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%.

As an emerging distributed computing paradigm, edge computing (EC) brings computing and storage capabilities to the network edge to support low-latency services. Service providers can cache popular data on edge servers to greatly reduce service latency and improve service quality. However, data popularity in EC environments is often dynamic and stochastic. Given the limited storage resources of edge servers, when data popularity changes, data may need to be migrated from edge servers with lower data popularity to those with higher data popularity. Data migration between edge servers can effectively reduce the high transmission costs associated with transferring data from the cloud to the edge. Most existing data migration approaches are designed for static EC environments, assuming that user access patterns and data demands remain constant over time. Even in dynamic EC environments, many approaches oversimplify data popularity by treating it as static, failing to model its fluctuations over time. In realistic EC scenarios, unexpected situations, such as sudden shifts in user preferences for specific data at certain times, can cause significant fluctuations in data popularity, leading to increased data retrieval latency and inefficient storage management. To address this uncertainty, this paper proposes a popularity uncertainty-aware online data migration approach that combines distributed robust optimization and Lyapunov optimization to minimize the data retrieval latency and migration cost for service providers. Specifically, it handles data popularity uncertainty in a data-driven manner through robust optimization. Then, it employs Lyapunov optimization to decompose the continuous optimization problem into multiple single-slot online optimization problems. Extensive experiments on a widely used real-world dataset confirm the effectiveness of this approach and its significant advantages over state-of-the-art approaches.

Path planning is a basic requirement for Autonomous Underwater Vehicles (AUVs) to accomplish underwater missions. However, previous studies often have limitations, such as ignoring the basic condition that the AUV operates in an ocean current environment and discretizing its actions without considering the action space, which results in the simulation being far from the actual situation. To solve the above problems, this paper proposes a method of using a Parametrized Deep Q-Network (PDQN) to output hybrid actions for path planning, which can output a hybrid action space based on the current local observation, flexibly avoid obstacles under limited sensor observations, and realize the refined operation of AUV actions. According to the setup of the simulation environment, the AUV needs to visit multiple target nodes underwater and decelerate within the communication range of the nodes to have enough time to communicate with the nodes. The PDQN enables the AUV to easily learn the connection between the current state and discrete actions. It outputs the corresponding continuous actions based on the current discrete actions, which realizes a time-saving strategy of accelerating and then decelerating among the nodes. Meanwhile, we also utilize the actual current data and terrain data to restore the simulation environment as accurately as possible, and the simulation results prove the superiority and robustness of the algorithm.