Retrieval-Augmented Generation (RAG) systems retrieve relevant documents from a corpus database to mitigate issues like hallucination, outdated knowledge, and limited domain coverage. While enhancing large language models (LLMs) performance, RAG also introduces a new attack surface: adversaries can inject trigger-embedded malicious documents into the corpus database, potentially causing the LLM to produce attacker-controlled outputs.

Incomplete scenario is a prevalent, practical, yet challenging setting in Multimodal Recommendations (MMRec), where some item modalities are missing due to various factors. Recently, a few efforts have sought to improve the recommendation accuracy by exploring generic structures from incomplete data. However, two significant gaps persist: 1) the difficulty in accurately generating missing data due to the limited ability to capture modality distributions; and 2) the critical but overlooked visibility bias, where items with missing modalities are more likely to be disregarded due to the prioritization of items' multimodal data over user preference alignment. This bias raises serious concerns about the fair treatment of items. To bridge these two gaps, we propose a novel Modality-Diffused Counterfactual (MoDiCF) framework for incomplete multimodal recommendations. MoDiCF features two key modules: a novel modality-diffused data completion module and a new counterfactual multimodal recommendation module. The former, equipped with a particularly designed multimodal generative framework, accurately generates and iteratively refines missing data from learned modality-specific distribution spaces. The latter, grounded in the causal perspective, effectively mitigates the negative causal effects of visibility bias and thus assures fairness in recommendations. Both modules work collaboratively to address the two aforementioned significant gaps for generating more accurate and fair results. Extensive experiments on three real-world datasets demonstrate the superior performance of MoDiCF in terms of both recommendation accuracy and fairness. The code and processed datasets are released at https://github.com/JinLi-i/MoDiCF.

With the increasing prevalence of Web-based platforms handling vast amounts of user data, machine unlearning has emerged as a crucial mechanism to uphold users' right to be forgotten, enabling individuals to request the removal of their specified data from trained models. However, the auditing of machine unlearning processes remains significantly underexplored. Although some existing methods offer unlearning auditing by leveraging backdoors, these backdoor-based approaches are inefficient and impractical, as they necessitate involvement in the initial model training process to embed the backdoors. In this paper, we propose a TAilored Posterior diffErence (TAPE) method to provide unlearning auditing independently of original model training. We observe that the process of machine unlearning inherently introduces changes in the model, which contains information related to the erased data. TAPE leverages unlearning model differences to assess how much information has been removed through the unlearning operation. Firstly, TAPE mimics the unlearned posterior differences by quickly building unlearned shadow models based on first-order influence estimation. Secondly, we train a Reconstructor model to extract and evaluate the private information of the unlearned posterior differences to audit unlearning. Existing privacy reconstructing methods based on posterior differences are only feasible for model updates of a single sample. To enable the reconstruction effective for multi-sample unlearning requests, we propose two strategies, unlearned data perturbation and unlearned influence-based division, to augment the posterior difference. Extensive experimental results indicate the significant superiority of TAPE over the state-of-the-art unlearning verification methods, at least 4.5x efficiency speedup and supporting the auditing for broader unlearning scenarios.

In unsupervised graph anomaly detection, existing methods usually focus on detecting outliers by learning local context information of nodes, while often ignoring the importance of global context. However, global context information can provide more comprehensive relationship information between nodes in the network. By considering the structure of the entire network, detection methods are able to identify potential dependencies and interaction patterns between nodes, which is crucial for anomaly detection. Therefore, we propose an innovative graph anomaly detection framework, termed CoCo (Context Correlation Discrepancy Analysis), which detects anomalies by meticulously evaluating variances in correlations. Specifically, CoCo leverages the strengths of Transformers in sequence processing to effectively capture both global and local contextual features of nodes by aggregating neighbor features at various hops. Subsequently, a correlation analysis module is employed to maximize the correlation between local and global contexts of each normal node. Unseen anomalies are ultimately detected by measuring the discrepancy in the correlation of nodes’ contextual features. Extensive experiments conducted on six datasets with synthetic outliers and five datasets with organic outliers have demonstrated the significant effectiveness of CoCo compared to existing methods.

Most existing methods for visual domain adaptation need to convert high-order tensors into one-order high-dimensional vectors through naive vectorization operations. However, they not only destroy the internal spatial structure within the original high-order tensors, but also result in exponentially increasing model parameters. To address these problems, we propose a novel method for visual domain adaptation by representing tensorial features in tensor-train subspace in this paper. Specifically, we firstly provide a theoretical deduction by constructing a tensor-train subspace and proving its linearity and left-orthogonality. Secondly, to extract common tensorial features between source and target domains, we formulate the visual domain adaptation problem into an optimization problem that models the aforementioned common tensor-train subspace between two domains, as well as their corresponding projections. Thirdly, we design a tensor-train subspace representation algorithm (TTSR) to solve the multiple variables optimization problem by optimizing its sub-problems iteratively, so as to process high-order tensorial features. Finally, we evaluate the performance of our proposed TTSR algorithm by conducting extensive experiments on three popular public datasets. The experimental results demonstrate that the TTSR algorithm can improve the classification accuracy of unlabeled target domain than that of baseline algorithms.

Federated learning (FL) is an emerging paradigm using a parameter server (PS) to coordinate multiple decentralized clients for training a common model without exposing their raw data. Despite its amazing capability in preserving data privacy, FL confronts two significant challenges that have not been sufficiently addressed by existing works, which are: 1) heterogeneous data distributed on clients given that the PS cannot alter data locations. 2) limited computation resources of FL clients who may conduct model training with mobile devices. To tackle these challenges, we propose a novel federated one-shot transferring of intermediate features (FedOTF) algorithm. More specifically, FedOTF consists of two stages: feature extraction and model reconstruction. To overcome challenge 1, clients in stage 1 only collaboratively train a small model in a federated manner, with the objective to extract features. In stage 2, intermediate features (generated by the small model trained by stage 1) are transferred from clients to the PS so that a large model can be trained to overcome challenge 2. In addition, the design of FedOTF is robust, which can flexibly diminish the number of communication rounds of stage 1 when network capacity is limited, and reduce the amount of exposed features when privacy is concerned. To verify the superiority of FedOTF, we conduct comprehensive experiments with real datasets. The experiment results demonstrate that FedOTF can significantly improve the model utility of FL because a better model can be finally obtained by the PS without incurring heavy computational load on clients. Besides, we conduct robustness evaluation of FedOTF which can achieve stable performance when varying network capacity and privacy requirement.

Federated learning (FL) is a machine learning framework for collaborative data mining in many scenarios (e.g. Internet of Things) due to its privacy-preserving feature. However, various attacks arise security concerns of FL, such as poisoning, backdoor, and DDoS attacks. Several blockchain-based FL schemes strengthen credibility and security without considering the increased communication overhead. Some existing work compresses local updated gradients to sign vectors to lower communication overhead at the expense of model accuracy. To address the above concerns, this paper offers a blockchain-based asynchronous SignSGD (BASS) scheme. A novel asynchronous sign aggregation algorithm is introduced to ensure model accuracy even if the local updated gradients are compressed to sign vectors. Considering the unstable network connection on IoT, a consensus algorithm that elects multiple leader nodes enables reliable global model aggregation. The introduced blockchain improves credibility and security without downgrading efficiency. Empirical studies show that BASS outperforms other schemes in efficiency, model accuracy, and security.

The widespread use of mobile devices propels the development of new-fashioned video applications like 3D (3-Dimensional) stereo video and mobile cloud game via web or App, exerting more pressure on current mobile access network. To address this challenge, we adopt the crowdsourcing paradigm to offer some incentive for guiding the movement of recruited crowdsourcing users and facilitate the optimization of the movement control decision. In this paper, based on a practical 4G (4th-Generation) network throughput measurement study, we formulate the movement control decision as a cost-constrained user recruitment optimization problem. Considering the intractable complexity of this problem, we focus first on a single crowdsourcing user case and propose a pseudo-polynomial time complexity optimal solution. Then, we apply this solution to solve the more general problem of multiple users and propose a graph-partition-based algorithm. Extensive experiments show that our solutions can improve the efficiency of real-time D2D communication for mobile videos.

The low-rank matrix completion problem is fundamental in both machine learning and computer vision fields with many important applications, such as recommendation system, motion capture, face recognition, and image inpainting. In order to avoid solving the rank minimization problem which is NP-hard, several surrogate functions of the rank have been proposed in the literature. However, the matrix restored from the optimization problem based on the existing surrogate functions seriously deviates from the original one. In this paper, we first design a new non-convex Schatten capped$p$norm which generalizes several existing non-convex matrix norms and balances between the rank and the nuclear norm of the matrix. Then, a matrix completion method based on the Schatten capped$p$norm is proposed by exploiting the framework of the alternating direction method of multipliers. Meanwhile, the Schatten capped$p$norm regularized least squares subproblem is analyzed in detail and is solved explicitly. Finally, we evaluate the performance of the proposed matrix completion method based on extensive experiments in the field of image inpainting. All the experimental results demonstrate that the proposed method can indeed improve the accuracy of matrix completion compared with the existing methods.

Twitter sentiment analysis has become a hot research topic in recent years. Most of existing solutions to Twitter sentiment analysis basically only consider textual information of Twitter messages, and struggle to perform well when facing short and ambiguous Twitter messages. Recent studies show that sentiment diffusion patterns on Twitter have close relationships with sentiment polarities of Twitter messages. Therefore, in this paper, we focus on how to fuse textual information of Twitter messages and sentiment diffusion patterns to obtain better performance of sentiment analysis on Twitter data. To this end, we first analyze sentiment diffusion by investigating a phenomenon called sentiment reversal, and find some interesting properties of sentiment reversals. Then, we consider the inter-relationships between textual information of Twitter messages and sentiment diffusion patterns, and propose an iterative algorithm called SentiDiff to predict sentiment polarities expressed in Twitter messages. To the best of our knowledge, this work is the first to utilize sentiment diffusion patterns to help improve Twitter sentiment analysis. Extensive experiments on real-world dataset demonstrate that compared with state-of-the-art textual information based sentiment analysis algorithms, our proposed algorithm yields PR-AUC improvements between 5.09 and 8.38 percent on Twitter sentiment classification tasks.

Malware is pervasive in networks, and poses a critical threat to network security. However, we have very limited understanding of malware behavior in networks to date. In this paper, we investigate how malware propagates in networks from a global perspective. We formulate the problem, and establish a rigorous two layer epidemic model for malware propagation from network to network. Based on the proposed model, our analysis indicates that the distribution of a given malware follows exponential distribution, power law distribution with a short exponential tail, and power law distribution at its early, late and final stages, respectively. Extensive experiments have been performed through two real-world global scale malware data sets, and the results confirm our theoretical findings.

In this paper, we propose a novel model for web-based database systems based on the multicast and anycast' protocols. In the model, we design a middleware, castway, which locates between the database server and the Web server. Every castway in a distributed system operates as a multicast node and an anycast node independently, respectively. The proposed mechanism can balance the workload among the distributed database servers, and offers the "best" server to serve for a query. Three algorithms are employed for the model: the requirement-based probing algorithm for anycast routing, the atomic multicast update algorithm for database synchronization, and the job deviation algorithm for system workload balance. The simulations and experiments show that the proposed model works very well.