In recent years, Large Language Model-based Recommendation Systems (LLMRec) have achieved remarkable advancements. By designing effective prompts, LLMs can understand user interests and either directly generate recommendations based on users' interests or assist traditional recommendation systems in enhancing recommendation performance. However, existing LLMRec predominantly leverage LLMs as static tools for user interest understanding, ignoring the dynamic behavioral patterns of LLMs when simulating human personalities. While some studies explore LLMs' human-like behaviors in conversational scenarios, their impacts on recommendation systems remain underexplored. To fill these gaps, this paper aims to investigate how LLM-simulated personalities interact with users of varying personalities in recommendation systems, particularly focusing on whether LLM-simulated personalities exhibit behaviors align with Social Homophily Theory which is well-documented in human social interactions but unexplored in LLMRec. We conducted experiments on two public datasets, inducing LLMs to simulate different human personalities based on the Big Five Personality Traits through prompt engineering, and incorporating the generated user profiles into the Knowledge Augmented Recommendation framework to evaluate performance across various recommendation models. Experimental results demonstrate that inducing LLMs to simulate personality in the KAR framework significantly influences recommendation performance. Moreover, LLM-simulated personalities exhibit a preference for users with similar personalities, mirroring the Social Homophily Theory observed in human behavior.

Sequential recommendation systems aim to predict users' future interests based on their historical interactions and are widely applied in domains such as e-commerce and social media. However, in real-world scenarios, recommendation data often exhibits a long-tail distribution, resulting in extremely sparse interactions for a large number of users. Traditional methods face significant challenges in modeling such user preferences, thereby compromising recommendation performance and personalization. In recent years, Large Language Models (LLMs) have shown great promise in sparse scenarios due to their powerful language understanding and generation capabilities, offering new opportunities for user interest modeling. Nevertheless, existing approaches often rely on a single prompt or focus solely on item-side semantic information, neglecting users' multi-interest structures and collaborative behavior signals, which limits the reasoning potential of LLMs. To address these issues, this paper proposes the PERec (Prompt-enhanced Semantic Modeling with Large Language Models for Long-tail Recommendation) framework, aiming to enhance semantic modeling and recommendation performance for long-tail users. PERec builds a prompt-driven multi-interest semantic enhancement structure. Specifically, it first generates multiple personalized prompts from users' behavioral histories to guide LLMs in learning diverse preference-aware semantic representations. It then constructs semantic and collaborative dual-view representations and employs a contrastive alignment mechanism to enhance inter-view consistency. Finally, a view fusion strategy is used to generate robust user representations, thereby improving recommendation performance in sparse longtail scenarios. We conduct empirical studies on three real-world datasets-Amazon Beauty, Amazon Clothing, and Yelp. The results demonstrate that PERec significantly outperforms representative baseline methods in both overall recommendation performance and long-tail user accuracy, verifying its effectiveness and practical value in sparse recommendation settings.

Conversational Recommender Systems (CRS) enable more accurate and personalized recommendation through multi-turn interactions by dynamically capturing user preferences. However, the data imbalance in CRSs, due to the inconsistent distribution of user interaction histories, may lead to biased treatment for disadvantaged user group. This paper systematically investigates the discriminate problems in CRS from the user's perspective, called as User-Oriented Fairness (UOF). Through comprehensive empirical analysis across multiple dimensions, we reveal significant disparities in recommendation quality and representation learning among user groups with varying interaction frequencies. To address these challenges, we propose a user fairness enhancement framework, named DareCRS, which is a model-agnostic framework. The novelty of DareCRS lies in two key components: (1) At the data level, it integrates LLM-powered Dialog Augmentation with a hybrid sampling strategy to effectively mitigate data sparsity for disadvantaged users; (2) At the representation level, it introduces a Representation Equalization mechanism based on optimal transport between user groups, aiming to enhance the embedding space of disadvantaged users via knowledge transfer. Extensive experiments on two public datasets show that DareCRS achieves an average improvement of 5.37% in recommendation performance and an average enhancement of 43.14% in recommendation fairness across user groups compared to current state-of-the-art baseline methods. Our code and data are available at: https://anonymous.4open.science/r/DareCRS-8B94.

Abstract Coloring an animation sketch sequence is a challenging task in computer vision since the information contained in line sketches is too sparse, and the colors need to be uniform between continuous frames. Many the existing colorization algorithms can only be applied to one image and can be considered color filling algorithms. Such algorithms only provide a color result that fits within a reasonable range and can not be applied to the coloring of frame sequences. This paper proposes an end‐to‐end two‐stage optical flow colorization network to solve the animation frame sequence colorization problem. The first stage of the network finds the direction of the color pixel flow from the detail change between a given reference frame and the next frame of line artwork and then completes the initial coloring process. The second stage of the network performs color correction and clarifies the output of the first stage. Since our algorithm does not directly colorize the image but finds the path of the color change to colorize it, it ensures a consistent color space for the sequence frames after colorization. We conduct experiments on an animation dataset, and the results show that our algorithm is effective. The code is available at https://github.com/silenye/Colorization .

Deep learning based recommender systems(DLRS) as one of the up-and-coming recommender systems, and their robustness is crucial for building trustworthy recommender systems. However, recent studies have demonstrated that DLRS are vulnerable to data poisoning attacks. Specifically, an unpopular item can be promoted to regular users by injecting well-crafted fake user profiles into the victim recommender systems. In this paper, we revisit the data poisoning attacks on DLRS and find that state-of-the-art attacks suffer from two issues: user-agnostic and fake-user-unitary or target-item-agnostic, reducing the effectiveness of promotion attacks. To gap these two limitations, we proposed our improved method Generate Targeted Attacks(GTA), to implement targeted attacks on vulnerable users defined by user intent and sensitivity. We initialize the fake users by adding seed items to address the cold start problems of fake users so that we can implement targeted attacks. Our extensive experiments on two real-world datasets demonstrate the effectiveness of GTA.

Existing work on social robotic persuasion (SRP) has provided ample knowledge that can benefit the design of social robotics. However, as recent research points out, this knowledge is not presented in a format that effectively supports design practice. Research on translational science provides a theoretical foundation for connecting research to design practice, but the process of knowledge appropriation remains under-explored. In this paper, we present the development and evaluation of the SRPD (social robot persuasive design) cards and the corresponding Roundtable and Spotlight generative methods. Our results show that the SRPD cards can benefit the ideation phase in social robotic design and that our generative methods can optimize participants’ experience in a brainstorming activity.

Social robotic (SR) applications are expected to advance in the near future, which is indicated by the level of autonomy of SR, and is supported by the maturing of key technologies. However, technical, research-practice communication, and ethical challenges remain in the design of such advanced SR applications. In this paper, we start with our research cases through which we encountered and explored the challenges facing the design practice of advanced SR applications. Then we present an overview of recent design research endeavors in the HRI field and our research cases to demonstrate design research’s characteristics, focus, and approaches. Finally, we discuss how the eliciting, speculative, and communicative approaches of design research can support user investigations with the hindrances of technical limitations, communicate between basic & applied research and design practice, and engage in ethical explorations of advanced SR applications.

The federated framework is actively applied in knowledge graph fusion research to obtain a complete knowledge graph without exposing data privacy. It can help local clients learn the knowledge graph embeddings in other clients without revealing data privacy. However, current federated-based knowledge graph embedding frameworks cannot exploit both entity and relation embeddings and may not prevent partial triples from being reconstructed. This paper proposes a novel framework named Federated Multi-server knowledge graph embedding (FedM), which creatively utilizes uploaded entity and relation embeddings while preventing privacy leakage. Expressly, we first set up two central servers for entity and relation embeddings to aggregate and share client-uploaded embeddings. Secondly, we design a knowledge graph secure aggregation algorithm to address the potential privacy concerns in FedM. We conduct comparative experiments on an empirical dataset (divided into three federated datasets) with four commonly-used knowledge graph embedding methods to evaluate the performance of our proposed framework. In addition, our proposed FedM framework is generally superior to the latest baseline frameworks on both privacy preservation and link prediction tasks.

Point of Interest (POI) recommendation system rec-ommends places in which users have never been to but may be in-terested. Traditionally, it centrally collects contextual information and interaction data to model users' preferences, which raises many privacy concerns. The current studies habitually sacrifice the recommendation performance to cope with privacy con-cerns. To protect users' privacy while ensuring the performance of the POI recommendation system, we propose a Ranking-based Federated POI Recommendation with Geographic Effect (RFPG). The RFPG allows users to reserve their private data on local devices to secure privacy. It adaptively constructs an active region to model the geographic effect, enhancing users' personalized preference modeling. In addition, we design a probability-based negative sampling method to protect privacy further and improve recommendation performance. This method calculates the probability of a POI being a negative sample through POI geographic distribution, then combines the positive samples to construct a local triplet training dataset. Theoretical analysis and experiments on two real datasets demonstrate that our proposed RFPG improves the performance of the POI recommendation while protecting users' private data.

Recommender systems are collecting user data to provide better personalized services. However, centralized collection and analysis of users' private data will raise privacy concerns and legal risks. The emergence of federated learning enables training a machine learning model from highly decentralized data. This paper extends the Neural Collaborative Filtering (NCF) method using a federated setting and proposes a privacy-preserving federated recommender system named FedNCF, which can train NCF models without needing to know user's private data. We apply differential privacy to the computed gradients to prevent inference attacks. To improve the utility of differential privacy, we propose an adaptive differential privacy approach that meticulously adjusts the noise scale in each iteration controlled by a decay rate. Our proposed approach provides an explicit mathematical expression to estimate the user's privacy loss by truncated Concentrated Differential Privacy (tCDP). Extensive experiments and analysis demonstrate that FedNCF can achieve competitive performance with the centralized NCF meanwhile effectively protect users' privacy.

Existing federated social recommendation systems mainly focus on protecting users' data privacy, and the hetero-geneity of data distribution among users is less mentioned. A key reason for the poor performance of federated social recommendation systems compared to traditional ones is the heterogeneity of the local data in the federated setting. Also, it severely limits the development of federated social recommendation systems. This paper proposes a federated social recommendation framework based on Contrastive Learning. We use contrastive learning to minimize the distance between a user and his trusted users on the user level's feature space and maximize the consistency between local and global item embeddings for item embedding. Corrections for updates to user embeddings and item embeddings alleviate the performance impact of heterogeneous data distributions. Experiments on three datasets show that our approach significantly improves performance over existing methods.

This paper presents a qualitative study that uses video prototypes and interviews to explore older adults’ acceptance, needs, and design requirements towards a social robotic application for physical rehabilitation. Our study identified the benefits of applying social robots (SR) in physical rehabilitation. Further, we discovered participants’ preference for an anthropomorphic social robot design. The data revealed a desire for social interaction could increase motivation for older adults to engage in an active lifestyle and social robot acceptance. However, participants showed low motivation for technology adoption and negatively anthropomorphize the social robot, which lowers acceptance for their application. This work complements the current user-centered explorations with SR in rehabilitation, and provides considerations for SR design for rehabilitative applications.

As people enjoy the personalized services recommended by Recommender Systems (RSs), the privacy disclosure risk increases with frequent interactions. Malicious adversary often collects public information online to infer private information for illicit profit. As privacy concerns grew, researchers introduced data obfuscation into recommender systems. However, there still exists several limitations in current work. First, although the existing methods effectively reduce the risk of privacy disclosure, they can be detrimental to the quality of the recommendation service. Second, a range of practical issues under the application of recommendation systems are not considered, e.g., long-tail, density, etc. To address those challenges, we propose a novel framework named Want User Defending Inference (WUDI), a high-performance privacy-preserving debiased framework based on data obfuscation. Unlike the original strategies, i.e., adding or removing user ratings, we introduced some novel strategies to generate an obfuscated matrix. Firstly, we define a new method called Cluster Recommend for alleviating the long-tail skewness and data sparsity in RSs. Then we investigate the gender bias in obfuscation and apply a bias mitigating strategy to RSs. Experiments on public datasets demonstrate that WUDI can outperform the state-of-the-art baselines in obfuscation.

The cross-domain sequential recommendation aims to predict the next item based on a sequence of recorded user behaviours in multiple domains. We propose a novel Cross-domain Sequential Recommendation approach with Graph-Collaborative Filtering (CsrGCF) to alleviate the sparsity issue of user-interaction data. Specifically, we design time-aware and relation-aware graph attention mechanisms with collaborative filtering to exploit high-order behaviour patterns of users for promising results in both domains. Time-aware Graph Attention mechanism (TGAT) is designed to learn the inter-domain sequence-level representation of items. Relationship-aware Graph Attention mechanism (RGAT) is proposed to learn collaborative items' and users' feature representations. Moreover, to simultaneously improve the recommendation performance in the two domains, a Cross-domain Feature Bidirectional Transfer module (CFBT) is proposed, transferring user's common sharing features in both domains and retaining user's domain-specific features in a specific domain. Finally, cross-domain and sequential information jointly recommend the next items that users like. We conduct extensive experiments on two real-world datasets that show that CsrGCF outperforms several state-of-the-art baselines in terms of Recall and MRR. These demonstrate the necessity of exploiting high-order behaviour patterns of users for a cross-domain sequential recommendation. Meanwhile, retaining domain-specific features is an important step in the process of cross-domain feature bidirectional transferring.

In recent years, the number of traffic accident deaths due to distracted driving has been increasing dramatically. Fortunately, distracted driving can be detected by the rapidly developing deep learning technology. Nevertheless, considering that real-time detection is necessary, three contradictory requirements for an optimized network must be addressed: a small number of parameters, high accuracy, and high speed. We propose a new D-HCNN model based on a decreasing filter size with only 0.76M parameters, a much smaller number of parameters than that used by models in many other studies. D-HCNN uses HOG feature images, L2 weight regularization, dropout and batch normalization to improve the performance. We discuss the advantages and principles of D-HCNN in detail and conduct experimental evaluations on two public datasets, AUC Distracted Driver (AUCD2) and State Farm Distracted Driver Detection (SFD3). The accuracy on AUCD2 and SFD3 is 95.59% and 99.87%, respectively, higher than the accuracy achieved by many other state-of-the-art methods.

As the number of online job postings and users grows dramatically, the accuracy and explainability of personjob fit systems are of increasing concern. An explainable personjob fit system can show reasons when making recommendations to both Human Resources and Job Seekers, building trust between uses and recommendation system while providing accurate recommendation results. However, the existing research on content-based person-job fit mainly focuses on 1) dealing with unstructured statements without effectively using structured information in resumes and jobs, and 2) the explanations of the model stay at the level of giving a few sentences, which leads to a lack of explanations. In this paper, we propose an explainable person-job fit model based on the attention mechanism. We model the resume text through a hierarchical attention mechanism and capture the semantic connections between the resume, structured job text, and unstructured job text through a collaborative attention mechanism to better model the job content and provide both structured and unstructured levels of recommendation explanation. Experiments on a large real dataset show that our model outperforms existing baseline models and provides job recommendation reasons at both levels.

Abstract Researchers usually face difficulties in finding scientific papers relevant to their research interests due to increasing growth. Recommender systems emerge as a leading solution to filter valuable items intelligently. Recently, deep learning algorithms, such as convolutional neural network, improved traditional recommendation technologies, for example, the graph‐based or content‐based methods. However, existing graph‐based methods ignore high‐order association between users and items on graphs, and content‐based methods ignore global features of texts for explicit user preferences. Therefore, this paper proposes a Content‐based and knowledge Graph‐based Paper Recommendation method (CGPRec), which uses a two‐layer self‐attention block to obtain global features of texts for more complete explicit user preferences, and proposes an improved graph convolutional network for modeling high‐order associations on the knowledge graph to mine implicit user preferences. And the knowledge graph in this paper is constructed with concept nodes, user nodes, paper nodes, and other meta‐data nodes. Experimental results on a public dataset, CiteULike‐a, and a real application log dataset, AHData, show that our model outperforms compared with baseline methods.