obtained. Furthermore, the top M third vectors ranked by similarity can be selected from the k second vectors, and the third vectors can be scored based on different dimensions of the target vector to obtain M candidate fingerprint scores. If the score of the target candidate fingerprint with the highest score among all candidate fingerprint scores is greater than or equal to a preset threshold, the vector corresponding to the target candidate fingerprint with the highest score can be used as the first vector.

[0072] In one embodiment, the third vector is scored based on different dimensions of the target vector, including: configuring a scoring card for the third vector; comparing the different dimensions of the third vector and the target vector one by one, and updating the scoring card according to whether the values ​​of the third vector and the target vector in each dimension are the same; after the scoring of each dimension of the third vector is completed, the current score of the scoring card is used as the candidate fingerprint score of the third vector.

[0073] In this embodiment, in the process of scoring the third vector, a scoring card can be configured for the third vector, and the different dimensions of the third vector and the target vector can be compared one by one. If the two vectors have the same value in a certain dimension, the dimension is scored. After the scores of all dimensions of the third vector are completed, the current score is used as the final score of the third vector, and the final score is the candidate fingerprint score of the third vector. For example, the most similar top 50 candidate inventory fingerprints are quickly calculated through the vector database, and then the similarity of these candidate fingerprints and the current device is finally sorted through the scoring card. The scoring card is equivalent to a weight for each dimension. For example, whether the dimension A of the current device and the dimension A of the candidate device 1 are exactly the same, if they are the same, the weight of dimension A is obtained, and if they are different, no score is obtained. Finally, all the scores obtained are accumulated. If there are 50 candidate fingerprints, 50 scoring calculations are performed. If the candidate fingerprint with the highest score exceeds the preset threshold, it is considered to be successfully retrieved, and this inventory fingerprint can be sent to the requested device to be retrieved.

[0074] In one embodiment, before calculating the K second vectors closest to the target vector from a preset vector database through the K nearest neighbor algorithm, the method also includes: obtaining a multidimensional vector database; based on a word vector model, average-pooling each multidimensional vector in the multidimensional vector database to obtain a vector database.

[0075] In this embodiment, in order to simplify the subsequent vector comparison process, each multidimensional vector in the multidimensional vector database can be averaged and pooled. For example, if a device fingerprint has 100 dimensions, then there are 100 128-dimensional vectors corresponding to it. These 100 vectors can be averaged bit by bit (for example, the first digit of each vector is taken, and the 100 first digits are averaged to obtain the first digit of the final vector), and they can be simplified into a final 128-dimensional vector to represent the device fingerprint information, thereby obtaining a database of all simplified vectors.

[0076] In one embodiment, the method further includes: when the score of the target candidate fingerprint with the highest score is less than a preset threshold, issuing a new device fingerprint for the device to be retrieved.

[0077] In this embodiment, if the score of the target candidate fingerprint with the highest score is less than the preset threshold, it means that there is no fingerprint information matching the target vector in the vector database, that is, the device to be retrieved is considered to be a new device, and a new device fingerprint needs to be issued for the device to be retrieved.

[0078] In one embodiment, after selecting the first M third vectors from the K second vectors, the method further includes: encrypting features based on N feature vectors in the device fingerprint information of the device to be retrieved to obtain encrypted features; determining a feature vector identical to the encrypted feature from a vector database; and using the feature vector as the third vector.

[0079] In this embodiment, after selecting the first M third vectors from the K second vectors, feature encryption can also be performed based on the N feature vectors in the device fingerprint information of the device to be retrieved to obtain encrypted features. For example, the message digest algorithm MD5 encryption is performed based on 10 feature vectors in the device fingerprint information of the device to be retrieved to obtain an encrypted hash value, and the feature vector identical to the encrypted hash value is determined from the vector database. The feature vector with the same hash value is also used as the third vector, and subsequent scoring is performed together with the selection of the first M third vectors from the K second vectors to determine whether it can be retrieved.

[0080] Step 205: Send the target fingerprint information corresponding to the first vector to the device to be retrieved.

[0081] The method can obtain the device fingerprint information of the device to be retrieved based on the device fingerprint retrieval request of the device to be retrieved, and convert the device fingerprint information into a target vector based on a pre-trained word vector model. Since the comparison between vectors can be performed quickly and accurately, the first vector whose similarity with the target vector meets the preset conditions can be quickly and accurately determined from a preset vector database, and the target fingerprint information corresponding to the first vector is promptly sent to the device to be retrieved, thereby improving the speed and accuracy of device fingerprint retrieval.

[0082] In a specific embodiment, a device fingerprint retrieval method includes:

[0083] 1. Train the word2vec model using inventory device fingerprint data

[0084] Here you can use the word2vec algorithm module provided by spark to concatenate the dimensions of the inventory fingerprint with spaces into a piece of text for training. After completion, you will get a 128-dimensional vector corresponding to each value.

[0085] 2. Inventory fingerprints are vectorized and stored in the vector database

[0086] The BES vector database provided by Baidu Cloud can be used here. Through the word2vec model trained in the previous step, each dimension of the inventory fingerprint can be converted into a 128-dimensional vector. After that, average pooling is performed to obtain a unique 128-dimensional vector to represent the fingerprint, and then the fingerprint ID and the 128-dimensional vector are stored in the database.

[0087] 3. Input Request Vectorization

[0088] Through the word2vec model trained in the first step, the device fingerprint retrieval request received in the subsequent online application is vectorized, which is equivalent to receiving the detailed dimensional information of the specific device to be retrieved. Through the model + average pooling, the device to be retrieved can be converted into a 128-dimensional vector.

[0089] 4. Input vector to do nearest neighbor query through vector database

[0090] Input this vector into the BES vector database provided by Baidu Cloud, and request to query the 50 most similar inventory fingerprints and return 50 inventory fingerprint IDs.

[0091] 5. Calculate the score for each neighbor inventory fingerprint

[0092] Through the scoring card, the neighboring fingerprints and the input fingerprint to be retrieved are matched one by one, and the total score is calculated based on whether the values ​​of each dimension are exactly the same. The one with the highest total score is found, and its score is judged whether it is higher than the preset retrieval threshold. If it is exceeded, it is considered that the input device is retrieved by the device with the highest score. If not, it is considered that the input device is a new device and a new fingerprint is reissued.

[0093] In this embodiment, retrieval is performed in a vectorized manner, and the retrieval efficiency is significantly improved, and the retrieval can be completed within 50ms, and is not easily affected by the instability caused by missing dimensions or small changes.

[0094] Based on the same technical concept, the second embodiment of the present application provides a device fingerprint retrieval device, such as Figure 3 , the device comprises:

[0095] The first acquisition module 301 is used to obtain a device fingerprint retrieval request for the device to be retrieved;

[0096] A second acquisition module 302 is used to acquire the device fingerprint information of the device to be retrieved based on the device fingerprint retrieval request;

[0097] A conversion module 303, configured to convert the device fingerprint information into a target vector based on a pre-trained word vector model;

[0098] A determination module 304 is used to determine a first vector whose similarity with the target vector satisfies a preset condition from a preset vector database;

[0099] The retrieval module 305 is used to send the target fingerprint information corresponding to the first vector to the device to be retrieved.

[0100] The device can obtain device fingerprint information of the device to be retrieved based on a device fingerprint retrieval request of the device to be retrieved, and convert the device fingerprint information into a target vector based on a pre-trained word vector model. Since the comparison between vectors can be performed quickly and accurately, the first vector whose similarity with the target vector meets preset conditions can be quickly and accurately determined from a preset vector database, and the target fingerprint information corresponding to the first vector is promptly sent to the device to be retrieved, thereby improving the speed and accuracy of device fingerprint retrieval.

[0101] like Figure 4 As shown, an embodiment of the present application provides an electronic device, including a processor 111, a communication interface 112, a memory 113 and a communication bus 114, wherein the processor 111, the communication interface 112, and the memory 113 communicate with each other through the communication bus 114.

[0102] Memory 113, used for storing computer programs;

[0103] In one embodiment of the present application, the processor 111 is used to execute the program stored in the memory 113 to implement the device fingerprint retrieval method provided by any of the above method embodiments, including:

[0104] Obtain a device fingerprint retrieval request for the device to be retrieved;

[0105] Acquire the device fingerprint information of the device to be retrieved based on the device fingerprint retrieval request;

[0106] Converting the device fingerprint information into a target vector based on a pre-trained word vector model;

[0107] Determine a first vector from a preset vector database, the similarity of which with the target vector satisfies a preset condition;

[0108] The target fingerprint information corresponding to the first vector is sent to the device to be retrieved.

[0109] The communication bus mentioned in the above terminal can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0110] The communication interface is used for communication between the above terminal and other devices.

[0111] The memory may include a random access memory (RAM) or a non-volatile memory, such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0112] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0113] An embodiment of the present application further provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the device fingerprint retrieval method provided by any of the aforementioned method embodiments is implemented.

[0114] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0115] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a general hardware platform, and of course, by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the relevant technology can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0116] It should be understood that the terms used herein are only for the purpose of describing specific example embodiments and are not intended to be limiting. Unless the context clearly indicates otherwise, the singular forms "one", "an" and "said" as used herein may also be meant to include plural forms. The terms "include", "comprise", "contain", and "have" are inclusive, and therefore specify the existence of stated features, steps, operations, elements and / or parts, but do not exclude the existence or addition of one or more other features, steps, operations, elements, parts, and / or combinations thereof. The method steps, processes, and operations described herein are not interpreted as necessarily requiring them to be performed in the specific order described or illustrated, unless the execution order is clearly indicated. It should also be understood that additional or alternative steps may be used.

[0117] It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application. In the description, the suffixes such as "module", "component" or "unit" used to represent the elements are only used to facilitate the description of the present application and have no specific meaning in themselves. Therefore, "module", "component" or "unit" can be used in a mixed manner.

[0118] The above description is only a specific implementation of the present application, so that those skilled in the art can understand or implement the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest range consistent with the principles and novel features applied for herein.

Claims

1. A device fingerprint retrieval method, characterized in that: The method comprises: Obtain a device fingerprint retrieval request for the device to be retrieved; Acquire the device fingerprint information of the device to be retrieved based on the device fingerprint retrieval request; Converting the device fingerprint information into a target vector based on a pre-trained word vector model; Determine a first vector from a preset vector database whose similarity with the target vector satisfies a preset condition; The target fingerprint information corresponding to the first vector is sent to the device to be retrieved.

2. The method according to claim 1, characterized in that Determining a first vector from a preset vector database whose similarity to the target vector meets a preset condition, comprising: Calculate K second vectors closest to the target vector from the preset vector database by using a K nearest neighbor algorithm; Select the first M third vectors from the K second vectors; wherein M is less than K; Scoring the third vector based on different dimensions of the target vector to obtain M candidate fingerprint scores; When the target candidate fingerprint score with the highest score among all the candidate fingerprint scores is greater than or equal to a preset threshold, the vector corresponding to the target candidate fingerprint score is used as the first vector.

3. The method according to claim 2, characterized in that Before calculating the K second vectors closest to the target vector from the preset vector database by using the K nearest neighbor algorithm, the method further includes: Obtain a multidimensional vector database; Based on the word vector model, each multidimensional vector in the multidimensional vector database is averaged and pooled to obtain the vector database.

4. The method according to claim 2, characterized in that: The method further comprises: When the score of the target candidate fingerprint with the highest score is less than a preset threshold, a new device fingerprint is issued for the device to be retrieved.

5. The method according to claim 2, characterized in that: Scoring the third vector based on different dimensions of the target vector includes: configuring a scorecard for the third vector; Comparing different dimensions of the third vector and the target vector one by one, and updating the scorecard according to whether the values ​​of the third vector and the target vector in each dimension are the same; After all dimensions of the third vector are scored, the current score of the score card is used as the candidate fingerprint score of the third vector.

6. The method according to claim 2, characterized in that After selecting the first M third vectors from the K second vectors, the method further includes: Perform feature encryption based on N feature vectors in the device fingerprint information of the device to be retrieved to obtain encrypted features; Determine a feature vector identical to the encryption feature from the vector database; and use the feature vector as the third vector.

7. The method according to claim 1, characterized in that The device fingerprint information is converted into a target vector based on a pre-trained word vector model, including: Get historical device fingerprint data; Training an initial word vector model according to the historical device fingerprint data to obtain the word vector model; The device fingerprint information is input into the word vector model to obtain a target vector output by the word vector model.

8. A device fingerprint retrieval device, characterized in that: The device comprises: A first acquisition module is used to obtain a device fingerprint retrieval request for a device to be retrieved; A second acquisition module, configured to acquire the device fingerprint information of the device to be retrieved based on the device fingerprint retrieval request; A conversion module, used to convert the device fingerprint information into a target vector based on a pre-trained word vector model; A determination module, used to determine a first vector whose similarity with the target vector satisfies a preset condition from a preset vector database; The retrieval module is used to send the target fingerprint information corresponding to the first vector to the device to be retrieved.

9. An electronic device, characterized in that: It includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus; Memory, used to store computer programs; The processor is used to implement the device fingerprint retrieval method described in any one of claims 1 to 7 when executing the program stored in the memory.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the device fingerprint retrieval method according to any one of claims 1 to 7 is implemented.

Images