Cold-start Users Research Articles

MFPCDR: A Meta-Learning-Based Model for Federated Personalized Cross-Domain Recommendation

Cross-domain recommendation systems frequently require the use of rich source domain information to improve recommendations in the target domain, thereby resolving the data sparsity and cold-start problems, whereas the majority of existing approaches frequently require the centralized storage of user data, which poses a substantial risk of privacy breaches. Compared to traditional recommendation systems with centralized data, federated recommendation systems with multiple clients trained collaboratively have significant privacy benefits in terms of user data. While users’ interests are often personalized, meta-learning can be used to learn users’ personalized preferences, and personalized preferences can help models make recommendations in cold-start scenarios. We use meta-learning to learn the personalized preferences of cold-start users. Therefore, we offer a unique meta-learning-based federated personalized cross-domain recommendation model that discovers the personalized preferences for cold-start users via a server-side meta-recommendation module. To avoid compromising user privacy, an attention mechanism is used on each client to find transferable features that contribute to knowledge transfer while obtaining embeddings of users and items; each client then uploads the weights to the server. The server accumulates weights and delivers them to clients for update. Compared to traditional recommendation system models, our model can effectively protect user privacy while solving the user cold-start problem, as we use an attention mechanism in the local embedding module to mine the source domain for transferable features that contribute to knowledge transfer. Extensive trials on real-world datasets have demonstrated that our technique effectively guarantees speed while protecting user privacy.

Intelligent Innovation Platform Design of Chinese Film and Television Music Based on Multi-Dimensional Information Fusion Algorithm

Experienced changes in recording techniques and musical styles. With the development of the times, digital technology is constantly developing and progressing. In the aspect of film and television music creation, the application of digital technology has been innovated accordingly. Using multi-source information fusion to build a music feature system to solve the problem of music recommendation cold start Users provide personalized music recommendations. The optimized high-concurrency Cuckoo Hashing algorithm is implemented on the DPDK platform, and the algorithm is optimized for large page memory, batch processing and preprocessing combined with DPDK features. Music information retrieval is expanded from the traditional retrieval based on keywords such as file name and artist name. At the same time, the evaluation of the example model of the system also verifies that the multi-information source fusion method proposed in this paper can optimize the precision and recall rate of the system.

A Deep Reinforcement Learning Recommender System With Multiple Policies for Recommendations

Deep reinforcement learning (DRL) based recommender systems are suitable for user cold-start problems as they can capture user preferences progressively. However, most existing DRL-based recommender systems are suboptimal, since they use the same policy to suit the dynamics of different users. We reformulate recommendation as a multitask Markov Decision Process, where each task represents a set of similar users. Since similar users have closer dynamics, a task-specific policy is more effective than a single universal policy for all users. To make recommendations for cold-start users, we use a default policy to collect some initial interactions to identify the user task, after which a task-specific policy is employed. We use Q-learning to optimize our framework and consider the task uncertainty by the mutual information regarding tasks. Experiments are conducted on three real-world datasets to verify the effectiveness of our proposed framework.

A Multi-strategy-based Pre-training Method for Cold-start Recommendation

The cold-start issue is a fundamental challenge in Recommender Systems. The recent self-supervised learning (SSL) on Graph Neural Networks (GNNs) model, PT-GNN, pre-trains the GNN model to reconstruct the cold-start embeddings and has shown great potential for cold-start recommendation. However, due to the over-smoothing problem, PT-GNN can only capture up to 3-order relation, which cannot provide much useful auxiliary information to depict the target cold-start user or item. Besides, the embedding reconstruction task only considers the intra-correlations within the subgraph of users and items, while ignoring the inter-correlations across different subgraphs. To solve the above challenges, we propose a multi-strategy-based pre-training method for cold-start recommendation (MPT), which extends PT-GNN from the perspective of model architecture and pretext tasks to improve the cold-start recommendation performance. 1 Specifically, in terms of the model architecture, in addition to the short-range dependencies of users and items captured by the GNN encoder, we introduce a Transformer encoder to capture long-range dependencies. In terms of the pretext task, in addition to considering the intra-correlations of users and items by the embedding reconstruction task, we add an embedding contrastive learning task to capture inter-correlations of users and items. We train the GNN and Transformer encoders on these pretext tasks under the meta-learning setting to simulate the real cold-start scenario, making the model able to be easily and rapidly adapted to new cold-start users and items. Experiments on three public recommendation datasets show the superiority of the proposed MPT model against the vanilla GNN models, the pre-training GNN model on user/item embedding inference, and the recommendation task.

User structural information in priority-based ranking for top-N recommendation

The recommender system is a set of data recovery tools and techniques used to recommend items to users based on their selection. To improve the accuracy of the recommendation, the use of additional information (e.g., social information, trust, item tags, etc.) in addition to user-item ranking data has been an active area of research for the past decade. In this paper, we present a new method for recommending top-N items, which uses structural information and trust among users within the social network and extracts the implicit connections between users and uses them in the item recommendation process. The proposed method has seven main steps: (1) extract items liked by neighbors, (ii) constructing item features for neighbors, (iii) extract embedding trust features for neighbors, (iv) create user-feature matrix, (v) calculate user’s priority, (vi) calculate item’s priority and finally, (vii) recommend top-N items. We implement the proposed method with three datasets for recommendations. We compare our results with some advanced ranking methods and observe that the accuracy of our method for all users and cold-start users improves. Our method can also create more items for cold-start users in the list of recommended items.

Meta-GRS: A Graph Neural Network for Cross-Domain Recommender System via Meta-Learning

Despite the fact that Recommender Systems have been researched and developed for quite a while, they still hold some certain challenges, two of which are Cold Start and Data Sparsity. Being introduced as a savior, the Cross-domain Recommender system (CDR) owns the capacity to transfer knowledge across domains which makes it become the practical solution for the mentioned issues. In CDR approaches, the family of graph-based solutions is very effective, which builds graphs to illustrate the relationship between users, items, and other factors and learn their representations through graph representation learning. However, most graph-based approaches focus on extracting either domain-specific or domain-shared features and do not have the mechanism to prevent the transfer of private features, which can degrade the quality of vector representations. Moreover, the knowledge transfer process built on overlapping users makes the model biased toward these users, thus downgrading the performance on cold-start users. This paper proposes a meta-GRS framework that uses Graph Neural Network with Meta-Learning and Adversarial Learning for cross-domain recommendation. In meta-GRS, the representative user and item embeddings are improved thanks to the features extracted by the private and cross-domain graphs. Domain-shared features are learned under adversarial learning such that the domain discriminator is unable to determine whether the domain they came from can ensure the positive transfer. To optimize the model performance effectively, we use a Dynamic Weight Averaging algorithm to learn loss weights automatically. The model's parameters are optimized under the optimization-based meta-learning method provides our model the capability of generalization to the new users. Experiments on practical datasets illustrate that the meta-GRS leads the chart in the comparison of other state-of-the-art baselines in recommendation accuracy.

Adversarial Auto-encoder Domain Adaptation for Cold-start Recommendation with Positive and Negative Hypergraphs

This article presents a novel model named Adversarial Auto-encoder Domain Adaptation to handle the recommendation problem under cold-start settings. Specifically, we divide the hypergraph into two hypergraphs, i.e., a positive hypergraph and a negative one. Below, we adopt the cold-start user recommendation for illustration. After achieving positive and negative hypergraphs, we apply hypergraph auto-encoders to them to obtain positive and negative embeddings of warm users and items. Additionally, we employ a multi-layer perceptron to get warm and cold-start user embeddings called regular embeddings. Subsequently, for warm users, we assign positive and negative pseudo-labels to their positive and negative embeddings, respectively, and treat their positive and regular embeddings as the source and target domain data, respectively. Then, we develop a matching discriminator to jointly minimize the classification loss of the positive and negative warm user embeddings and the distribution gap between the positive and regular warm user embeddings. In this way, warm users’ positive and regular embeddings are connected. Since the positive hypergraph maintains the relations between positive warm user and item embeddings, and the regular warm and cold-start user embeddings follow a similar distribution, the regular cold-start user embedding and positive item embedding are bridged to discover their relationship. The proposed model can be easily extended to handle the cold-start item recommendation by changing inputs. We perform extensive experiments on real-world datasets for both cold-start user and cold-start item recommendations. Promising results in terms of precision, recall, normalized discounted cumulative gain, and hit rate verify the effectiveness of the proposed method.

A Temporal and Social Network-based Recommender using Graph Clustering

Recommendation Systems (RSs) have significant applications in many industrial systems. The duty of a recommender algorithm is to operate available data (users/items contextual data and rating (or purchase) the consumption history for items), as well as to provide a recommendation list for any target user. The recommended items should be selected so that the target user is compelled to give them positive reviews. In this manuscript, we propose a novel of RS algorithm that makes advantage of user-user trust relationships, rating histories, and their frequency of occurrence. We also provide a brand new overlapping community detection algorithm. The information about the users’ community structure is used to handle the cold-start and sparsity problems. We compare the performance of the proposed RS algorithm with a number of state-of-the-art algorithms on the extended Epinions dataset, which has both information on trust relations and the timing of the ratings. Numerical simulations reveal the superiority of the proposed algorithm over others. We also investigate how the algorithms perform when only cold-start users and items are considered. As a cold-start user (item) we consider those that have made (received) less than five ratings. The experiments show significant outperformance of the proposed algorithm over others, which is mainly due to the use of information on overlapping community structures between users.

Personalized recommender systems based on social relationships and historical behaviors

Recommender systems have a wide range of applications in the age suffering information overload. A promising way to design better recommender systems in the presence of ubiquitous social media is to utilize social relationships in recommendation algorithms, named social recommendation. One critical challenge in social recommendation is how to mine valuable information intrinsic to social relationships and integrate such information into the algorithm design. In this paper, we argue that both social relationships and historical behaviors are affected by the same implicit factors. For example, due to the existence of implicit factors such as peer influence or common interests in social networks, users with similar implicit factors will have a high probability to become friends and collect similar objects. Accordingly, we propose a recommendation algorithm that jointly utilizes social relationships and historical behaviors, based on the extended linear optimization technique. We test the performance of our algorithm for four groups of users on real networks, including all users, active users, inactive users and cold-start users. Results show that, in all the above four scenarios, the proposed algorithm performs overall best subject to accuracy and diversity metrics compared with the benchmarks. In particular, the algorithm remarkably improves the recommendation performance for cold-start users. Further analysis shows that the contribution of social relationships depends on the coupling strength between social relationships and historical behaviors.

Implementation of Personalized Information Recommendation Platform System Based on Deep Learning Tourism

In order to provide tourists with better tourism services, a system method of personal information recommendation platform based on deep learning tourism is proposed. The system includes noise reduction autoencoder, feature extraction module, data preprocessing module, recommendation calculation module, expert evaluation module, recommendation result output module, customer feedback module, and storage module. The personal information recommendation platform system based on deep learning tourism of the present invention enables tourists to obtain tourism information conveniently and quickly through scientific information organization and presentation form and helps tourists to better arrange tourism plans and form tourism decisions. By effectively aggregating multiple neighborhoods of nodes, embedding high-order collaboration information into the node embedding vector, obtaining the potential preferences of users, solving the problems of user data sparse and cold start, and finally through experimental analysis, a research method is proposed. It is used to build the model of tourist attraction recommendation system. Experimental results show that the proposed method for cold-start user recommendation has the best performance in terms of accuracy, recall, and normalized loss cumulative gain, and it is 17.9% higher than BPR in recall rate Recall@5 and 11.8% higher in accuracy rate. It is proved that the system has a significant impact on the diversity and novelty of tourist attraction recommendation.

Deep transfer learning with multimodal embedding to tackle cold-start and sparsity issues in recommendation system.

Recommender systems (RSs) have become increasingly vital in the modern information era and connected economy. They play a key role in business operations by generating personalized suggestions and minimizing information overload. However, the performance of traditional RSs is limited by data sparseness and cold-start issues. Though deep learning-based recommender systems (DLRSs) are very popular, they underperform when considering rating matrices with sparse entries. Despite their performance improvements, DLRSs also suffer from data sparsity, cold start, serendipity, and generalizability issues. We propose a multistage model that uses multimodal data embedding and deep transfer learning for effective and personalized product recommendations, and is designed to overcome data sparsity and cold-start issues. The proposed model includes two phases. In the first—offline—phase, a deep learning technique is implemented to learn hidden features from a large image dataset (targeting new item cold start), and a multimodal data embedding is used to produce dense user feature and item feature vectors (targeting user cold start). This phase produces three different similarity matrices that are used as inputs for the second—online—phase to generate a list of top-n relevant items for a target user. We analyzed the accuracy and effectiveness of the proposed model against the existing baseline RSs using a Brazilian E-commerce dataset. The results show that our model scored 0.5882 for MAE and 0.4011 for RMSE which is lower than baseline RSs which indicates that the model achieved an improved accuracy and was able to minimize the typical cold start and data sparseness issues during the recommendation process.

ImposeSVD: Incrementing PureSVD For Top-N Recommendations for Cold-Start Problems and Sparse Datasets

Abstract In this paper, we introduced two novel collaborative filtering techniques for recommendation systems in cases of various cold-start situations and incomplete datasets. The first model establishes an asymmetric weight matrix between items without using item meta-data and eradicates the disadvantages of neighborhood approaches by automatic determination of threshold values. Our first model, z-scoREC, is also regarded as a pure deep-learning model because it performs like a vanilla auto-encoder in transforming column vectors with z-score normalization similar to batch normalization. With the second model, ImposeSVD, we aimed to enhance the shortcomings of the PureSVD in cases of cold-start and incomplete data by preserving its straightforward implementation and non-parametric form. The ImposeSVD model relies on the z-scoREC, produces synthetic new predictions for the users by decomposing the latent factors from the imposed matrix. We evaluated our method on the well-known datasets and found out that our method was outperforming similar approaches in the specific scenarios including recommendations for cold-start users, strength in cold-start systems, and diversification of long-tail item recommendations in lists. Our z-scoREC model also outperformed familiar neighbor-based approaches when operated as a recommender system and gave a closer appearance to the decomposition methods despite its simple and rigid cost framework.

Robust Collaborative Filtering Recommendation With User-Item-Trust Records

The ever-increasing popularity of recommendation systems allows users to find appropriate services without excessive effort. However, due to the unstable and complex network environment, the historical behavior data of users are quite sparse in most cases. The inherent drawbacks render preference prediction infeasible for cold-start users and have become a crucial issue to be resolved in recommendation systems. To deal with the problems, we first present a Trust-based Collaborative Filtering (TbCF) algorithm to perform basic rating prediction in a manner consistent with the existing CF methods. Then, we propose the Hybrid Collaborative Filtering Recommendation approach with User-Item-Trust Records ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\text {UIT}_{\text {hybrid}}$ </tex-math></inline-formula> ), a novel approach that incorporates user trust into the existing CF-based methods in a harmonious way to supplement rating information. <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\text {UIT}_{\text {hybrid}}$ </tex-math></inline-formula> employs multiple perspectives to extract proper services and achieves a good tradeoff between the robustness, accuracy, and diversity of the recommendation. We conduct extensive real-world experiments on the Epinions data set to demonstrate the feasibility and efficiency of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\text {UIT}_{\text {hybrid}}$ </tex-math></inline-formula> .

A Reliable and Lightweight Trust Inference Model for Service Recommendation in SIoT

In the era of Internet of Things (IoT), millions of heterogeneous IoT devices generate an explosion of data and services waiting to be discovered. The convergence of IoT with social networks (SIoT) interconnects multiple IoT applications and alleviates the common data sparsity and cold start problems in traditional recommendation systems. However, the social trust relationships may also be very sparse, which affects the accuracy of trust-based recommendation systems. Meanwhile, mobile devices have limited resources and are more vulnerable to malicious attacks in the IoT environment. In order to complete the trust relationship and further improve the trust-based recommendation performance, we propose a reliable and lightweight trust inference model for service recommendation in SIoT, called <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">TIRec</i> . First, we obtain a comprehensive weighted centrality metric (LGWC) considering both local and global contexts. Based on this, we propose a corresponding lightweight trust path selection algorithm. Then, we present a reliable trust inference calculation algorithm consist of trust propagation and aggregation strategy, which can efficiently resist two common malicious attacks. Finally, we incorporate the rating, direct trust, and indirect trust together into the matrix factorization model, and integrate the influence of truster and trustee to obtain the synthetic model for rating predication. To the best of our knowledge, this article is the first to integrate trust inference algorithm into the trust-based recommendation systems. The extensive experiments are conducted on three real-world data sets, and the results show that our <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">TIRec</i> model performs better than other advanced recommendation models in both “all users” view and “cold start users” view.

A Recommendation Model for College English Digital Teaching Resources Using Collaborative Filtering and Few-Shot Learning Technology

This study designs and implements a digital English instructional resource management recommendation system based on collaborative filtering technology based on CF research and the construction of digital English instructional resources. This study designs the system with B/S structure combined with hierarchical design architecture, plans the overall design goal, architecture design, compilation structure, and key technologies, and designs and implements the system's core modules on the basis of fully analysing the functional and nonfunctional requirements of personalized educational resource recommendation system. Furthermore, the traditional CF has been improved to address scalability, data sparseness, and user cold start in the recommendation process. The evaluation results show that this algorithm has a recall rate of 96.37% and a system resource recommendation accuracy rate of 95.31%, both of which are higher than the traditional method's 6.37%. This algorithm can overcome the drawbacks of traditional algorithms, improve recommendation accuracy, and efficiently provide high-quality English teaching resources. Educators and students will be more likely to find high-quality digital English instructional resources if they use the instructional resource system proposed in this study.

SMINet: State-Aware Multi-Aspect Interests Representation Network for Cold-Start Users Recommendation

Online travel platforms (OTPs), e.g., bookings.com and Ctrip.com, deliver travel experiences to online users by providing travel-related products. Although much progress has been made, the state-of-the-arts for cold-start problems are largely sub-optimal for user representation, since they do not take into account the unique characteristics exhibited from user travel behaviors. In this work, we propose a State-aware Multi-aspect Interests representation Network (SMINet) for cold-start users recommendation at OTPs, which consists of a multi-aspect interests extractor, a co-attention layer, and a state-aware gating layer. The key component of the model is the multi-aspect interests extractor, which is able to extract representations for the user's multi-aspect interests. Furthermore, to learn the interactions between the user behaviors in the current session and the above multi-aspect interests, we carefully design a co-attention layer which allows the cross attentions between the two modules. Additionally, we propose a travel state-aware gating layer to attentively select the multi-aspect interests. The final user representation is obtained by fusing the three components. Comprehensive experiments conducted both offline and online demonstrate the superior performance of the proposed model at user representation, especially for cold-start users, compared with state-of-the-art methods.

A Dynamic Meta-Learning Model for Time-Sensitive Cold-Start Recommendations

We present a novel dynamic recommendation model that focuses on users who have interactions in the past but turn relatively inactive recently. Making effective recommendations to these time-sensitive cold-start users is critical to maintain the user base of a recommender system. Due to the sparse recent interactions, it is challenging to capture these users' current preferences precisely. Solely relying on their historical interactions may also lead to outdated recommendations misaligned with their recent interests. The proposed model leverages historical and current user-item interactions and dynamically factorizes a user's (latent) preference into time-specific and time-evolving representations that jointly affect user behaviors. These latent factors further interact with an optimized item embedding to achieve accurate and timely recommendations. Experiments over real-world data help demonstrate the effectiveness of the proposed time-sensitive cold-start recommendation model.

Combining feature importance and neighbor node interactions for cold start recommendation

Cold start recommendation usually views preference embedding as a missing problem because there is not any historical interaction. Existing approaches on graph neural networks for cold start users/items build the attribute embedding of each node through simply concatenating multiple features equally, and then reconstruct node preference embedding from its attribute embedding through a mapping function which is learned from warm users/items. However, these approaches do not consider the different contributions of features for building the attribute embedding. In addition, they assume the neighbors of a target node are independent and ignore interactions between the neighbor nodes when building the mapping function between the attribute embedding and the preference embedding. These two limitations reduce the effectiveness of their performance. To overcome these limitations, we propose a novel framework called Feature Importance and Neighbor node Interactions graph neural network (FINI) that exploits feature weights and interactions between neighbor nodes. The core ideas of the proposed method are as follows. First, it designs a global–local contexts attention mechanism in the attribute encoding layer, which can dynamically learn the importance of the attributes of each node and improve the expression of the feature embeddings. Second, it proposes a mixed interaction mechanism to augment the weighted information of neighbor node interactions in the neighbor interaction layer, which can strengthen the expressive capability of the user/item embeddings and further improve the quality of the mapping function for cold start users/items. Additionally, we also combine the rating prediction loss and mimic loss as the total loss to train the network in the prediction layer for model training. To assess the performance of the FINI, both cold start users and cold start items recommendation are considered. The results demonstrate FINI outperforms the state-of-the-art approaches for cold start recommendation and gains significant improvements in terms of metric evaluations.

RecMem: Time Aware Recommender Systems Based on Memetic Evolutionary Clustering Algorithm.

Nowadays, the recommendation is an important task in the decision-making process about the selection of items especially when item space is large, diverse, and constantly updating. As a challenge in the recent systems, the preference and interest of users change over time, and existing recommender systems do not evolve optimal clustering with sufficient accuracy over time. Moreover, the behavior history of the users is determined by their neighbours. The purpose of the time parameter for this system is to extend the time-based priority. This paper has been carried out a time-aware recommender systems based on memetic evolutionary clustering algorithm called RecMem for recommendations. In this system, clusters that evolve over time using the memetic evolutionary algorithm and extract the best clusters at every timestamp, and improve the memetic algorithm using the chaos criterion. The system provides appropriate suggestions to the user based on optimum clustering. The system uses optimal evolutionary clustering using item attributes for the cold-start item problem and demographic information for the cold start user problem. The results show that the proposed method has an accuracy of approximately 0.95, which is more effective than existing systems.

Enhancing Personalized Recommendation by Transductive Support Vector Machine and Active Learning

As an important component of information service networks, personalized recommendation technology provides users with better options and enables them to obtain information anytime and anywhere. Collaborative filtering (CF) is a successful and widely used form of this technology. However, the traditional CF recommendation algorithm is ineffective in environments with frequent entry of new users and high levels of data sparsity. For new users in the system, few or no scores, labels, or other such information is available, leading to the user cold start problem. Simultaneously, data sparsity leads to the selection of unreasonable neighbors, which reduces the recommendation accuracy. In addition, the traditional CF recommendation algorithm ignores the inherent connections between users’ preferences and their basic information (such as demographics). Users with similar demographic information are likely to have similar preferences, which can serve as a good basis for finding neighbors. To address the aforementioned problems, we propose a recommendation model that combines active learning (AL) and a semi-supervised transductive support vector machine (TSVM). To enable neighbors to be found quickly and accurately, similar users are clustered together on the basis of their basic information. Then, the TSVM-based classifier is trained on each cluster. To improve the quality of sample labeling and thus the classifier performance, an active learning method based on the distance strategy and a multiclassifier voting mechanism is implemented. Finally, the TSVM-based recommendation model is trained on the labeled samples. The extensive experiments conducted using a real data set from MovieLens demonstrate that the proposed model effectively alleviates the aforementioned cold start and data sparsity problems.