Federated learning (FL) is a promising paradigm for collaborative intelligence in low-altitude economy, enabling unmanned aerial vehicle (UAV) swarms to perform deep learning tasks (e.g., logistics, emergency rescue) while preserving data sovereignty. However, limited communication channels and heterogeneous computation capabilities among UAVs cause significant FL aggregation delays. To reduce convergence time, we propose an asynchronous FL (AFL) framework for UAV swarms, integrating over-the-air computation (AirComp) to boost communication efficiency via simultaneous transmission. To address signal distortion in AirComp, we design an objective function solved via an aggregation scheduling algorithm, which transforms the nonconvex problem into two convex subproblems tackled via alternating optimization, to derive optimal aggregation strategies and beamforming vectors. Besides, to mitigate model staleness in AFL, which causes gradient divergence and slow convergence, we propose a self-adaptive aggregation scheme with staleness awareness, enabling UAVs to adjust local models autonomously without information from other UAVs. Simulation results show that our scheme not only reduces staleness impact but also leverages stale parameters, helping AFL outperform synchronous FL in convergence speed and accuracy. Overall, our study presents an effective AFL framework, a fast aggregation scheduling algorithm, and a self-adaptive aggregation scheme for UAV swarms, accelerating global model convergence while reducing energy expenditure.

Unmanned Aerial Vehicle (UAV) swarms are key enablers for cooperative intelligent tasks in Internet of Everything (IoE) environments, like environmental monitoring and disaster response. Meanwhile, Digital Twin (DT) technology is rapidly advancing in academia and industry due to its potential to design, test, and optimize evolving IoE networks. While DT-empowered UAV swarms provide a compelling solution for complex tasks, challenges remain in jointly designing swarm control and service provision strategies, especially in dynamic IoE environments. To harness the potential of UAV swarm intelligence, we propose a DT-empowered cluster formation technique employing the Over-the-Air Computation (AirComp) transmission scheme for energy-efficient IoE data aggregation, where UAV and IoE device clusters are dynamically constructed and adjusted. Furthermore, we consider the limited power resources by maximizing energy efficiency, which leads to a joint cluster formation, multi-UAV trajectory design, and power allocation problem under the constraint of AirComp signal distortion. To efficiently solve this mixed-integer nonlinear fractional programming problem, we develop a low-complexity iterative algorithm based on the block coordinate descent method. Simulation results show that the proposed DT-empowered AirComp cluster formation technique achieves up to 42% higher energy efficiency and 34% greater throughput than the baselines, validating its potential for efficient data aggregation tasks.

Facial expressions can provide a better understanding of people’s mental status and attitudes towards specific things. However, facial occlusion in real world is an unfavorable phenomenon that greatly affects the performance of facial expression recognition models. Recent works addressing the occlusion problem have primarily relied on attention mechanisms or occlusion discarding methods that focus on non-occluded regions of the face. However, these methods have not achieved a good balance between occlusion robustness and model efficiency. In this paper, we propose a simple and efficient model, called FERMixNet, for occluded facial expression recognition. The model incorporates a novel facial mixing augmentation strategy (FERMix) that generates new training samples by simulating real-world facial occlusion and preserving high expression-related semantic information. By co-training the original and newly generated samples, the model’s occlusion robustness is improved without increasing its complexity during inference. Additionally, to further enhance the model’s occlusion robustness, we include mid-level representation learning in the network to learn the discriminative non-occluded local features of the samples with low computational cost. Extensive experiments on four public facial occlusion datasets: Occlusion-RAF-DB, Occlusion-FERPlus and FED-RO show that the proposed model achieves state-of-the-art results which demonstrates the good robustness of our method for occluded facial expression recognition. Meanwhile, the proposed model also achieves state-of-the-art results on the in-the-wild facial expression datasets RAF-DB, AffectNet-8, and AffectNet-7. It proves that the proposed model has good application prospects in real world.

Facial Expression Recognition (FER) is crucial for accurately understanding human emotions and intentions, making it a fundamental task in computer vision. Some studies have shown differences in facial expression patterns across different age groups. However, existing research on FER often overlooks the impact of age differences on model performance, leading to significant age bias issues and negatively affecting the overall performance of the models. To address this problem, we propose a FER framework based on Age-Group Expression Feature Learning (AEFL). Firstly, we adopt an age-group distribution guidance multi-branch to extract age-group expression features, thereby preserving the unique expression information within each age group. Next, we capture shared expression information among different age groups by learning global features from samples across all age groups. Lastly, we introduce an age-guided attention mechanism to enhance the representation of age-group information in the global features, mitigating the suppression of disadvantaged age-group information by advantaged age-group information in the global features. Experimental results demonstrate that our method significantly improves the recognition performance of disadvantaged age groups without compromising the performance of advantaged age groups, thus enhancing the overall model performance. Moreover, our method achieves state-of-the-art results on multiple large-scale facial expression benchmarks (RAF-DB, AffectNet-7, AffectNet-8).

Evolutionary multi-objective optimization (EMO) algorithms have been demonstrated to be effective in solving multi-criteria decision-making problems. In real-world applications, analysts often employ several algorithms concurrently and compare their solution sets to gain insight into the characteristics of different algorithms and explore a broader range of feasible solutions. However, EMO algorithms are typically treated as black boxes, leading to difficulties in performing detailed analysis and comparisons between the internal evolutionary processes. Inspired by the successful application of visual analytics tools in explainable AI, we argue that interactive visualization can significantly enhance the comparative analysis between multiple EMO algorithms. In this paper, we present a visual analytics framework that enables the exploration and comparison of evolutionary processes in EMO algorithms. Guided by a literature review and expert interviews, the proposed framework addresses various analytical tasks and establishes a multi-faceted visualization design to support the comparative analysis of intermediate generations in the evolution as well as solution sets. We demonstrate the effectiveness of our framework through case studies on benchmarking and real-world multi-objective optimization problems to elucidate how analysts can leverage our framework to inspect and compare diverse algorithms.

Owing to its ability to offer collaborative data utilization while ensuring data privacy, federated learning (FL) provides a promising paradigm to enable cooperative intelligent tasks across multiple low-earth orbit (LEO) satellites, such as carbon estimation, traffic surveillance, and forest fire detection. Although the advantages of pushing intelligence to satellites are multi-fold, limited communication channels along with the rigid global model aggregation conditions result in dramatic convergence delays. In order to reduce the convergence time, we propose an asynchronous FL framework in LEO satellite networks by exploiting multiple high-altitude platforms for model aggregation, where the advanced over-the-air computation (AirComp) transmission scheme is utilized for the sake of further reducing energy consumption. Considering the practical constraint of AirComp signal distortion, the objective function of optimizing FL performance is carefully formulated and solved by the proposed quantity-quality jointed linkage search algorithm. Simulation results demonstrate that our proposed asynchronous FL framework outperforms the conventional synchronous FL framework by a decline of 30.07% in convergence time at most. It also provides an average increase of 110% and 580%, respectively, in terms of throughput and energy efficiency in all scenarios considered. Overall, our study presents a beneficial asynchronous FL framework and a fast aggregation scheduling algorithm in LEO satellite networks, accelerating the convergence of the global model with reduced energy expenditure.

Facial expression recognition has received extensive attention in recent years due to its important applications in many fields. Most expression samples used in research are relatively easy to analyze emotions because they have explicit expressions with strong intensities. However, in situations such as video question and answer, business negotiation, polygraph detection in the security field, autism treatment and medical escort, emotions are expressed in suppressed manners with low intensive expression or subtle expressions, making it difficult to estimate emotions accurately. In these situations, how to effectively extract expression features from facial expression images is a critical problem that affects the accuracy of subtle expression recognition. To address this problem, we propose an end-to-end group adaptive attention model for subtle expression recognition. Cropping an image into several regions of interest (ROI) according to the correlations between facial skeleton and emotions, the proposed model analyses the relationship among regions of interest, and mutual relations between local regions and the holistic region. Using the region group adaptive attention mechanism, the model effectively trains the convolutional neural network to efficiently extract facial expressions representing features and increases the accuracy and robustness of the recognition, particularly in some subtle facial expression circumstances. To improve the ability of different regional features to discriminate expressions, a group adaptive loss function is introduced to verify and improve estimation accuracy. Extensive experiments are conducted on the existing public face datasets CK+, JAFFE, KDEF and the self-collected subtle expression dataset SFER. Results show that the proposed model achieves accuracies of 99.59%, 95.20%, and 93.47% with datasets CK+, JAFFE, and KDEF, respectively. The proposed model thus generally achieves better performance in facial expression recognition than other methods.

Damaged road extraction is a challenging task in the field of remote sensing. Some existing methods include the step to extract road from pre- and post-disaster remote sensing images of the same area. In practice, it often occurs that one of these two images is missing. To solve this problem, we use CoCosNet, the model for exemplar-based image translation, to translate pre-disaster images to simulated post-disaster ones. Then we use D-LinkNet, the state-of-the-art method in road extraction, to extract road from the pre- and post-disaster images of the same area. We extract damaged road area by comparing pre-disaster road masks with post-disaster ones and output the damage level by calculating the proportion of the damaged road area. Finally, we evaluate the damaged road extraction accuracy. Experimental results on simulated post-disaster images prove the effectiveness of the simulation method and the framework for damaged road extraction and damage level evaluation.