Unmanned aerial vehicles (UAVs), characterized by their agility, affordability, and flexible deployment, exhibit significant advantages in scenarios such as disaster monitoring, target tracking, and environmental data collection. However, the limited onboard energy of UAVs poses a major challenge for long-duration or large-scale missions. To address this issue, this paper proposes a dynamic trajectory planning framework for cooperative task search involving multiple UAVs. First, a UAV capability evaluation approach is developed to assess the relative performance of heterogeneous UAVs. Next, a density-aware clustering mechanism is employed to partition the search region based on spatial distance and regional density. After clustering, a sequential matching strategy is employed to assign UAVs with higher capabilities to larger or more complex task regions, ensuring efficient resource utilization. The problem is then formulated as a combinatorial optimization task, and a pointer network is designed to generate UAV trajectories. The network is trained using deep reinforcement learning to produce near-optimal paths, thereby minimizing the overall system cost. Experimental results confirm that the proposed method can substantially lower total task execution expenditure.

Modeling 3D open-vocabulary language fields is challenging yet highly anticipated. Despite great progress, existing approaches heavily rely on a large number of training views to construct language-embedded 3D scenes, which is unfortunately impractical in real-world scenarios. This paper introduces SOVGaussian, the first method for few-shot novel view open-vocabulary language querying. We introduce a depth-constrained neural language field to mitigate the geometry degradation caused by overfitting training views. Rather than straightforwardly using dense depth maps for loosely accurate supervision, Language-Aware Depth Distillation (LAD) based on open-vocabulary object masks is proposed, ensuring intra-object geometric accuracy within the language field. To further refine the language-geometry consistency of the language field, we propose a novel Language-Guided Outlier Pruning (LOP) strategy, which identifies floating 3D Gaussian primitives overfitting training views based on their language-grouped densities. Our comprehensive experiments demonstrate that SOVGaussian is able to reconstruct a superior scene representation from few-shot images, outperforming existing state-of-the-art methods and achieving significantly better performance on novel view language querying and synthesis.

Advanced driver assistance systems require a comprehensive understanding of the driver’s mental/physical state and traffic context but existing works often neglect the potential benefits of joint learning between these tasks. This paper proposes MMTL-UniAD, a unified multi-modal multitask learning framework that simultaneously recognizes driver behavior (e.g., looking around, talking), driver emotion (e.g., anxiety, happiness), vehicle behavior (e.g., parking, turning), and traffic context (e.g., traffic jam, traffic smooth). A key challenge is avoiding negative transfer between tasks, which can impair learning performance. To address this, we introduce two key components into the framework: one is the multi-axis region attention network to extract global context-sensitive features, and the other is the dual-branch multimodal embedding to learn multi-modal embeddings from both task-shared and task-specific features. The former uses a multi-attention mechanism to extract task-relevant features, mitigating negative transfer caused by task-unrelated features. The latter employs a dual-branch structure to adaptively adjust task-shared and task-specific parameters, enhancing cross-task knowledge transfer while reducing task conflicts. We assess MMTL-UniAD on the AIDE dataset, using a series of ablation studies, and show that it outperforms state-of-the-art methods across all four tasks. The code is available on https://github.com/Wenzhuo-Liu/MMTL-UniAD.

Bowen Wang, Haiyuan Wan, Liwen Shi, Chen Yang, Peng He, Yue Ma, Haochen Han, Wenhao Li, Tiao Tan, Yongjian Li, Fangming Liu, Gong Yifan, Sheng Zhang. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025.

Diffusion transformer (DiT) architecture catches much attention in image generation, which achieves better fidelity, performance, and diversity. However, most existing DiT-based image generation methods are global-aware synthesis and regional prompt control is less explored. In this paper, we propose a coarse-to-fine generation pipeline for regional prompt-following generation. Specifically, we first leverage the powerful large language model (LLM) to generate the high-level description of image (such as content, topic, and objects) and low-level description of image (such as details and style). Then we explore the influence of cross-attention layers in different depths. We discover that deeper layers always responsible for the high-level content control, while the shallow layers handles low-level content control. The various prompts are injected into the proposed regional cross-attention control in order for course-to-fine generation. Using the proposed pipeline, we improve the controllability of DiT-based image generation. Extensive quantitative and qualitative results demonstrate that our pipeline enables to improve the generated performance. Our codes are available at https://github.com/ZhenXiong-dl/ICASSP2025-RCAC.

Due to the birth of various new Internet of Things devices, the rapid increase of users, and the limited coverage of infrastructure, computing resources will inevitably become insufficient. Therefore, we consider an unmanned aerial vehicle (UAV)–assisted mobile edge computing system with multiple users, an edge server, a remote cloud server, and an UAV. A UAV, as a relay node, can provide users with extensive communications and certain computing capabilities. Our proposed scheme aims to optimize the unloading decision of the tasks among all users and the allocation of computing and communication resources to minimize overall energy consumption and costs of computing and maximum delay. To solve the joint optimization problem, we propose an efficient USS algorithm, which includes a UAV position optimization algorithm, semi-qualitative relaxation method, and self-adaptive adjustment method. Our numerical results show that the proposed algorithm can significantly reduce the unloading cost of multi-user tasks compared with four other unloading decisions, such as traditional cloud computing, which uses only the edge server.