Content processing method and device, readable medium, electronic equipment and program product

Abstract

A content processing method and apparatus, a readable medium, an electronic device and a program product, the content processing method comprising: acquiring a plurality of first attribute words for describing a target object, the plurality of first attribute words being obtained based on target content associated with the target object; and classifying the plurality of first attribute words at least based on category information of a category to which the target object belongs through a machine learning model to obtain a target attribute word group, and performing induction based on semantics of the first attribute words in the target attribute word group to obtain a target attribute word used for describing the target object. According to the method, the attribute words are subjected to finer-dimensional semantic understanding in combination with the categories to which the objects belong, so that accurate semantic definition and analysis can be performed on the attribute words in different categories of scenes, accurate classification and semantic induction can be conveniently performed on the attribute words, more accurate and more representative attribute words are obtained, and data redundancy is avoided.

Description

Technical Field

[0001] The present disclosure relates to the fields of large language models, machine learning models, and computer technologies. Specifically, it relates to a content processing method, apparatus, readable medium, electronic device, and program product. Background Art

[0002] In related technologies, object analysis can be performed based on the vocabulary extracted from object-associated content. However, during the vocabulary extraction process, there may be problems such as inaccurate extracted vocabulary or a large number of similar words. For example, when extracting attribute words of the content-associated object from the content, a large number of inaccurate or highly similar attribute words appear, causing data redundancy and thus affecting the efficiency and accuracy of subsequent object analysis. Summary of the Invention

[0003] This Summary of the Invention section is provided to introduce concepts in a concise form, which will be described in detail in the subsequent Detailed Description section. This Summary of the Invention section is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0004] In a first aspect, the present disclosure provides a content processing method, the content processing method comprising: obtaining a plurality of first attribute words for describing a target object, the plurality of first attribute words being obtained based on target content associated with the target object; classifying the plurality of first attribute words by a machine learning model at least based on category information of the category to which the target object belongs to obtain a target attribute word group, and generalizing based on the semantics of the first attribute words in the target attribute word group to obtain a target attribute word for describing the target object.

[0005] In a second aspect, the present disclosure provides a content processing apparatus, the content processing apparatus comprising: an obtaining module, configured to obtain a plurality of first attribute words for describing a target object, the plurality of first attribute words being obtained based on target content associated with the target object; a model processing module, configured to classify the plurality of first attribute words by a machine learning model at least based on category information of the category to which the target object belongs to obtain a target attribute word group, and generalizing based on the semantics of the first attribute words in the target attribute word group to obtain a target attribute word for describing the target object.

[0006] In a third aspect, the present disclosure provides a computer-readable medium, on which a computer program is stored, and when the program is executed by a processing device, the steps of the method described in the first aspect are implemented.

[0007] Fourthly, the present disclosure provides an electronic device, including: a storage device storing a computer program thereon; a processing device configured to execute the computer program in the storage device to implement the steps of the method in the first aspect.

[0008] Fifthly, the present disclosure provides a computer program product including a computer program, which when executed by a processor implements the steps of the method in the first aspect.

[0009] Through the above technical solutions, multiple first attribute words can be classified by a machine learning model based on the category information of the target object's category, so that the semantic understanding of the attribute words can be carried out in a finer dimension in combination with the object's category, and thus the semantic definition and analysis of the attribute words in different category scenarios can be accurately carried out, facilitating the accurate classification of the attribute words. Furthermore, induction is performed based on the semantics of the first attribute words in the target attribute phrase to obtain the target attribute words for describing the target object, so as to obtain more accurate and representative attribute words and avoid data redundancy. In the object analysis scenario, the efficiency and accuracy of object analysis can be improved based on the induced attribute words.

[0010] Other features and advantages of the present disclosure will be described in detail in the subsequent specific implementation section. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Combined with the drawings and referring to the following specific implementation manners, the above and other features, advantages and aspects of the embodiments of the present disclosure will become more obvious. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic, and the original and elements are not necessarily drawn to scale. In the drawings: Figure 1 is a flowchart of a content processing method shown according to an exemplary embodiment of the present disclosure; Figure 2 is a schematic diagram of an attribute word recall process shown according to an exemplary embodiment of the present disclosure; Figure 3 is a schematic diagram of a content processing method process shown according to an exemplary embodiment of the present disclosure; Figure 4 is a schematic diagram of a reward model for scoring shown according to an exemplary embodiment of the present disclosure; Figure 5 is a schematic diagram of a preset attribute definition shown according to an exemplary embodiment of the present disclosure; Figure 6 is a schematic diagram of a mental word extraction process shown according to an exemplary embodiment of the present disclosure; Figure 7It is a structural block diagram of a content processing device shown according to an exemplary embodiment of the present disclosure; Figure 8 It is a schematic structural diagram of an electronic device shown according to an exemplary embodiment of the present disclosure. Detailed implementation manners

[0012] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although some embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. On the contrary, these embodiments are provided to more thoroughly and completely understand the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not used to limit the protection scope of the present disclosure.

[0013] It should be understood that the various steps recorded in the method embodiments of the present disclosure can be executed in different orders and / or executed in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this regard.

[0014] As used herein, the term "including" and its variants are open-ended, that is, "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0015] It should be noted that the concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence relationship of the functions executed by these devices, modules or units.

[0016] It should be noted that the modifications of "one" and "multiple" mentioned in the present disclosure are illustrative rather than restrictive. Those skilled in the art should understand that, unless clearly stated otherwise in the context, it should be understood as "one or more".

[0017] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only for illustrative purposes and are not used to limit the scope of these messages or information.

[0018] It can be understood that before using the technical solutions disclosed in the embodiments of the present disclosure, the types, usage scopes, usage scenarios, etc. of the personal information involved in the present disclosure should be informed to users and the authorization of users should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0019] For example, when receiving an active request from a user, a prompt message is sent to the user to clearly prompt the user that the operation requested by the user will require obtaining and using the user's personal information. Thus, the user can autonomously choose whether to provide personal information to software or hardware such as an electronic device, an application program, a server, or a storage medium that performs the operations of the present disclosure technical solution according to the prompt message.

[0020] As an optional but non-limiting implementation manner, the manner of sending a prompt message to the user in response to receiving an active request from the user may be, for example, in the form of a pop-up window, and the prompt message may be presented in text in the pop-up window. In addition, the pop-up window may also carry a selection control for the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0021] It can be understood that the above notification and obtaining user authorization process is only illustrative and does not limit the implementation manner of the present disclosure. Other manners that meet relevant laws and regulations can also be applied to the implementation manner of the present disclosure.

[0022] At the same time, it can be understood that the data involved in the present technical solution (including but not limited to the data itself, the acquisition or use of the data) should comply with the requirements of corresponding laws, regulations and related regulations.

[0023] In the related art, in the case of user authorization, object analysis is performed based on the vocabulary extracted from the object-associated content. However, in the vocabulary extraction process, since different words may have different meanings in different fields, in the face of complex and diverse semantics, it is difficult for the model to capture the meanings of words such as polysemous words and synonyms in different contexts, and thus the extracted attribute words are inaccurate and cannot include the common features of all words. In addition, there may be some attribute words with similar semantics among the obtained attribute words. For example, in the food industry scenario, attribute words with similar meanings such as "gathering / making friends" and "gift / present", for the brand side, these words represent the same attribute, which will instead cause data redundancy and thus affect the efficiency and accuracy of subsequent object analysis.

[0024] In addition, although the large language model can understand certain induction rules and output induction results that conform to the rules, it is easy to ignore the preference differences of users for different output results. For example, for the business logics of different business scenarios, users want to obtain different induction results. For example, taking the group attribute classification in the food field as an example, for the attribute words extracted from "eating melon seeds", "melon seeds", "people who love to eat melon seeds", "eating melon seeds", etc., the large language model will output "melon seeds" as the induced attribute word. However, from the user's perspective, "people who love to eat melon seeds" is more in line with the induction result desired by the user, because "people who love to eat melon seeds" can more accurately reflect the attribute characteristics of the group attribute classification.

[0025] In view of this, the present disclosure provides a content processing method, apparatus, readable medium, electronic device, and program product to solve the above technical problems.

[0026] The following further explains the embodiments of the present disclosure with reference to the accompanying drawings.

[0027] Figure 1 is a flowchart of a content processing method shown according to an exemplary embodiment of the present disclosure. Referring to Figure 1 , the method may include the following steps: S101: Obtain a plurality of first attribute words for describing a target object, where the plurality of first attribute words are obtained based on target content associated with the target object.

[0028] In this embodiment, the target object may be a brand or a category to which the brand belongs, or it may be a commodity or a live broadcast room. Of course, it may also be other objects, and the embodiments of the present disclosure do not impose any restrictions on this. Among them, the category to which the brand belongs may refer to the industry to which the brand belongs, such as industries like cosmetics, electronics, food, automobiles, or smart homes. The target content may be content in the form of short videos, long videos, graphics, texts, audio, etc. obtained under the authorization of the user. Taking the content associated with the brand as an example, it may include promotional videos, promotional articles, and other promotional content released by the brand side or creators such as anchors and bloggers who cooperate with the brand side on social media platforms, as well as comment content on the promotional content by users, evaluation content after users purchase the brand's products, and search content of users on social media platforms. The embodiments of the present disclosure do not impose any restrictions on this.

[0029] S102: Classify the plurality of first attribute words at least based on the category information of the category to which the target object belongs through a machine learning model to obtain a target attribute word group, and summarize based on the semantics of the first attribute words in the target attribute word group to obtain a target attribute word for describing the target object.

[0030] In this embodiment, when the target object is a brand, the target attribute words of the target object can be a summary of various description contents of the brand or product from the perspective of the brand side, such as the product functions of the brand, the origin of the product, the certification information of the brand, etc., or a summary of various description contents of the brand or product from the perspective of the user, such as brand impression, product evaluation, etc. The target attribute words can also be called the mental words of brand mind. Among them, brand mind is used to represent the cognition and impression formed by consumers on the brand, and the mental words are the concrete expressions of brand mind, which are the key information that the brand hopes consumers will remember, covering aspects such as the personality characteristics of the brand, product advantages, and brand style. For example, the first attribute words can be the original attribute words directly extracted from the target content, such as "0 sugar", "sugar-free", "no added sugar", and the target attribute words are the attribute words obtained by summarizing the first attribute words and capable of covering the semantic meaning of the first attribute words. For example, the target attribute word obtained by summarizing the above first attribute words can be "zero sugar", etc. The embodiments of the present disclosure do not make any restrictions on this.

[0031] Correspondingly, when the target object is an industry, the target attribute words of the target object can be a summary of various description contents of the brands in this industry for the brand or product, or a summary of various description contents of the brands in this industry for the brand or product by users, and can also be called the mental words of industry mind.

[0032] Subsequently, based on brand mind or industry mind, it can help the brand side to more comprehensively and accurately understand the cognitive information of consumers on the brand or industry, and then optimize the promotion content and improve the content promotion efficiency.

[0033] By using the above method, the machine learning model can classify multiple first attribute words based on the category information of the target object category, so as to be able to conduct a more detailed semantic understanding of the attribute words in combination with the object category, and thus be able to accurately define and analyze the semantic meaning of the attribute words in different category scenarios, facilitating the accurate classification of the attribute words. Then, based on the semantic meaning of the first attribute words in the target attribute word group, a summary is made to obtain the target attribute words for describing the target object, so as to obtain more accurate and representative attribute words and avoid data redundancy. In the object analysis scenario, based on the summarized attribute words, the efficiency and accuracy of object analysis can be improved.

[0034] To facilitate the understanding of the content processing method provided by the present disclosure, the possible implementation manners in the present disclosure are described below.

[0035] In a possible implementation, the method further includes: for each first attribute word, performing a recall in an attribute word library based on the first attribute word to obtain candidate attribute words, and performing a consistency check on the candidate attribute words and the first attribute word, where the consistency check is used to verify whether the semantics of the candidate attribute words and the first attribute word are consistent. Classifying a plurality of first attribute words through a machine learning model based at least on the category information of the target object to which the target object belongs to obtain a target attribute word group, and generalizing based on the semantics of the first attribute words in the target attribute word group to obtain a target attribute word for describing the target object, including: classifying all the first attribute words that fail the consistency check through a machine learning model based at least on the category information of the target object to which the target object belongs to obtain a target attribute word group, and generalizing based on the semantics of the first attribute words in the target attribute word group to obtain a target attribute word for describing the target object.

[0036] In this embodiment, a content acquisition task can be set to periodically batch acquire content related to the target object from a social media platform. Then, attribute words for describing the target object can be extracted from the content through a large language model. However, the extracted attribute words may have already been acquired and the target attribute words have been extracted in previous content acquisition tasks. To avoid redundant generalization processing of the attribute words, attribute word recall can be performed before processing.

[0037] Exemplarily, as Figure 2 shown, the attribute word library is used to store the target attribute words obtained by previous generalization. For each original attribute word, performing an attribute word recall in the attribute word library to obtain candidate attribute words, and then the original attribute word and the candidate attribute words can be subjected to a consistency check based on a consistency check model. If the semantics are inconsistent, the consistency check fails, indicating that the attribute word library does not include attribute words with similar or the same semantics as the original attribute word. The original attribute words that fail the consistency check can be generalized to obtain new target attribute words, avoiding redundant generalization processing of existing attribute words and wasting resources. Furthermore, the original attribute words that fail the consistency check can be accumulated, and clustering and generalization processing can be performed after obtaining a batch of original attribute words.

[0038] It should be noted that the attribute words in the attribute word library can be stored in vector form, and then the original attribute words can be vectorized to obtain corresponding word vectors before recall. The embodiments of the present disclosure do not impose any restrictions on this.

[0039] Through the above processing link of attribute word recall, while realizing the dynamic update of attribute words, the consistency and measurability of the attribute words in the attribute word library can be ensured, data redundancy can be avoided, and the efficiency of subsequent data processing and analysis can be improved. For example, when a brand party promotes content, it can help the brand party understand the growth of the attribute word association volume and improve business practicability.

[0040] In a possible way, there are multiple candidate attribute words, and each candidate attribute word is associated with a recall score used to characterize the matching degree between the candidate attribute word and the first attribute word. The consistency check of the candidate attribute word and the first attribute word includes: sorting the multiple candidate attribute words in descending order of the recall score, and respectively performing the consistency check on the multiple candidate attribute words and the first attribute word. The method further includes: in response to the candidate attribute word passing the consistency check, determining the candidate attribute word that passes the consistency check as the target attribute word for describing the target object.

[0041] Exemplarily, for the same original attribute word, multiple candidate attribute words may be recalled. To improve the efficiency of the consistency check, the multiple candidate attribute words can be sorted in reverse order of the recall score, so that the consistency check can be performed sequentially in order. Thus, the consistency check can be performed starting from the candidate attribute word with the highest matching degree, so as to more quickly determine the candidate attribute word with the highest matching degree with the original attribute word as the target attribute word. And after determining that there is a candidate attribute word that passes the consistency check, the subsequent consistency check process and the induction process can be stopped, improving the data processing efficiency.

[0042] In a possible way, the machine learning model classifies multiple first attribute words at least based on the category information of the category to which the target object belongs to obtain a target attribute word group, and performs induction based on the semantics of the first attribute words in the target attribute word group to obtain the target attribute word for describing the target object, including: the machine learning model performs the following processing on the multiple first attribute words: respectively performing vector processing on the multiple first attribute words based on the category information of the category to which the target object belongs to obtain multiple target word vectors; classifying the multiple first attribute words according to the multiple target word vectors to obtain a target attribute word group; performing induction based on the semantics of the first attribute words in the target attribute word group to obtain the target attribute word for describing the target object.

[0043] Exemplarily, to improve the accuracy of semantic understanding, the machine learning model can utilize the powerful semantic understanding and reasoning ability of the large language model, simulate the reasoning process of humans through the chain-of-thought technology, and perform vector processing on the first attribute word in combination with the category information of the category to which the target object belongs. The obtained word vector can provide richer and finer-grained semantic information compared to the word vector obtained solely based on the attribute word, facilitating the model to more accurately capture the commonalities of multiple attribute words and more clearly demarcate the semantic boundaries.

[0044] Exemplarily, "smooth and tender" can be classified as a texture attribute in the food industry, and thus grouped with texture-related attribute words such as "smooth" into an attribute phrase. In the beauty industry, it can be classified as an efficacy attribute and grouped with efficacy-related attribute words such as "moist and tender" into an attribute phrase, and so on. The present disclosure does not limit this. By combining category information, the semantic understanding of attribute words can be more accurate, and thus the classification and induction of attribute words can be more accurate, resulting in more representative attribute words, such as attribute words that can cover the semantics of all attribute words within an attribute phrase, or attribute words that can cover the semantics of the most attribute words within an attribute phrase.

[0045] In a possible way, a machine learning model classifies multiple first attribute words at least based on the category information of the target object to obtain a target attribute phrase, including: the machine learning model processes the multiple first attribute words as follows: respectively performs vector processing on the multiple first attribute words based on the category information of the target object to obtain multiple first word vectors; classifies the multiple first attribute words according to the multiple first word vectors to obtain an initial attribute phrase; generalizes based on the semantics of the first attribute words in the initial attribute phrase to obtain multiple second attribute words for describing the target object; respectively performs vector processing on the multiple second attribute words based on the category information of the target object to obtain multiple second word vectors; and classifies the multiple second attribute words at least based on the multiple second word vectors to obtain the target attribute phrase.

[0046] Exemplarily, as Figure 3 shown, adding secondary clustering can reduce the repetition of the finally output attribute words. Because after the first clustering and generalization are completed, some attribute words with similar semantics will appear together. By further performing a clustering operation, the attribute words with similar semantics are grouped into one category and generalized to obtain the final target attribute words.

[0047] Exemplarily, as Figure 3 shown, an encoding model can be used to perform vector processing on attribute words to obtain corresponding word vectors, and then calculate the distance between each two word vectors. For example, the Euclidean distance can be calculated using the Cartesian product. The embodiments of the present disclosure do not limit this in any way. Furthermore, an edge can be constructed between two attribute words whose distance is less than a preset distance, so as to perform clustering based on the constructed edges. The attribute words obtained from the first classification and generalization continue to be vector-processed to obtain corresponding word vectors, and then distance calculation, edge construction, and classification and generalization are performed to obtain the final attribute words. The encoding processing and distance calculation of the two clusterings are similar, except that the clustering algorithms and clustering processes are different.

[0048] It should be understood that, as Figure 3As shown, the first threshold in the first clustering can be greater than the second threshold in the second clustering, indicating that during the second clustering process, the distance between the attribute words that can be clustered is closer, which means the semantics are closer, so that the second clustering of the attribute words with closer semantics can be achieved.

[0049] In a possible way, classifying multiple first attribute words according to multiple first word vectors to obtain an initial attribute word group, including: clustering multiple first attribute words based on multiple first word vectors through a first clustering algorithm to obtain an initial attribute word group, and the first clustering algorithm is used to cluster based on the modularity of multiple first word vectors. Classifying multiple second attribute words based on at least multiple second word vectors to obtain a target attribute word group, including: clustering the second attribute words based on multiple second word vectors through a second clustering algorithm to obtain a target attribute word group, and the second clustering algorithm is used to cluster based on the connected components of multiple second word vectors.

[0050] Exemplarily, as Figure 3 shown, during the first clustering, an algorithm that can cluster based on the modularity of multiple first word vectors can be used, such as the LP (Label Propagation) unsupervised clustering algorithm. The LP unsupervised clustering algorithm is suitable for clustering large-scale similar data. During the second clustering, an algorithm that can cluster based on the connected components of multiple second word vectors can be used, such as the connected components unsupervised algorithm. The connected components unsupervised algorithm can further cluster high-similarity data based on the connected components, and can be specifically selected according to requirements. The embodiments of the present disclosure do not make any restrictions on this.

[0051] It should be understood that first clustering based on the LP unsupervised clustering algorithm can quickly cluster the first attribute words that are large-scale and have similarity. The data volume of the second clustering is smaller than that of the first clustering, but the similarity between the second attribute words is higher. If clustering continues based on the similarity between the attribute words, the effect is not good. Therefore, secondary clustering can be performed based on the connected components, so that the attribute words can be clustered and summarized based on different dimensions, improving the accuracy of attribute word summarization.

[0052] In a possible way, classifying multiple second attribute words based on at least multiple second word vectors to obtain a target attribute word group, including: determining the semantic similarity between every two second attribute words through a similarity verification model, and the similarity verification model is used to determine the semantic similarity between the input attribute words; classifying multiple second attribute words according to multiple second word vectors and the semantic similarity between every two second attribute words to obtain a target attribute word group.

[0053] Exemplarily, the similarity verification model can be implemented based on a trained large language model. For example, the model can be trained with sample attribute words marked with semantic similarity. The embodiments of the present disclosure do not impose any restrictions on this.

[0054] It should be understood that simply measuring the semantic distance between two attribute words based on the vector distance between them may not conform to the semantic distance in the business scenario. For example, "good complexion" and "nice complexion" may have a relatively large distance in vector measurement, but in business, these two words can be considered as the same attribute word.

[0055] Exemplarily, as Figure 3 shown, during the second clustering, the semantic similarity between every two second attribute words can be determined based on the similarity verification model. Then, classification and induction can be carried out by combining the semantic similarity and the word vector distance to obtain the target attribute word. For example, "good complexion" and "nice complexion" may have a relatively large distance based solely on the word vector distance. After adding the similarity verification, their semantic similarity is high, so the distance between them can be reduced, and they can finally be clustered into the same attribute word group, thereby reducing the phenomenon of duplicate attribute words.

[0056] The above clustering process can be implemented based on a trained clustering model. The embodiments of the present disclosure do not impose any restrictions on this.

[0057] In a possible way, the machine learning model includes a large language model. Induction is carried out based on the semantics of the first attribute words in the target attribute word group to obtain the target attribute word for describing the target object, including: using the large language model to perform induction based on the semantics of the first attribute words in the attribute word group to obtain the target attribute word for describing the target object; wherein, the training process of the large language model includes: obtaining sample attribute word groups; using the initial large language model to perform induction based on the semantics of the first sample attribute words in the sample attribute word groups to obtain second sample attribute words; using the trained reward model to score the second sample attribute words to obtain the target score; using the sample attribute word groups, the second sample attribute words, and the target score to train the initial large language model, and obtaining the trained large language model when the preset training completion condition is met.

[0058] Exemplarily, the sample attribute word groups can be obtained by clustering multiple sample attribute words after vector processing by the above encoding model, or multiple similar sample attribute words can be manually processed as a sample attribute word group. The number of sample attribute word groups can be multiple. The embodiments of the present disclosure do not impose any restrictions on this.

[0059] In this embodiment, the above large language model is used to perform semantic induction on the attribute words within the attribute phrase group. The sample attribute phrase group is input into the initial large language model to obtain the second sample attribute word obtained by the initial large language model through semantic induction. Then, the trained reward model is used to score the second sample attribute word. Then, the model training continues by combining the sample attribute phrase group, the second sample attribute word, and the target score until the preset training completion condition is met, and a trained large language model is obtained. The preset training completion condition can be set according to requirements. For example, it can be set as an accuracy threshold or the number of training rounds, etc. The embodiments of the present disclosure do not impose any restrictions on this.

[0060] By introducing a reward model to score the attribute words induced by the large language model, compared with the method of manual scoring, it can quickly score the attribute words induced by the large language model, so as to generate a large-scale training data set for the large language model and improve the model optimization efficiency.

[0061] The above process can be implemented by an induction model. Its core task is to group words with similar semantics through clustering analysis and semantic understanding, and extract a core word that can best summarize the common features of the elements within the word package. This process is similar to finding the "greatest common divisor". For example, an attribute word that can cover the semantics of all attribute words within the attribute phrase group, or an attribute word that can cover the semantics of the most attribute words within the attribute phrase group is induced.

[0062] It should be understood that the induction model usually combines natural language processing technologies and large language models, such as clustering algorithms, word vector representations, and semantic analysis, to identify the similarities and correlations between words. To improve accuracy, the induction model can utilize the powerful semantic understanding and reasoning capabilities of the large language model to simulate the reasoning process of humans through the chain of thought, that is, the above process of semantic understanding by combining classification information, and optimize the induction model through direct preference optimization based on the reward model to make it more in line with human preferences and business requirements.

[0063] In a possible way, the trained reward model is used to score the second sample attribute word to obtain the target score, including: determining the sample classification information corresponding to the sample attribute phrase group, where the sample classification information is determined based on the preset attribute definition and the first sample attribute word in the sample attribute phrase group. The preset attribute definition is obtained by defining the attributes under each classification, and the attributes under each classification are obtained by classifying the attributes of the category to which the attribute description object belongs; the reward model is used to score the second sample attribute word based on the semantic matching degree between the sample classification information and the second sample attribute word to obtain the target score, and the target score is positively correlated with the semantic matching degree.

[0064] Exemplarily, such as Figure 4As shown, the reward model can be a large language model. Through model training, the reward model can score the induced attribute words based on user preferences or actual business logic, and then continue to train the induction model based on the scores, so that the induction results of the induction model increasingly conform to the diverse needs of user preferences and actual business logic, achieving direct preference optimization of the induction model and improving the accuracy of attribute word induction.

[0065] Exemplarily, as Figure 4 shown, the reward model can be instructed by prompt words to score the second sample attribute word based on the semantic matching degree between the sample classification information of the sample attribute phrase and the second sample attribute word. For example, for the group attribute classification in the food industry, assuming the sample attribute phrase includes "eating melon seeds", "melon seeds", "people who love eating melon seeds", "eating melon seeds", compared with the attribute word induced for "melon seeds", the attribute word induced for "people who love eating melon seeds" better conforms to the definition of group attribute classification, so the score is higher.

[0066] Direct preference optimization of the induction model based on the reward model can improve the acceptance of the induced attribute words at the business level and reduce their understanding difficulty, thereby improving the accuracy of attribute word induction.

[0067] In this embodiment, preset attribute definitions can be pre-defined for attributes in different category scenarios. For example, different attribute classifications can be set for different industries such as beauty, food, electronic products, and automobiles. For example, the beauty industry can involve product functions and ingredients, the food industry can involve flavors and textures, and automobiles can involve vehicle functions and performance, etc. The embodiments of the present disclosure do not impose any restrictions on this.

[0068] In a possible way, the classifications in the preset attribute definition include at least one of a first classification, a second classification, and a third classification; the first classification includes a first subclass for classifying object attributes for describing a first object and / or a second subclass for classifying product attributes associated with the first object; the second classification includes at least one of a third subclass for classifying characteristic attributes for generally describing the first object, a fourth subclass for classifying functional attributes of a product associated with the first object, and a fifth subclass for classifying feedback attributes of a user on a product associated with the first object; the third classification includes at least one of a sixth subclass for classifying sub-characteristic attributes under the characteristic attributes, a seventh subclass for classifying sub-functional attributes under the functional attributes, and an eighth subclass for classifying sub-feedback attributes under the feedback attributes.

[0069] In this embodiment, the first classification, the second classification, and the third classification in the preset attribute definition can be configured according to requirements. For example, one-level classification can be configured, or multi-level classification can be configured. Specifically, it can be configured according to requirements, and the embodiments of the present disclosure do not impose any restrictions on this.

[0070] Exemplarily, as Figure 5 shown, taking the extraction of brand mind in the automotive industry as an example, the first subclass in the first classification can be the attribute classification for the brand dimension, and the second subclass can be the attribute classification for the product dimension. The third subclass in the second classification can be the attribute classification for brand characteristics under the brand dimension, the fourth subclass can be the attribute classification for product functions under the product dimension, and the fifth subclass can be the attribute classification for user feedback under the product dimension. The first classification and the second classification can be general classifications across the industry, while the third classification can be set differently for different industries to set corresponding attribute classifications for different industries, improve domain adaptability, and improve the accuracy of attribute word induction.

[0071] Exemplarily, continuing to refer to Figure 5 , the third classification can include attribute classifications for different product functions such as power, endurance, handling, and space of automobiles, can also include attribute classifications for different user feedback such as user evaluations, product pronouns, and usage scenarios, and can also include attribute classifications for different brand characteristics such as brand identity recognition, brand certification, brand style, and brand story. Specifically, it can be configured according to requirements, and the embodiments of the present disclosure do not impose any restrictions on this.

[0072] In this implementation, as Figure 6 shown, the preset attribute definition can be input into the first large language model to semantically understand the content from different sources based on the preset attribute definition, extract the first attribute words that meet the attribute definition, and then the similar attribute words can be clustered through a clustering model to obtain attribute groups, and the semantic induction of the attribute words within the attribute groups can be performed through the second large language model to obtain the final brand mind.

[0073] By adopting the above method, through the above attribute word recall link, secondary clustering and induction process, and direct preference optimization, more accurate and representative attribute words can be refined. For example, brand mind words can help brand owners clarify the core memory points of the brand in the minds of users, optimize content creation to help brand owners conduct more accurate, efficient, and differentiated brand promotion on social media, understand user needs. For example, the attribute words extracted based on the user's search content can directly reflect user needs, assist brand promotion by improving the content connection under popular search terms, or further analyze the preference and reputation of different products or different contents based on the attribute words to help the brand distinguish negative evaluations and optimize them in a timely manner, and at the same time adjust the product promotion direction and strengthen the advantages and highlights in positive evaluations.

[0074] Based on the same concept, an embodiment of the present disclosure also provides a content processing device, as Figure 7 shown. The content processing device 700 includes: An acquisition module 701, configured to acquire a plurality of first attribute words for describing a target object, where the plurality of first attribute words are obtained based on target content associated with the target object; A model processing module 702, configured to classify the plurality of first attribute words through a machine learning model at least based on category information of a category to which the target object belongs, obtain a target attribute word group, and perform induction based on semantics of first attribute words in the target attribute word group to obtain a target attribute word for describing the target object.

[0075] Optionally, the model processing module 702 is configured to: Perform the following processing on the plurality of first attribute words through the machine learning model: Perform vector processing on the plurality of first attribute words respectively based on category information of a category to which the target object belongs to obtain a plurality of target word vectors; Classify the plurality of first attribute words according to the plurality of target word vectors to obtain a target attribute word group; Perform induction based on semantics of first attribute words in the target attribute word group to obtain a target attribute word for describing the target object.

[0076] Optionally, the model processing module 702 is configured to: Perform the following processing on the plurality of first attribute words through a machine learning model: Perform vector processing on the plurality of first attribute words respectively based on category information of a category to which the target object belongs to obtain a plurality of first word vectors; Classify the plurality of first attribute words according to the plurality of first word vectors to obtain an initial attribute word group; Perform induction based on semantics of first attribute words in the initial attribute word group to obtain a plurality of second attribute words for describing the target object; Perform vector processing on the plurality of second attribute words respectively based on category information of a category to which the target object belongs to obtain a plurality of second word vectors; Classify the plurality of second attribute words at least based on the plurality of second word vectors to obtain a target attribute word group.

[0077] Optionally, the model processing module 702 is configured to: Cluster the multiple first attribute words based on the multiple first word vectors through a first clustering algorithm to obtain the initial attribute phrase group, where the first clustering algorithm is used to cluster based on the modularity of the multiple first word vectors; Cluster the second attribute words based on the multiple second word vectors through a second clustering algorithm to obtain the target attribute phrase group, where the second clustering algorithm is used to cluster based on the connected components of the multiple second word vectors.

[0078] Optionally, the model processing module 702 is used to: Determine the semantic similarity between every two of the second attribute words through a similarity verification model, where the similarity verification model is used to determine the semantic similarity between the input attribute words; Classify the multiple second attribute words according to the multiple second word vectors and the semantic similarity between every two of the second attribute words to obtain the target attribute phrase group.

[0079] Optionally, the machine learning model includes a large language model, and the model processing module 702 is used to: Induce based on the semantics of the first attribute words in the attribute phrase group through the large language model to obtain target attribute words for describing the target object; The content processing device 700 further includes a training module, and the training module is used to: Obtain sample attribute phrase groups; Induce based on the semantics of the first sample attribute words in the sample attribute phrase group through an initial large language model to obtain second sample attribute words; Score the second sample attribute words through a trained reward model to obtain a target score; Train the initial large language model with the sample attribute phrase group, the second sample attribute words, and the target score, and obtain a trained large language model when a preset training completion condition is met.

[0080] Optionally, the training module is used to: Determine the sample classification information corresponding to the sample attribute phrase group, where the sample classification information is determined based on a preset attribute definition and the first sample attribute words in the sample attribute phrase group, the preset attribute definition is obtained by defining the attributes under each classification, and the attributes under each classification are obtained by classifying the attributes of the category to which the attribute description object belongs; Score the second sample attribute words through the reward model based on the semantic matching degree between the sample classification information and the second sample attribute words to obtain the target score, and the target score is positively correlated with the semantic matching degree.

[0081] Optionally, the content processing device 700 further includes a verification module, and the verification module is configured to: For each of the first attribute words, recall candidate attribute words based on the first attribute word in the attribute word library, and perform consistency verification on the candidate attribute words and the first attribute word, where the consistency verification is used to verify whether the semantics of the candidate attribute words and the first attribute word are consistent; The model processing module 702 is configured to: Classify the candidate attribute words that fail the consistency verification at least based on the category information of the category to which the target object belongs through a machine learning model to obtain a target attribute word group, and perform induction based on the semantics of the first attribute word in the target attribute word group to obtain a target attribute word for describing the target object.

[0082] Optionally, there are multiple candidate attribute words, and each candidate attribute word is associated with a recall score used to characterize the matching degree between the candidate attribute word and the first attribute word, and the verification module is configured to: Perform consistency verification on the multiple candidate attribute words and the first attribute word respectively in the order from largest to smallest of the recall scores; The content processing device 700 further includes a determination module, and the determination module is configured to: In response to the candidate attribute word passing the consistency verification, determine the candidate attribute word that passes the consistency verification as the target attribute word for describing the target object.

[0083] Based on the same concept, an embodiment of the present disclosure further provides a computer-readable medium, on which a computer program is stored, and when the program is executed by a processing device, the steps of any of the above content processing methods are implemented.

[0084] Based on the same concept, an embodiment of the present disclosure further provides an electronic device, which may include: A storage device on which a computer program is stored; A processing device configured to execute the computer program in the storage device to implement the steps of any of the above content processing methods.

[0085] Based on the same concept, an embodiment of the present disclosure further provides a computer program product, including a computer program, and when the computer program is executed by a processor, the steps of any of the above content processing methods are implemented.

[0086] Next, refer to Figure 8, which shows a schematic structural diagram of an electronic device 800 suitable for implementing the embodiments of the present disclosure. The terminal devices in the embodiments of the present disclosure may include, but are not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Tablet Computers), PMPs (Portable Multimedia Players), in-vehicle terminals (such as in-vehicle navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 8 The electronic device shown is merely an example and should not impose any limitations on the functions and usage scope of the embodiments of the present disclosure.

[0087] As Figure 8 shown, the electronic device 800 may include a processing device (such as a central processing unit, a graphics processing unit, etc.) 801, which may perform various appropriate actions and processes according to the programs stored in the read-only memory (ROM) 802 or the programs loaded from the storage device 808 into the random access memory (RAM) 803. In the RAM 803, various programs and data required for the operation of the electronic device 800 are also stored. The processing device 801, the ROM 802, and the RAM 803 are connected to each other through a bus 804. The input / output (I / O) interface 805 is also connected to the bus 804.

[0088] Generally, the following devices may be connected to the I / O interface 805: an input device 806 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 807 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 808 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 809. The communication device 809 may allow the electronic device 800 to communicate with other devices wirelessly or wiredly to exchange data. Although Figure 8 shows the electronic device 800 having various devices, it should be understood that it is not required to implement or have all the shown devices. More or fewer devices may be alternatively implemented or had.

[0089] Particularly, according to the embodiments of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, the embodiments of the present disclosure include a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program contains program codes for performing the methods shown in the flowcharts. In such an embodiment, the computer program may be downloaded and installed from the network through the communication device 809, or installed from the storage device 808, or installed from the ROM 802. When the computer program is executed by the processing device 801, the above-mentioned functions defined in the methods of the embodiments of the present disclosure are executed.

[0090] It should be noted that the above-mentioned computer-readable medium in the present disclosure can be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of the computer-readable storage medium can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, the computer-readable storage medium can be any tangible medium that contains or stores a program, and this program can be used by or in combination with an instruction execution system, apparatus, or device. In the present disclosure, the computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer-readable signal medium can also be any computer-readable medium other than the computer-readable storage medium, and this computer-readable signal medium can send, propagate, or transmit a program for use by or in combination with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0091] In some embodiments, communication can be carried out using any currently known or future-developed network protocol such as HTTP (HyperText Transfer Protocol), and can be interconnected with digital data communication in any form or medium (e.g., a communication network). Examples of communication networks include local area networks (“LAN”), wide area networks (“WAN”), the Internet (e.g., the Internet), and end-to-end networks (e.g., ad hoc end-to-end networks), as well as any currently known or future-developed network.

[0092] The above-mentioned computer-readable medium can be included in the above-mentioned electronic device; it can also exist separately and not be assembled into the electronic device.

[0093] The above computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to: obtain a plurality of first attribute words for describing a target object, where the plurality of first attribute words are obtained based on target content associated with the target object; classify the plurality of first attribute words at least based on category information of a category to which the target object belongs through a machine learning model to obtain a target attribute word group, and perform induction based on the semantics of the first attribute words in the target attribute word group to obtain a target attribute word for describing the target object.

[0094] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages or combinations thereof. The programming languages include, but are not limited to, object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, executed as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (for example, by using an Internet service provider to connect through the Internet).

[0095] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code that contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, may be implemented by a dedicated hardware-based system for performing the specified functions or operations, or may be implemented by a combination of dedicated hardware and computer instructions.

[0096] The modules involved in the embodiments described in the present disclosure may be implemented in software or in hardware. Among them, the name of the module does not constitute a limitation to the module itself in some cases.

[0097] The functions described above in this document can be performed, at least in part, by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that can be used include: Field Programmable Gate Arrays (FPGAs), Application Specific Integrated Circuits (ASICs), Application Specific Standard Products (ASSPs), Systems on Chip (SOCs), Complex Programmable Logic Devices (CPLDs), and the like.

[0098] In the context of this disclosure, a machine-readable medium can be a tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a Read-Only Memory (ROM), an Erasable Programmable Read-Only Memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0099] The above description is only a preferred embodiment of the present disclosure and an explanation of the applied technical principles. Those skilled in the art should understand that the scope of the disclosure involved in the present disclosure is not limited to the technical solutions formed by the specific combination of the above technical features, and should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosure concept. For example, the technical solutions formed by mutually replacing the above features with the technical features (but not limited to) having similar functions disclosed in the present disclosure.

[0100] In addition, although the operations are depicted in a particular order, this should not be construed as requiring that the operations be performed in the particular order shown or in sequential order. In certain environments, multitasking and parallel processing may be advantageous. Similarly, although a number of specific implementation details are included in the above discussion, these should not be construed as limiting the scope of the present disclosure. Certain features described in the context of separate embodiments can also be implemented combinatorially in a single embodiment. Conversely, the various features described in the context of a single embodiment can also be implemented separately or in any suitable sub-combination in multiple embodiments.

[0101] Although the subject matter has been described in language specific to structural features and / or methodological logical acts, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. On the contrary, the specific features and acts described above are merely example forms for implementing the claims. Regarding the devices in the above embodiments, the specific manner in which each module performs operations has been described in detail in the embodiments related to the method, and will not be elaborated here.

Claims

1. A content processing method, characterized in that, The content processing method includes: Obtaining a plurality of first attribute words for describing a target object, where the plurality of first attribute words are obtained based on target content associated with the target object; Classifying the plurality of first attribute words at least based on category information of a category to which the target object belongs through a machine learning model to obtain a target attribute word group, and generalizing based on semantics of first attribute words in the target attribute word group to obtain a target attribute word for describing the target object.

2. The content processing method according to claim 1, wherein The step of classifying the plurality of first attribute words at least based on category information of a category to which the target object belongs through a machine learning model to obtain a target attribute word group, and generalizing based on semantics of first attribute words in the target attribute word group to obtain a target attribute word for describing the target object includes: Performing the following processing on the plurality of first attribute words through the machine learning model: Performing vector processing on the plurality of first attribute words respectively based on category information of a category to which the target object belongs to obtain a plurality of target word vectors; Classifying the plurality of first attribute words according to the plurality of target word vectors to obtain a target attribute word group; Generalizing based on semantics of first attribute words in the target attribute word group to obtain a target attribute word for describing the target object.

3. The content processing method according to claim 1, wherein The step of classifying the plurality of first attribute words at least based on category information of a category to which the target object belongs through a machine learning model to obtain a target attribute word group includes: Performing the following processing on the plurality of first attribute words through a machine learning model: Performing vector processing on the plurality of first attribute words respectively based on category information of a category to which the target object belongs to obtain a plurality of first word vectors; Classifying the plurality of first attribute words according to the plurality of first word vectors to obtain an initial attribute word group; Generalizing based on semantics of first attribute words in the initial attribute word group to obtain a plurality of second attribute words for describing the target object; Performing vector processing on the plurality of second attribute words respectively based on category information of a category to which the target object belongs to obtain a plurality of second word vectors; Classifying the plurality of second attribute words at least based on the plurality of second word vectors to obtain a target attribute word group.

4. The content processing method according to claim 3, characterized in that, The step of classifying the plurality of first attribute words according to the plurality of first word vectors to obtain an initial attribute word group includes: Clustering the plurality of first attribute words based on the plurality of first word vectors through a first clustering algorithm to obtain the initial attribute word group, where the first clustering algorithm is used to perform clustering based on modularity of the plurality of first word vectors; The step of classifying the plurality of second attribute words at least based on the plurality of second word vectors to obtain a target attribute word group includes: Clustering the second attribute words based on the plurality of second word vectors through a second clustering algorithm to obtain the target attribute word group, where the second clustering algorithm is used to perform clustering based on connected components of the plurality of second word vectors.

5. The content processing method according to claim 3, wherein The step of classifying the plurality of second attribute words at least based on the plurality of second word vectors to obtain a target attribute word group includes: Determine the semantic similarity between every two of the second attribute words through a similarity verification model, which is used to determine the semantic similarity between the input attribute words; Classify the multiple second attribute words according to the multiple second word vectors and the semantic similarity between every two of the second attribute words, and obtain a target attribute word group.

6. The content processing method according to any one of claims 1-5, characterized in that The machine learning model includes a large language model. The induction based on the semantics of the first attribute words in the target attribute word group to obtain the target attribute words for describing the target object includes: Through the large language model, induce the target attribute words for describing the target object based on the semantics of the first attribute words in the attribute word group; Among them, the training process of the large language model includes: Obtain sample attribute word groups; Through the initial large language model, induce the second sample attribute words based on the semantics of the first sample attribute words in the sample attribute word group; Score the second sample attribute words through the trained reward model to obtain a target score; Use the sample attribute word group, the second sample attribute words, and the target score to train the initial large language model, and obtain a trained large language model when the preset training completion condition is met.

7. The content processing method according to claim 6, characterized in that The scoring the second sample attribute words through the trained reward model to obtain a target score includes: Determine the sample classification information corresponding to the sample attribute word group. The sample classification information is determined based on the preset attribute definition and the first sample attribute words in the sample attribute word group. The preset attribute definition is obtained by defining the attributes in each classification, and the attributes in each classification are obtained by classifying the attributes of the category to which the attribute description object belongs; Score the second sample attribute words through the reward model based on the semantic matching degree between the sample classification information and the second sample attribute words to obtain the target score, and the target score is positively correlated with the semantic matching degree.

8. The content processing method according to any one of claims 1-5, characterized in that The method further includes: For each of the first attribute words, recall in the attribute word library based on the first attribute word to obtain candidate attribute words, and perform consistency verification on the candidate attribute words and the first attribute word. The consistency verification is used to verify whether the semantics of the candidate attribute words and the first attribute word are consistent; The classification of the multiple first attribute words through the machine learning model at least based on the category information of the target object to obtain a target attribute word group, and the induction based on the semantics of the first attribute words in the target attribute word group to obtain the target attribute words for describing the target object includes: Through the machine learning model, classify all the first attribute words that have not passed the consistency verification at least based on the category information of the target object to obtain a target attribute word group, and induce the target attribute words for describing the target object based on the semantics of the first attribute words in the target attribute word group.

9. The content processing method according to claim 8, characterized in that, There are multiple candidate attribute words, and each of the candidate attribute words is associated with a recall score for characterizing the matching degree between the candidate attribute word and the first attribute word. The consistency check of the candidate attribute word and the first attribute word includes: Performing a consistency check on each of the multiple candidate attribute words and the first attribute word in descending order of the recall score; The method further includes: In response to the candidate attribute word passing the consistency check, determining the candidate attribute word that passes the consistency check as the target attribute word for describing the target object.

10. A content processing device, characterized in that, The content processing device includes: An acquisition module for acquiring multiple first attribute words for describing a target object, where the multiple first attribute words are obtained based on the target content associated with the target object; A model processing module for classifying the multiple first attribute words at least based on the category information of the category to which the target object belongs through a machine learning model to obtain a target attribute word group, and performing induction based on the semantics of the first attribute words in the target attribute word group to obtain a target attribute word for describing the target object.

11. A computer-readable medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processing device, the steps of the method according to any one of claims 1-9 are implemented.

12. An electronic device, characterized in that, Including: A storage device on which a computer program is stored; A processing device for executing the computer program in the storage device to implement the steps of the method according to any one of claims 1-9.

13. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, the steps of the method according to any one of claims 1-9 are implemented.

Images