Rapid labeling method and system for checking article and label information

Abstract

The method comprises the following steps: extracting key information from original information of an article and a label, converting the key information into a plurality of computable features, and combining the plurality of computable features into two feature vectors; carrying out normalization processing on the feature vectors, converting the feature vectors into quantum states, carrying out similarity calculation on the quantum states, outputting a matching degree value according to a similarity calculation result, and generating a matching result of article and label checking by utilizing the matching degree value; according to a matching result of article and label checking, selecting whether to construct an optimization problem of article and label checking, if so, solving the optimization problem, and readjusting the matching result of article and label checking, otherwise, directly outputting the matching result to a checking result; through the quantum computing technology, the matching efficiency and accuracy in a large-scale data scene are improved, and the urgent requirements of modern logistics and supply chains for efficient and accurate labeling are met.

Description

Technical Field

[0001] The present invention relates to the field of information verification and quantum computing technology, and in particular to a rapid labeling method and system for verifying information between an object and a label. Background Art

[0002] In recent years, with the rapid development of logistics, warehousing and retail industries, the importance of commodity identification technology has become increasingly prominent. In order to ensure the efficient management of commodities during storage and circulation, the matching and verification of items and label information has become a key task. Traditional identification methods mainly rely on barcode scanning, manual verification, and rule algorithm calculation. These technologies have a certain degree of reliability in realizing basic identification functions, especially in small and medium-sized logistics scenarios. However, with the rapid increase in the types and quantities of items, especially in e-commerce and large-scale supply chain management, traditional identification and verification methods have gradually exposed their shortcomings. For example, traditional verification methods often rely on static rules and single-threaded processing, which cannot effectively adapt to the explosive growth of data scale. At the same time, the information matching between labels and items often requires a lot of manual intervention, which makes it difficult to balance the identification speed and accuracy, further limiting the overall efficiency and adaptability of the system. In addition, traditional methods show obvious limitations when dealing with high-dimensional, multi-feature items and label information.

[0003] The specific shortcomings are as follows: First, when faced with the fusion of multiple types of data (such as numerical, text, etc.), traditional data processing methods are difficult to achieve efficient and unified feature extraction. This limitation directly leads to a decrease in the accuracy of matching items and labels, especially when the data noise is high or the data distribution is uneven. Secondly, the existing verification algorithms generally use linear similarity metrics in feature matching, ignoring the nonlinear relationship between multi-dimensional features, making the matching results unreliable in complex scenarios. More importantly, as the scale of the system expands, traditional optimization algorithms often find it difficult to achieve a global optimal solution when solving the problem of allocating items and labels due to excessive computational complexity, resulting in a decrease in the efficiency of resource allocation and an increase in the mismatch rate. The above-mentioned shortcomings not only affect the efficiency of checking items and labels, but also limit the applicability of the system in dynamic and high-concurrency environments.

[0004] Therefore, a new method is urgently needed to effectively integrate multiple feature information, improve verification efficiency and accuracy, and achieve rapid optimization allocation in the event of matching failure, so as to meet the urgent needs of modern logistics and supply chains for efficient and accurate labeling. Summary of the invention

[0005] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0006] In view of the above existing problems, the present invention is proposed. Therefore, the present invention provides a rapid labeling method for checking information of an article and a label, so as to solve the problems mentioned in the background technology.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In a first aspect, the present invention provides a rapid labeling method for checking information of an item and a label, comprising:

[0009] Extract key information from original information of the item and the tag, process the key information, convert it into multiple computable features, and combine the multiple computable features into two feature vectors;

[0010] By normalizing the feature vector, converting the normalized feature vector into a quantum state, performing similarity calculation on the quantum state, outputting a matching value according to the similarity calculation result, and using the matching value to generate a matching result of the object and the label verification;

[0011] According to the matching result of the item and label verification, choose whether to construct an optimization problem for the item and label verification. If the optimization problem for the item and label verification is constructed, solve the optimization problem and readjust the matching result of the item and label verification. Otherwise, directly output the matching result to the verification result to complete the rapid labeling of the item and label information.

[0012] As a preferred solution of the rapid labeling method for checking the information of articles and labels of the present invention, the key information is processed and converted into a plurality of computable features, including:

[0013] For the unique identifiers in the key information, i.e., the IDs of the items and tags, a standard hash mapping method is used to convert the IDs of the items and tags into corresponding unique values;

[0014] For the numerical information contained in the key information, the equal-width binning method is used to divide the numerical information into several equal-number intervals. Each interval contains the same number of data points. The result of the equal-width binning is a numerical feature.

[0015] For the text information in the key information, a word vector model is used to convert the text information into a numerical vector.

[0016] As a preferred solution of the rapid labeling method for checking the information of articles and labels of the present invention, the multiple computable features are combined into two feature vectors, including:

[0017] The unique value converted from ID, the result of equal-width binning, and the numerical vector converted from text information are combined into two feature vectors;

[0018] The two feature vectors are represented as feature vectors of the object and the label respectively, and the feature vectors of the object and the label are taken as classical vectors, and their dimensions are set to be consistent.

[0019] As a preferred solution of the rapid labeling method for checking the information of the article and the label of the present invention, wherein: by normalizing the feature vector, the normalized feature vector is converted into a quantum state, including:

[0020] Normalizing each numerical vector in the two eigenvectors, and setting the modulus of each numerical vector to 1;

[0021] Using the characteristics of classical vectors, quantum feature mapping is performed on each numerical vector in the feature vectors of objects and tags to obtain a feature mapping function;

[0022] Based on the feature mapping function, quantum states of feature vectors of objects and tags are generated for classical vectors respectively.

[0023] As a preferred solution of the rapid labeling method for checking the information of an object and a label of the present invention, wherein: similarity calculation is performed on the quantum state, a matching value is output according to the calculation result of the similarity, and a matching result of checking the object and the label is generated by using the matching value, including:

[0024] Calculate the inner product of the quantum state of the object and the tag, that is, measure the degree of overlap between the quantum state of the object and the quantum state of the tag. If the degree of overlap is infinitely close to 1, it means that there is feature similarity between the object and the tag. If the degree of overlap is infinitely close to 0, it means that there is no feature similarity between the object and the tag.

[0025] Output the matching value based on the feature similarity between the item and the label;

[0026] Collect historical feature similarities between items and tags, count the feature similarity distribution between all items and tags, obtain the average value and fluctuation range of feature similarity, and use the average value and fluctuation range of feature similarity as the matching threshold;

[0027] If the outputted matching value exceeds the matching threshold, it is considered that the object and the tag are matched successfully, otherwise it is considered that the object and the tag are matched unsuccessfully.

[0028] As a preferred solution of the rapid labeling method for checking the information of the article and the label of the present invention, wherein: according to the matching result of the article and the label, it is selected whether to construct an optimization problem for checking the article and the label; if the optimization problem for checking the article and the label is constructed, the optimization problem is solved and the matching result of the article and the label is readjusted; otherwise, the matching result is directly output to the checking result, including:

[0029] If the item and the label match fails, an optimization problem for checking the item and the label is constructed, wherein the goal of the optimization problem is to reallocate the item and the label;

[0030] The optimization problem is regarded as a quantum annealing problem, by collecting pairs of items and labels that fail to match, establishing feature similarities between the items and labels that fail to match, and defining binary variables;

[0031] By using the similarity values ​​and binary variables between the failed matching items and the labels, an optimization objective function of the total cost of failed matching is constructed, and constraints of the optimization objective function are given;

[0032] The optimization objective function and the constraint conditions are converted into the Hamiltonian of the quantum system to simulate the annealing process.

[0033] As a preferred solution of the rapid labeling method for checking article and label information of the present invention, the simulated annealing process includes:

[0034] The quantum system is set to start in a high-energy state, that is, a matching scheme that randomly assigns items to labels, and each binary variable is initialized to a random value;

[0035] The quantum system is set to gradually reduce the Hamiltonian of the quantum system, and the global optimal solution is searched through the quantum tunneling effect. When the quantum system converges to a low-energy state, the matching solution between the currently allocated items and the labels is output as the optimal solution.

[0036] According to the optimization scheme, the matching value between the object and the label is recalculated, and the optimized matching value is marked as the matching result of the final object and label information verification.

[0037] In a second aspect, the present invention provides a rapid labeling system for checking article and label information, comprising:

[0038] An object and tag information feature extraction module is configured to extract key information from the original information of the object and the tag, process the key information, convert it into a plurality of computable features, and combine the plurality of computable features into two feature vectors;

[0039] The object and label quantum feature calculation module is configured to perform normalization processing on the feature vector, convert the normalized feature vector into a quantum state, perform similarity calculation on the quantum state, output a matching value according to the similarity calculation result, and generate a matching result of the object and label verification using the matching value;

[0040] The item and label matching optimization module is configured to select whether to construct an optimization problem for item and label verification based on the matching result of the item and label verification. If the optimization problem for item and label verification is constructed, the optimization problem is solved and the matching result of the item and label verification is readjusted. Otherwise, the matching result is directly output to the verification result to complete the rapid labeling of item and label information.

[0041] In a third aspect, the present invention provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, wherein: the processor implements any step of the above method when executing the computer program.

[0042] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: the computer program implements any step of the above method when executed by a processor.

[0043] Compared with the prior art, the invention has the following beneficial effects:

[0044] 1. The present invention can effectively extract and express the multi-dimensional features of objects and tag information by adopting the standardized processing of feature vectors, and the similarity calculation of matching degree combined with the quantum state overlap can significantly improve the matching accuracy in complex scenes;

[0045] 2. By introducing quantum computing technology, especially quantum state mapping and similarity calculation, the efficiency bottleneck of traditional algorithms in high-dimensional feature processing is solved; compared with the traditional linear similarity calculation method, this solution can use quantum parallel processing capabilities to quickly complete the task of matching objects and tags, greatly reducing calculation time and improving verification efficiency;

[0046] 3. For item and label pairs that failed to match, an optimization problem was constructed and solved using the quantum annealing algorithm, which can quickly realize the redistribution of items and labels, avoiding the waste of resources caused by local optimal traps in traditional algorithms, while ensuring the global optimality of the allocation plan and reducing the mismatch rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0048] Figure 1 This is an overall flow chart of a rapid labeling method for checking article and label information according to an embodiment of the present invention;

[0049] Figure 2 A comparison chart of matching accuracy and error rate of a fast labeling method for checking information of an object and a label according to an embodiment of the present invention;

[0050] Figure 3 A comparison chart of the allocation efficiency and optimization success rate of a rapid labeling method for verifying item and label information according to an embodiment of the present invention. DETAILED DESCRIPTION

[0051] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0052] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0053] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0054] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0055] At the same time, in the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0056] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0057] Example 1

[0058] Reference Figure 1 , which is the first embodiment of the present invention, provides a rapid labeling method for checking information of an item and a label, comprising:

[0059] S1. Extract key information from the original information of the object and the tag, process the key information, convert it into multiple computable features, and combine the multiple computable features into two feature vectors;

[0060] Specifically, key information includes: item information: unique identification (item ID), item category (such as electronic products, food), physical size (such as length, width) and storage location, etc.; label information: unique identification (label ID), label content (such as product description, barcode) and timestamp, etc.;

[0061] Furthermore, for the unique identifiers in the key information, i.e., the IDs of the items and tags, a standard hash mapping method is used to convert the IDs of the items and tags into corresponding unique numerical values;

[0062] Specifically, the standard hash mapping method can be expressed by a mathematical formula:

[0063] Hash(ID)=mod(SHA256(ID),10 n )

[0064] Among them, Hash(ID) represents the hash value of the converted item ID or tag ID, SHA256 represents the SHA-256 hash algorithm, and other hash algorithms can also be used, such as the MD5 hash algorithm, etc.; mod represents the modulus operation, which maps the hash result to the specified value range [0,10n ], n represents a power, and its value is at least 1 and an integer;

[0065] It should be noted that the irreversibility and low collision rate of the hash algorithm ensure the uniqueness of the numerical values ​​converted from different items and tags;

[0066] Furthermore, for the numerical information contained in the key information, an equal-width binning method is used to divide the numerical information into a number of equal intervals, each interval containing the same number of data points, and the result of the equal-width binning is directly used as the numerical feature;

[0067] Specifically, assuming that the current numerical information x = {x 1 ,x 2 ,…,x i The range of} is: [x min ,x max ], then it is divided into k intervals of equal number according to the rule, and the width of each interval can be expressed as:

[0068]

[0069] Furthermore, the boundaries of each interval can be expressed as:

[0070] Interval i = [x min +(i-1)·Δx,x min +i·Δx],i=1,2,…,k

[0071] Furthermore, the numerical information is classified into corresponding intervals, and each data point is marked with a corresponding interval number according to the range in which its value falls. For example, interval 1: [x min ,x min +Δx), interval 2: [x min +Δx,x min +2·Δx), [x min +2·Δx,x min +3·Δx), and so on;

[0072] It should be noted that, since the binned data has discrete characteristics, it is more suitable for processing by the quantum annealing algorithm, thereby accelerating the optimization solution process;

[0073] Furthermore, for the text information in the key information, a word vector model is used to convert the text information into a numerical vector;

[0074] Specifically, using a word vector model (such as Word2Vec or BERT), the text information is converted into a numerical vector, which can be expressed as:

[0075] y text=Word2Vec(text)

[0076] It should be noted that the advantage of converting text information into numerical vectors is that when the label or item information changes (new keywords, keyword loss, etc.), only the word vector model needs to be updated, so there is no need to build text information matching rules;

[0077] Furthermore, the unique value converted from the ID, the result of equal-width binning, and the numerical vector converted from the text information are combined into two feature vectors;

[0078] Specifically, the two feature vectors are the feature vector of the item and the feature vector of the label, and the mathematical expression of the combination is as follows:

[0079] X l =[Hash(ID L ),x,y text ]

[0080] X g =[Hash(ID g ),x,y text ]

[0081] Among them, X l is the feature vector of the label, ID L is the ID of the tag, X g is the feature vector of the item, ID g is the ID of the item;

[0082] It should be noted that, due to the combined feature vector operation, the original numerical features are transformed into numerical vector form;

[0083] Furthermore, the feature vectors of items and labels are taken as classic vectors, and their dimensions are set to be consistent;

[0084] It needs to be explained that, because the quantum state is a quantum representation of the classical vector in a high-dimensional space, in quantum computing, the quantum state often needs to input the classical vector as a basis; in addition, the classical vector can also be regarded as a point in the feature space, representing the feature combination of an object or a label; at the same time, stipulating the consistency of the dimensions of the two can ensure that each dimension has a corresponding relationship, that is, the symmetry of the features;

[0085] S2. Normalizing the feature vector, converting the normalized feature vector into a quantum state, performing similarity calculation on the quantum state, outputting a matching value according to the similarity calculation result, and using the matching value to generate a matching result of the object and the label verification;

[0086] Furthermore, each numerical vector in the two eigenvectors is normalized, and the modulus of each numerical vector is set to 1;

[0087] It should be noted that the modulus is defined as 1 because the norm of the quantum state requires that the modulus of the input vector is 1, otherwise the quantum state cannot be correctly constructed;

[0088] The specific, normalized mathematical expression is:

[0089]

[0090] Where X is the normalized feature vector of the item or label, and ‖X‖ represents the modulus of the numerical vector in the feature vector;

[0091] Furthermore, by using the characteristics of classical vectors, quantum feature mapping is performed on each numerical vector in the feature vectors of items and tags to obtain a feature mapping function;

[0092] Specifically, the characteristics of the classical vector refer to representing the original label feature vector and item feature vector as an ordered set of multiple scalars (values);

[0093] Specifically, quantum feature mapping is performed on each numerical vector, and the feature mapping function φ(X) is expressed as:

[0094] φ(X)=[cos(x i +b),sin(x i +b)]

[0095] Where, b is a randomly initialized mapping parameter;

[0096] Furthermore, based on the feature mapping function, the quantum states of the feature vectors of the object and the label are generated for the classical vectors respectively;

[0097] Specifically, the quantum state of the item feature vector |X g >For:

[0098]

[0099] Specifically, the quantum state of the label feature vector |X l >For:

[0100]

[0101] Among them, m and j represent the quantum state ground state, which is used to index the characteristic dimensions m and j; ||φ(X g )|| represents the modulus of the item feature vector mapping function, ||φ(X l )|| represents the modulus of the label feature vector mapping function;

[0102] Furthermore, the inner product of the quantum state of the object and the tag is calculated, that is, the degree of overlap between the quantum state of the object and the quantum state of the tag is measured. If the degree of overlap is infinitely close to 1, it means that there is feature similarity between the object and the tag. If the degree of overlap is infinitely close to 0, it means that there is no feature similarity between the object and the tag.

[0103] Specifically, the mathematical expression of the inner product of the quantum state of the object and the tag can be expressed as:

[0104] S(X g ,X l )=| <X g |X l >| 2

[0105] Among them, S represents the feature similarity between items and labels, | <X g |X l >| is represented as the inner product of the quantum states of the item and the tag;

[0106] Specifically, the feature similarity between the object and the tag indicates the actual matching degree between the two, i.e., the matching value; the higher the feature similarity, the higher the matching value;

[0107] Furthermore, the historical feature similarities of items and tags are collected, and the feature similarity distribution between all items and tags is statistically analyzed to obtain the average value and fluctuation range of the feature similarity, and the average value and fluctuation range of the feature similarity are used as the matching threshold;

[0108] Specifically, the average value of feature similarity represents the central tendency of the matching values ​​of all items and tags, while the fluctuation range represents the dispersion of the matching values;

[0109] Specifically, the matching threshold ∈ can be expressed as:

[0110] ε=μ-p·σ

[0111] Among them, μ represents the average value of feature similarity, σ represents the fluctuation range of feature similarity; p is 1 or 2; when p is 1, it is applicable to the exact matching scenario, and when p is greater than 1, it is applicable to the fuzzy matching scenario;

[0112] Furthermore, if the output matching value exceeds the matching threshold, it is considered that the item and the label are matched successfully, otherwise it is considered that the item and the label are matched unsuccessfully;

[0113] It should be noted that the use of quantum state inner product to calculate the degree of feature overlap can capture the high-dimensional nonlinear relationship between objects and tags, overcome the limitations of traditional linear methods (such as Euclidean distance, cosine similarity, etc.), and make matching calculations more accurate;

[0114] S3. According to the matching result of the item and the label, it is selected whether to construct an optimization problem for the item and the label. If the optimization problem for the item and the label is constructed, the optimization problem is solved and the matching result of the item and the label is readjusted. Otherwise, the matching result is directly output to the verification result to complete the rapid labeling of the item and the label information.

[0115] Furthermore, according to the matching result of the item and the label, if the item and the label fail to match, an optimization problem of the item and the label is constructed, and the goal of the optimization problem is to reallocate the item and the label;

[0116] Furthermore, the optimization problem is regarded as a quantum annealing problem. By collecting the pairs of items and labels that fail to match, the feature similarity between the items and labels that fail to match is established, and binary variables are defined;

[0117] It should be explained that quantum annealing is an optimization algorithm based on quantum mechanics, which is used to solve combinatorial optimization problems and can search for the global optimal solution by simulating the ground state of the quantum system;

[0118] Specifically, the items and label pairs that failed to match are collected as follows:

[0119] Fg = {g 1 ,g 2 ,…,g i}

[0120] Fl={l 1 ,l 2 ,…,l i}

[0121] Among them, Fg represents the set of items that failed to match, Fl represents the set of tags that failed to match; g represents items, l represents tags;

[0122] Specifically, the feature similarity between the failed matching items and the labels is established as: S(Fg, Fl);

[0123] Specifically, the binary variables defined are: z i ; If z i =1, put item g i Assigned to label l i , otherwise no allocation;

[0124] Furthermore, through the similarity values ​​and binary variables between the failed matching items and labels, an optimization objective function of the total cost of failed matching is constructed, and the constraints of the optimization objective function are given;

[0125] Specifically, the optimization objective function is expressed as:

[0126]

[0127] Among them, 1-S(Fg, Fl) represents the degree of matching failure (the lower the matching value, the higher the cost);

[0128] Specifically, the constraint rules are: each item can only be assigned to one label, and each label can only be assigned to one item;

[0129] Furthermore, the optimization objective function and constraints are converted into the Hamiltonian of the quantum system to simulate the annealing process;

[0130] Specifically, the Hamiltonian of the quantum system is expressed as:

[0131]

[0132] Among them, H total represents the total energy state of a quantum system;

[0133] Furthermore, the simulated annealing process is as follows:

[0134] The quantum system is set to start in a high-energy state, that is, a matching scheme that randomly assigns items to labels, and each binary variable is initialized to a random value;

[0135] The quantum system is set to gradually reduce the Hamiltonian of the quantum system, and the global optimal solution is searched through the quantum tunneling effect. When the quantum system converges to a low energy state (that is, when the total energy state reaches the lowest), the matching solution of the currently allocated items and labels is output as the optimal solution.

[0136] It should be noted that by allowing quantum systems to “travel” through local energy barriers through the quantum tunneling effect, lower energy states can be reached more quickly;

[0137] According to the optimization plan, the matching value between the object and the label is recalculated, and the optimized matching value is marked as the matching result of the final object and label information verification.

[0138] Furthermore, this embodiment also provides a rapid labeling system for checking item and label information, including:

[0139] An object and tag information feature extraction module is configured to extract key information from the original information of the object and the tag, process the key information, convert it into a plurality of computable features, and combine the plurality of computable features into two feature vectors;

[0140] The object and label quantum feature calculation module is configured to perform normalization processing on the feature vector, convert the normalized feature vector into a quantum state, perform similarity calculation on the quantum state, output a matching value according to the similarity calculation result, and generate a matching result of the object and label verification using the matching value;

[0141] The item and label matching optimization module is configured to select whether to construct an optimization problem for item and label verification based on the matching result of the item and label verification. If the optimization problem for item and label verification is constructed, the optimization problem is solved and the matching result of the item and label verification is readjusted. Otherwise, the matching result is directly output to the verification result to complete the rapid labeling of item and label information.

[0142] This embodiment also provides a computer device, which is applicable to the case of a rapid labeling method for checking information of an item and a label, and includes:

[0143] Memory and processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the rapid labeling method for verifying item and label information as proposed in the above embodiment.

[0144] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0145] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the fast labeling method for checking the information of an object and a label as proposed in the above embodiment is implemented.

[0146] The storage medium proposed in this embodiment and the data storage method proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be found in the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0147] Example 2

[0148] Reference Figure 2 and Figure 3 , which is the second embodiment of the present invention, and this embodiment provides a rapid labeling method for checking information of articles and labels, including: In order to verify the performance of the "rapid labeling method for checking information of articles and labels" of the present invention, the experiment is carried out in a simulated logistics system, and the specific experimental environment, parameter settings and processes are as follows:

[0149] Experimental environment: CPU: Intel Xeon Gold 6230 (2.10GHz, 20 cores); GPU: NVIDIA A100Tensor Core (40GB); Memory: 256GB; Quantum computing simulation environment: D-Wave Ocean SDK 5.4; System: Ubuntu 20.04; Data processing tools: Python 3.8, NumPy, Pandas; Quantum annealing solver: D-Wave QUBOSolver;

[0150] Data source: simulated a real logistics system, randomly generated 10,000 pairs of items and tags; item and tag ID: text type (generated using UUID); item size: continuous type (range [10,200] cm); tag information: text description (phrases generated based on a random vocabulary);

[0151] Experimental steps: Map the item and tag ID to a unique value through SHA-256 hash; use equal-width binning method for item size features to divide the numerical data into 10 intervals (each interval is 19 cm wide); use the pre-trained BERT model to convert the text description into a 768-dimensional numerical vector; combine the hash value, size binning result, and text vector into a unified classical feature vector; normalize the feature vector so that its modulus is 1; use the quantum feature mapping function to map the classical feature vector to a quantum state, and calculate the similarity between the item and the tag through the inner product of the quantum state (range: [0,1]); calculate the average similarity and standard deviation from the historical matching data; construct the QUBO problem (quantum optimization objective function) for the item and tag pairs that failed to match; use the D-Wave quantum annealing solver to optimize the allocation scheme and reallocate the failed pairs;

[0152] Comparison method: Traditional method, using a linear matching algorithm based on cosine similarity; The solution of the present invention: Similarity matching and dynamic optimization based on the inner product of quantum states;

[0153] First, through Figure 2The matching accuracy of the quantum method of the present invention is always higher than 97%, showing strong stability, and with the increase of data size (number of groups), the downward trend is very slow; while the matching accuracy of the traditional method is significantly lower than that of the quantum method, and the decline is more significant with the increase of data size, indicating that the traditional method is not adaptable enough to scenarios with large data volumes; Figure 2 The error rate (below) shows that the error rate of the method of the present invention remains at a very low level (<3%), and it still maintains high reliability even in large-scale data. In contrast, the error rate of the traditional method gradually increases, especially when the data scale is large (>8000 groups), the error rate is significantly higher than that of the quantum method; this shows that the scheme of the present invention shows superiority in high-dimensional features and multimodal data; secondly, reference Figure 3 The optimization time of the quantum method of the present invention is always short and the growth rate is slow, while the optimization time of the traditional method is obviously higher and grows more rapidly with the increase of the failure ratio, indicating that the allocation efficiency is obviously affected by the allocation scale; in addition, from Figure 3 The optimization success rate (below) shows that the optimization success rate of the method of the present invention is always maintained at a level close to 100%, and even when the failure ratio reaches 50%, it still shows extremely high robustness; while the optimization success rate of the traditional method decreases significantly with the increase of the failure ratio, especially when the failure ratio is high (>30%), the success rate is less than 85%, showing obvious limitations;

[0154] In summary, it can be proved that the method of the present invention can achieve better robustness and system performance than traditional methods even when the amount of data is large.

[0155] Those skilled in the art will appreciate that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the application can adopt the form of complete hardware embodiments, complete software embodiments, or embodiments in combination with software and hardware. Moreover, the application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program codes. The scheme in the embodiments of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0156] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0157] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0158] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0159] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0160] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A rapid labeling method for checking item and label information, characterized in that: include: Extract key information from original information of the item and the tag, process the key information, convert it into multiple computable features, and combine the multiple computable features into two feature vectors; By normalizing the feature vector, converting the normalized feature vector into a quantum state, performing similarity calculation on the quantum state, outputting a matching value according to the similarity calculation result, and using the matching value to generate a matching result of the object and the label verification; According to the matching result of the item and label verification, choose whether to construct an optimization problem for the item and label verification. If the optimization problem for the item and label verification is constructed, solve the optimization problem and readjust the matching result of the item and label verification. Otherwise, directly output the matching result to the verification result to complete the rapid labeling of the item and label information.

2. The rapid labeling method for checking article and label information as claimed in claim 1, characterized in that: The key information is processed and converted into multiple computable features, including: For the unique identifiers in the key information, i.e., the IDs of the items and tags, a standard hash mapping method is used to convert the IDs of the items and tags into corresponding unique values; For the numerical information contained in the key information, the equal-width binning method is used to divide the numerical information into several equal-number intervals. Each interval contains the same number of data points. The result of the equal-width binning is a numerical feature. For the text information in the key information, a word vector model is used to convert the text information into a numerical vector.

3. The rapid labeling method for checking article and label information as claimed in claim 2, characterized in that: Combining the plurality of computable features into two feature vectors includes: The unique value converted from ID, the result of equal-width binning, and the numerical vector converted from text information are combined into two feature vectors; The two feature vectors are represented as feature vectors of the object and the label respectively, and the feature vectors of the object and the label are taken as classical vectors, and their dimensions are set to be consistent.

4. The rapid labeling method for checking article and label information as claimed in claim 3, characterized in that: By normalizing the eigenvector, converting the normalized eigenvector into a quantum state, including: Normalizing each numerical vector in the two eigenvectors, and setting the modulus of each numerical vector to 1; Using the characteristics of classical vectors, quantum feature mapping is performed on each numerical vector in the feature vectors of objects and tags to obtain a feature mapping function; Based on the feature mapping function, quantum states of feature vectors of objects and tags are generated for classical vectors respectively.

5. The rapid labeling method for checking article and label information as claimed in claim 4, characterized in that: Performing similarity calculation on the quantum state, outputting a matching value according to the similarity calculation result, and using the matching value to generate a matching result of the object and the tag verification, including: Calculate the inner product of the quantum state of the object and the tag, that is, measure the degree of overlap between the quantum state of the object and the quantum state of the tag. If the degree of overlap is infinitely close to 1, it means that there is feature similarity between the object and the tag. If the degree of overlap is infinitely close to 0, it means that there is no feature similarity between the object and the tag. Output the matching value based on the feature similarity between the item and the label; Collect historical feature similarities between items and tags, count the feature similarity distribution between all items and tags, obtain the average value and fluctuation range of feature similarity, and use the average value and fluctuation range of feature similarity as the matching threshold; If the outputted matching value exceeds the matching threshold, it is considered that the object and the tag are matched successfully, otherwise it is considered that the object and the tag are matched unsuccessfully.

6. The rapid labeling method for checking article and label information as claimed in claim 5, characterized in that: According to the matching result of the item and the label verification, it is selected whether to construct an optimization problem for the item and the label verification. If the optimization problem for the item and the label verification is constructed, the optimization problem is solved and the matching result of the item and the label verification is readjusted. Otherwise, the matching result is directly output to the verification result, including: If the item and the label match fails, an optimization problem for checking the item and the label is constructed, wherein the goal of the optimization problem is to reallocate the item and the label; The optimization problem is regarded as a quantum annealing problem, by collecting pairs of items and labels that fail to match, establishing feature similarities between the items and labels that fail to match, and defining binary variables; By using the similarity values ​​and binary variables between the failed matching items and the labels, an optimization objective function of the total cost of failed matching is constructed, and constraints of the optimization objective function are given; The optimization objective function and the constraint conditions are converted into the Hamiltonian of the quantum system to simulate the annealing process.

7. The rapid labeling method for checking article and label information according to claim 6, characterized in that: The simulated annealing process comprises: The quantum system is set to start in a high-energy state, that is, a matching scheme that randomly assigns items to labels, and each binary variable is initialized to a random value; The quantum system is set to gradually reduce the Hamiltonian of the quantum system, and the global optimal solution is searched through the quantum tunneling effect. When the quantum system converges to a low-energy state, the matching solution of the currently allocated items and labels is output as the optimal solution. According to the optimization scheme, the matching value between the object and the label is recalculated, and the optimized matching value is marked as the matching result of the final object and label information verification.

8. A rapid labeling system for checking item and label information, based on the rapid labeling method for checking item and label information according to any one of claims 1 to 7, characterized in that: include: An object and tag information feature extraction module is configured to extract key information from the original information of the object and the tag, process the key information, convert it into a plurality of computable features, and combine the plurality of computable features into two feature vectors; The object and label quantum feature calculation module is configured to perform normalization processing on the feature vector, convert the normalized feature vector into a quantum state, perform similarity calculation on the quantum state, output a matching value according to the similarity calculation result, and generate a matching result of the object and label verification using the matching value; The item and label matching optimization module is configured to select whether to construct an optimization problem for item and label verification based on the matching result of the item and label verification. If the optimization problem for item and label verification is constructed, the optimization problem is solved and the matching result of the item and label verification is readjusted. Otherwise, the matching result is directly output to the verification result to complete the rapid labeling of item and label information.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

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 steps of the method according to any one of claims 1 to 7 are implemented.

Images