22 results about "Cold start recommendation" patented technology

The invention relates to the technical field of cold start, and discloses an image cold start recommendation sorting method and system, and the method comprises the steps: obtaining the three-dimensional point cloud data and RGB images of a target scene, and constructing a spatial topology layout relation graph, a fine-grained visual feature library and a multi-level feasibility constraint index vector; generating a multi-granularity representation vector based on the feature library, decomposing the multi-granularity representation vector into three types of independent attribute vectors through deentanglement learning, and generating a graph level representation vector; constructing a comprehensive scoring matrix, and outputting a final recommendation sequence through three-wheel progressive screening, multi-target Pareto sorting and visual confusion disambiguation; according to the method, the problems of visual feature semantic drifting and execution constraint mismatching in a cold start scene are effectively solved, and cognitive transition from visual similarity to function matching and then to an executable scheme is achieved.

The invention relates to a cold start recommendation method based on feature enhancement intention decoupling. A system model of the method comprises a feature enhancement module, an intention decoupling representation learning module, a cold start adaptation module and a recommendation prediction unit. The method comprises the following steps: firstly, carrying out adaptive weight learning on original features of a user and an article through a feature enhancement module, and dynamically adjusting the importance of different features; secondly, an improved capsule network is adopted to map the enhanced feature representation to a plurality of independent intention spaces, and fine-grained intention decoupling modeling is achieved; constructing an intention enhancement mechanism based on graph convolution, establishing an independent interaction sub-graph for each intention space, and learning intention-specific collaborative information; and finally, establishing a mapping relation from intention representation based on features to intention representation based on collaborative filtering through a cold start adaptation module, and realizing knowledge migration of a fine-grained intention level by adopting a bidirectional knowledge distillation strategy. According to the method, the problems of insufficient feature utilization, rough intention modeling and the like in traditional cold start recommendation are effectively solved, the recommendation accuracy and individuation degree are remarkably improved, and the method is particularly suitable for strict cold start scenes in which new users and new articles are completely lack of interaction history.

The invention relates to the technical field of cold start recommendation, in particular to a cold start recommendation method and system based on knowledge graph enhanced representation learning. Historical interaction behavior data of a user and an article and a structured knowledge graph are used as input data sources, entity relationship features, semantic attribute features and user interest distinguishing information are fused, and through a multi-stage feature modeling and sample optimization strategy, the problems of insufficient training signals and limited representation learning under a cold start condition are relieved. A structure and feature combined driven difficult negative sample mining mechanism is adopted, so that the sample discrimination is enhanced; in combination with a multi-hop neighbor feature aggregation method of relation specificity gating, semantic contribution weights of different relation types of neighbors to target node representation are effectively improved; and through a positive and negative view feature fusion strategy perceived by the interest boundary, discriminative modeling of the interest area and the non-interest area of the user is realized. And a dual-stage meta-learning optimization strategy is adopted, so that the rapid adaptation and generalization ability of the model to diversified cold start task distribution is improved.

The application provides a cold start recommendation method and a Thompson sampling recommendation method. The cold start recommendation method comprises the following steps: obtaining to-be-recommended information and a characteristic attribute of the to-be-recommended information; performing dimension classification on the characteristic attribute of the to-be-recommended information to obtain characteristic information of at least two dimensions; obtaining known information of a target user, the target user being a user without behavior data within a first preset time; selecting a plurality of recommended information from to-be-recommended information corresponding to each dimension characteristic information based on the association between the known information and the characteristic information of the at least two dimensions, and generating a recommendation pool based on the plurality of recommended information. The application performs information recommendation on the target user in the cold start period from the recommended information corresponding to the plurality of dimension characteristic information, improves the information recommendation hit rate, and enables the target user to quickly pass through the cold start period.

The invention relates to a novel cold start recommendation method and device, electronic equipment and a storage medium. The method comprises the steps of obtaining novel data to be recommended and micro-drama data of a user, wherein the novel data to be recommended and the micro-drama data comprise information of the same type; based on an NLP model, respectively performing deep semantic feature extraction on the micro-drama data and the to-be-recommended novel data, and generating a micro-drama text vector and a novel text vector; based on a pre-trained multi-label model, performing label prediction according to the novel text vector, the micro-episode text vector and the acquired micro-episode label to obtain a novel prediction label; and through a hybrid recommendation engine, obtaining a semantic similarity score of the novel prediction tag and the micro-drama tag and a novel historical behavior score of the user, and generating a novel recommendation list for recommendation according to the semantic similarity score and the novel historical behavior score. According to the invention, the accuracy of novel recommendation can be improved.

The present application belongs to the technical field of cold start of home maintenance recommendation system, and particularly relates to a household appliance maintenance cold start recommendation method based on an improved clustering algorithm and a Catboost model, comprising: fusing obtained community information and user information to form a new user portrait; statistically processing community users to divide the users into user group 1 and user group 2; for the user group 1, an improved clustering algorithm is used for clustering to obtain a similar user group, and the service with the most number of preferred products in the similar user group is recommended; for the user group 2, a Catboost model is used for classification prediction and recommendation according to the new user portrait. The present application recommends in a double-recommendation mode, uses easily-obtained community information as auxiliary information, expands the user dimension, and more information can make the classification model more easily learn the relationship between the characteristics and the classification prediction, thereby strengthening the accuracy of the classification prediction, so as to accurately recommend products for the users.

This invention discloses a cold-start recommendation method and system based on dynamic adaptive fusion of multimodal features. First, a multimodal semantic graph is constructed to extract features. Second, cold-start items are identified by calculating the deviation between the number of item interactions and the average number of item interactions. Third, a neighborhood set of cold-start items is extracted, and its average neighborhood similarity in the multimodal semantic graph is calculated. Next, base weights are calculated based on the number of item interactions, and a dynamic index is calculated by combining the average neighborhood similarity to determine the final weights. Then, the enhanced interaction matrix is ​​symmetrically normalized, and multi-signal extraction is performed on the symmetrically normalized enhanced interaction matrix. The signals are then fused to obtain the final preference probability. Finally, a Top-K recommendation list is generated based on the final preference probability as the final recommendation result. This invention can achieve high-precision, low-latency, and highly reliable personalized recommendations in cold-start scenarios, and is suitable for large-scale real-time recommendation platforms such as e-commerce and content distribution.

The application discloses a recommendation system and method based on meta-learning graph neural network for cold start, and belongs to the technical field of network recommendation. The technical problem to be solved by the application is how to overcome the defect that the cold start user interaction data is sparse and cannot accurately dynamically model user preferences. The technical solution adopted is that the system comprises a two-part graph construction layer, a graph embedding layer, a sequence coding layer and a meta-learning layer. The method is as follows: obtaining interaction data in a data set, constructing a user-item two-part graph using the interaction data; using a graph convolution network based on the two-part graph to construct high-order relationships between items in sequences, and generating accurate embedding representations of users and items; using the sequence coding layer to learn the transfer information of items within the sequence, and combining an attention mechanism to generate a dynamic interest representation for the user; and using the meta-learning layer to quickly adapt to the cold start recommendation task.

This invention relates to the field of cold start technology and discloses an image cold start recommendation and ranking method and system. By acquiring 3D point cloud data and RGB images of the target scene, a spatial topology layout relationship graph, a fine-grained visual feature library, and multi-level feasibility constraint index vectors are constructed. Based on the feature library, multi-granularity representation vectors are generated, which are then decomposed into three independent attribute vectors through de-entanglement learning and graph-level representation vectors are generated. A comprehensive scoring matrix is ​​then constructed, and after three rounds of progressive screening, multi-objective Pareto ranking, and visual confusion disambiguation, the final recommendation sequence is output. This invention effectively solves the problems of visual feature semantic drift and execution constraint mismatch in cold start scenarios, achieving a cognitive leap from "visual similarity" to "functional matching" and then to "executable solutions."

The application discloses a social prompt cold start recommendation method based on an SPGCLRec framework, and comprises the following steps: step one, constructing an interaction graph and pre-training a recommendation model; step two, optimizing a social prompt based on a diffusion model; step three, enhancing a cold start user representation; and step four, generating a cold start recommendation and jointly optimizing a model. The learnable social soft prompt designed by the application embeds social information into a representation space of a pre-training model, and can realize efficient parameter adaptation without fine-tuning core weights of the model, thereby making up for the deficiency of the prior art which only relies on a single auxiliary clue and realizing efficient parameter adaptation of the pre-training model. The diffusion model is used to filter noise social relationships in a latent vector space, thereby improving user representation accuracy. By introducing a contrast learning mechanism and introducing Gaussian noise in a latent space, the semantic consistency of the optimized social prompt and the item representation is strengthened, and the robustness of the user representation is improved.

The invention relates to the technical field of recommendation systems and multi-modal representation learning, in particular to a multi-modal recommendation semantic token modeling method based on residual vector quantization. The method aims at solving the problem that in a cold start scene, user-article interaction is sparse, multi-modal feature noise is large, alignment is difficult, and the recommendation effect is poor. According to the method, user behaviors, article images and text features are obtained, multi-modal features are mapped to a unified dimension and then input into a multi-level residual vector quantization self-encoding module, discrete semantic tokens are learned, collaborative filtering representation is obtained in combination with a user-article bigraph, fusion is carried out in a representation layer, and a user-article image is obtained. The user-article matching score is jointly optimized through the sorting-based loss function, the method is suitable for multi-modal recommendation systems such as e-commerce and content platforms, the accuracy and robustness of cold start recommendation can be improved, and the online calculation overhead can be reduced.

The invention belongs to the field of artificial intelligence recommendation systems, and particularly relates to a multi-scene recommendation method based on heterogeneous knowledge embedding fusion. According to the method, firstly, feature representation is carried out on user-article interaction data, meanwhile, modeling is carried out on entities and relationships in a knowledge graph by adopting the knowledge graph, and embedded representation of users, articles, entities and relationships is uniformly represented in a vector form, so that multi-scene recommendation is realized. According to the method, corresponding recommendation strategies are designed in three typical scenes of common recommendation, cold start recommendation and data sparse recommendation, so that the adaptability and generalization ability of a recommendation system are remarkably improved. Compared with an existing recommendation method which only depends on historical interaction records of the user, the method has the advantages that potential semantic information can be supplemented by utilizing the knowledge graph under the condition that user interaction data is insufficient or objects are not exposed, so that the cold start problem is relieved; under the condition of large-scale sparse interaction, the coverage and diversity of recommendation results are improved through a similarity modeling and knowledge completion mechanism.

The application discloses a user cold start recommendation system and method based on similarity matching, constructs a knowledge graph database according to users, products and the interaction relationship between the users and the products, and carries out standardization processing. The processed data is extracted to score information and mapped to a user like-dislike type space, so that the like-dislike product type of each user is determined, and a similarity matching algorithm is designed to find the old user with the highest similarity to the new user. The old user product type score vector and the new user product type vector score are processed by an association function and used as the recommendation basis of the system, the reasonable expansion of the new user data is completed, and the cold start problem of the recommendation system is effectively alleviated. The application solves the user cold start problem caused by the introduction of a new user into an industrial recommendation system. The application can be used for the scene of effectively recommending the product types that meet the preferences of new users in the industrial recommendation system.

The invention discloses a recommendation cold start enhancement method based on large language model semantic interaction knowledge distillation. The recommendation cold start enhancement method is characterized by comprising the steps that a closed source large language GPT-4o model is used for constructing a semantic interaction graph between articles in an off-line mode; designing dual-channel comparison learning loss, and distilling structured semantic relationship knowledge to a lightweight sequence recommendation model; and based on the sequence recommendation model, searching an article with a high attention score in the sequence, replacing the article with a cold start article according to the semantic similarity, judging whether a certain article is the next interaction object of the user or not through a fine-tuned Tonyuqian 7B model, and generating supplementary training data for secondary training of the sequence recommendation model. Compared with the prior art, the problems of semantic drift, logic fracture and situation mismatch in cold start recommendation are effectively relieved, recommendation precision and interpretability in new users, new articles and new fields are improved, training efficiency and online reasoning compatibility are considered, and the method is suitable for multi-scene cold start recommendation tasks of e-commerce, social contact, content platforms and the like.

The invention discloses a user and article semantic feature generation method for cold start recommendation, and belongs to the technical field of recommendation systems for semantic modeling based on a large language model. According to the method, after preprocessing operation of data extraction, data cleaning, user behavior sequence construction and validity check, article basic label generation and user basic label generation are realized through online LLM; based on the article basic label dictionary and the user basic label dictionary, performing fine tuning by constructing a training data set and using a LoRA method, setting training parameters and a loss function to perform training circulation, and outputting a local LLM after fine tuning; by generating the article multi-role advanced semantic tag and generating the user multi-role cognitive advanced semantic tag and outputting the article multi-role advanced semantic tag and the user multi-role cognitive advanced semantic tag, the cold start problem in the recommendation system is solved through the application of the method in the recommendation system, and the recommendation efficiency is improved. And particularly, the cold start problem of sparse user interaction data is solved.

The invention discloses a recommendation method and system, and the method comprises the steps: S1, obtaining unbiased features and auxiliary features of a user, carrying out the embedding processing of the unbiased features and the auxiliary features, and obtaining feature embedding and other feature embedding; s2, clustering all users in groups through a clustering module; s3, based on the distance from the user to each cluster, distributing a target cluster for the user, and obtaining a group representation of the target cluster; S4, generating a complete user representation containing a clustering center and a personalized offset feature; s5, inputting the complete user representation into a main recommendation model, and outputting a recommendation result; and S6, constructing inter-cluster correlation regular loss and main model basic loss, and completing cold start recommendation of the low-activity users. According to the method, grouping clustering of the users is firstly completed through unbiased features, interest migration from high-activity users to low-activity users is achieved, and therefore cold start recommendation of the low-activity users is achieved.

The application discloses a user cold start recommendation method based on individual and group adaptability meta-learning, comprising: inputting item interaction information into a related user identification component and a prior-based interest extraction component respectively; identifying related users showing the same interest for a specified item through the related user identification component, identifying related users with similar preferences and sharing similar knowledge by using a meta-path in a heterogeneous information network, obtaining prior knowledge corresponding to the found group, and realizing user group level adaptation; extracting the correlation relationship between different interests of the user and historical task knowledge through the prior-based interest extraction component, obtaining individual prior knowledge, and realizing user individual level adaptation; and guiding the global shared prior knowledge to adapt to the knowledge of the user by constructing an adapter for the user. The application has the advantages of simple implementation method, good generalization effect, high recommendation efficiency and precision, etc.

The invention provides a cold start recommendation model training method and device and a recommendation degree obtaining method and device, and relates to the technical field of computers, in particular to the technical fields of machine learning, intelligent search, intelligent recommendation, information flow and the like. The specific implementation scheme comprises the following steps: selecting a first target user from a plurality of cold start users of a target application; obtaining a first user side feature of the first target user; obtaining a first resource side feature of the first target resource; and training the first initial recommendation model based on the first user side feature and the first resource side feature to obtain a cold start recommendation model. By adopting the method and the device, after the cold start recommendation model is obtained, when the cold start recommendation model is applied to the resource recommendation scene of the start user, the recommendation degree for the candidate resources can be accurately obtained, so that the matching degree of the resource content distributed to the cold start user and the real interest of the resource content is improved, and the experience feeling of the cold start user is remarkably improved.

The application discloses an item cold start recommendation method based on a knowledge graph and meta learning, first, according to a user-item bipartite graph, user and item initialization embeddings are generated, and a meta embedding is constructed according to a self-attention encoder; secondly, in first-order propagation, for the problem of incomplete cold start of items, a meta aggregator is used to aggregate target items and neighbors to reconstruct the embedding of the items; finally, the probability of the user taking the target item is obtained through inner product, and the first N items are taken as the final recommendation result. The application combines the advantages of the knowledge graph and meta learning, integrates them into one, and more perfectly solves the item recommendation cold start problem, not only reduces noise and improves the recommendation result in the cold start scene through meta learning, but also explores hidden factors of user interaction with these items through the method of combining the knowledge graph, and digs out the potential interest of the user.

The application discloses a cold start recommendation method and system based on balanced sampling and embedded optimization meta-learning, and the method comprises the following steps: pre-processing interactive data; constructing a support set and a query set based on a score balanced sampling strategy; extracting seven-dimensional implicit task features to depict user preference patterns, including centralized mean, normalized standard deviation, truncated skewness, score range proportion, low score proportion, high score proportion and user activity; constructing an embedded optimization module, generating a task decision signal through a multi-source feature adapter and a funnel type fusion decoder, and introducing affine modulation dynamic correction of user and item original embedding; designing an adaptive deep recommendation network, and adopting a double-channel Dropout strategy to differentiate the adaptation of inner loop fitting and the generalization of outer loop. The application dynamically guides the feature embedding optimization based on the score distribution characteristics of user implicit statistics, realizes the adaptive correction of task preference, and improves the accuracy and robustness of recommendation in the cold start situation.

The application discloses a kind of multimedia-oriented generative generalization cold start recommendation method, comprising:1. the interactive record of user and product is used to construct score matrix;2. input layer is constructed by one-hot encoding mode, and user, product is mapped to different embedding space by combining the multimedia features of product;3. the embedding of user, product is optimized by bayesian ranking loss function;4. construct generative neural network, including: prior neural network, encoder, decoder;5. construct uniformity enhanced conditional variational autoencoder, to make the latent space of original variational autoencoder more uniform and distinguishable;6. new product embedding generation method based on clustering obtains new product generative embedding;7. the degree of love of user to new product is predicted by the way of vector dot product.The application can make recommendations for new products using multimedia information of products, thereby alleviating the cold start problem in the recommendation system.

The application discloses a multimedia generalization recommendation method based on invariant learning, comprising the following steps: 1. constructing heterogeneous data; 2. constructing a cold start recommendation network, including a representation generator, an embedding layer and a user collaborative representation matrix; 3. the representation generator processes multimedia original features to obtain representation vectors of each mode in a product set; 4. initializing a user collaborative matrix and obtaining a final representation vector of a user through the embedding layer; 5. constructing a weight set of mode fusion in M different environments by a cyclic mixing layer, which is used for generating a representation differentiated environment and optimizing an invariant representation; 6. calculating an invariant loss function according to a recommendation loss in different environments; 7. combining each loss function and performing multi-task learning on a recommended product network to update network parameters, thereby realizing new product recommendation. The application effectively improves the generalization ability of the model by using the idea of alignment and invariant learning, and relieves the cold start recommendation problem.