Retrieval-augmented generation (RAG) has shown strong potential for deploying large language models at the edge, yet existing designs largely rely on generic and monolithic knowledge bases that are poorly matched to the heterogeneous queries and resource-constrained edge computing environments. Through extensive empirical analysis, we find that domain-specialized knowledge bases, when deployed on individual edge servers, deliver substantially higher retrieval accuracy and generation quality than generic knowledge bases under identical resource budgets. Based on this, we propose SRAG, a distributed RAG system that enforces knowledge specialization at the edge. Each edge server maintains a domain-aware specialized knowledge base by retaining domain-aligned knowledge and decoupling out-of-domain content. SRAG uses a buffer-based knowledge migration mechanism to redistribute out-of-domain content to better-matched edge servers, enabling efficient global knowledge utilization without central coordination. To handle domain-mismatched queries, SRAG employs lightweight cross-node routing guided by compact metadata summaries, avoiding full knowledge replication. Together, these mechanisms form an end-to-end workflow for decentralized edge RAG. Experiments show that SRAG improves retrieval relevance, generation quality, and storage efficiency, while reducing end-to-end latency.

Large Language Models (LLMs) increasingly underpin intelligent web applications, from chatbots to search and recommendation, where efficient specialization is essential. Low-Rank Adaptation (LoRA) enables such adaptation with minimal overhead, while federated LoRA allows web service providers to fine-tune shared models without data sharing. However, in privacy-sensitive deployments, clients inject varying levels of differential privacy (DP) noise, creating privacy heterogeneity that misaligns individual incentives and global performance. In this paper, we propose WinFLoRA, a privacy-heterogeneous federated LoRA that utilizes aggregation weights as incentives with noise awareness. Specifically, the noises from clients are estimated based on the uploaded LoRA adapters. A larger weight indicates greater influence on the global model and better downstream task performance, rewarding lower-noise contributions. By up-weighting low-noise updates, WinFLoRA improves global accuracy while accommodating clients' heterogeneous privacy requirements. Consequently, WinFLoRA aligns heterogeneous client utility in terms of privacy and downstream performance with global model objectives without third-party involvement. Extensive evaluations demonstrate that across multiple LLMs and datasets, WinFLoRA achieves up to 52.58% higher global accuracy and up to 2.56× client utility than state-of-the-art benchmarks. Source code is publicly available at https://github.com/koums24/WinFLoRA.git.

Multimodal Retrieval-Augmented Generation (MM-RAG) extends the capabilities of Large Language Models (LLMs) by incorporating external image-text knowledge bases to handle various tasks. However, MM-RAG systems in open environments are highly vulnerable to retrieval poisoning attacks, i.e., adversaries can inject malicious image-text pairs that are retrieved and dominate the generation process, leading to incorrect or harmful outputs. Due to the unique challenges of image-text fusion and cross-modal interference, existing defenses for text-based RAG cannot be directly applied to multimodal scenarios. In this paper, we propose IRAG, the first robust defense framework specifically designed for MM-RAG. The core of IRAG lies in its hazard separation. This structured defense isolates potential contamination sources by leveraging redundancy and consensus, enhancing system robustness and ensuring reliable outputs even when portions of retrieved content are compromised. Extensive experiments conducted under the MMQA and WebQA and the BQI and ROTI poisoning schemes demonstrate that IRAG consistently restores system reliability: the normal answer accuracy improves by 15–30% (restoring it to pre-poisoning levels), while the poisoned answer rate is reduced to below 7%.

A reliable autonomous driving system requires a high-precision perception module. Collaborative perception is emerging as a web-scale information-sharing paradigm for autonomous driving, enabling multiple vehicles to collectively achieve a broader perception field than any single vehicle. However, existing approaches necessitate frequent one-to-many communication, which increases network load and leads to information redundancy. This paper presents Octopus, an innovative vehicle-to-road collaboration framework that leverages the computational capabilities of roadside units. Instead of frequent one-to-many communication, vehicles interact only with roadside units, which significantly reduces communication overhead and improves real-time processing efficiency. While this design alleviates communication burdens, vehicles may still struggle to achieve comprehensive situational awareness in highly dynamic environments. To further address this limitation, our framework incorporates global fusion results as prior knowledge, enabling closed-loop fusion to refine vehicle-side perception. Extensive experiments on OPV2V and V2V4Real datasets demonstrate that Octopus excels at collaborative perception, outperforming the state-of-the-art approach up to 11.58% on [email protected], 12.74% on [email protected] and 5514× reduction in communication volume.

Embedding-based recommender systems rely on historical interactions to model users, which poses challenges for recommending to new users, known as the user cold-start problem. Some approaches incorporate social networks to deduce preferences based on the social circles of cold-start users to solve the problem of sparse features. However, such methods have difficulty distinguishing between superficial correlations and causal relationships in social behaviors, leading to inaccuracies in predicting user preferences. To address the aforementioned issues, we propose the Causal Contrastive Learning Recommendation (C2lRec) framework. Specifically, we causally model the inference of hidden preferences from the feature and historical behavior of warm users and predict user interactions based on such preferences. The counterfactual inference is subsequently performed to intervene and extract interactions from historical behaviors of warm users that influence their preferences, designating as primary causal variables. Additionally, we utilize the primary causal variables from users within the social circle of cold-start users to substitute the missing historical interactions of cold-start users and employ a similar causal modeling approach to uncover hidden preferences as we do with warm users. Finally, we realize causal contrastive learning to enhance the distribution of cold-start users. Extensive experiments conducted on three public datasets demonstrate that the recommendation performance of C2lRec exceeds that of state-of-the-art methods.

Sequential recommendation models generate embeddings for items through the analysis of historical user-item interactions and utilize the acquired embeddings to predict user preferences. Despite being effective in revealing personalized preferences for users, these models heavily rely on user-item interactions. However, due to the lack of interaction information, new users face challenges when utilizing sequential recommendation models for predictions, which is recognized as the cold-start problem. Recent studies, while addressing this problem within specific structures, often neglect the compatibility with existing sequential recommendation models, making seamless integration into existing models unfeasible.To address this challenge, we propose CMCLRec, a Cross-Modal Contrastive Learning framework for user cold-start RECommendation. This approach aims to solve the user cold-start problem by customizing inputs for cold-start users that align with the requirements of sequential recommendation models in a cross-modal manner. Specifically, CMCLRec adopts cross-modal contrastive learning to construct a mapping from user features to user-item interactions based on warm user data. It then generates a simulated behavior sequence for each cold-start user in turn for recommendation purposes. In this way, CMCLRec is theoretically compatible with any extant sequential recommendation model. Comprehensive experiments conducted on real-world datasets substantiate that, compared with state-of-the-art baseline models, CMCLRec markedly enhances the performance of conventional sequential recommendation models, particularly for cold-start users.

The global deployment of the 5G network has led to a substantial increase in the deployment of edge servers to host web applications, catering to the growing demand for low service latency by edge web users. Yet, running edge servers 24/7 leads to enormous energy consumption and excessive carbon emissions. Energy-efficient edge resource provision is desired to achieve sustainable development goals in the new multi-access edge computing (MEC) architecture. Recently, several approaches have been proposed to solve the demand response problem for energy saving in cloud computing and MEC. However, accurate location information of edge web users should always be provided, which sacrifices users' privacy. To protect edge web users' location privacy while saving energy in MEC, we systematically formulate this location privacy-preserving edge demand response (LEDR) problem. To solve the LEDR problem effectively and efficiently, we propose a system named GEES by incorporating differential geo-obfuscation to secure user privacy while maximizing system utility and energy efficiency through inferences with theoretical analysis. Extensive and comprehensive experiments are conducted based on a synthetic real-world dataset, and the results demonstrate that GEES outperforms representative approaches by 23.02%, 31.47%, and 17.29% on average in terms of energy efficiency, user privacy and system utility.