Swapping and purification are the two fundamental building blocks for high-fidelity entanglement distribution in multi-hop quantum networks. Unfortunately, it is still a mystery how they intertwine with each other to affect the fidelity and cost of end-to-end entanglements. Current scheduling algorithms consider this problem under relatively limited assumptions and a critical yet unjustified conjecture. In this work, we first consider more general assumptions with operation failures and, accordingly, extend a tree-based modeling for joint swapping and purification. Then, we analytically prove the previous conjecture that the optimal strategy underBinary systemis always to purify the entanglements before any swapping. This sheds light on the protocol and device design for quantum networks. We then further propose a tree-based algorithm, which can efficiently schedule swapping and purification along a path for bothBinaryandWerner systems. Extensive simulations of the proposed method against state-of-the-art solutions show that our method uses fewer entanglements to establish qualified end-to-end entanglements, and thus achieves higher network throughput.

Space-air-ground integrated networks can provide better connectivity and improved performance to ground users and thus have been well-recognized as the innovative trend for future 6G and beyond wireless systems. However, such heterogeneous and complex integrated communication systems also pose many new challenges for joint performance optimization. In this paper, we study the data delivery optimization problem in a space-air-ground network, where joint resource management decisions on user association, bandwidth allocation, cache placement, and high-altitude platform (HAP) location selection have to be optimized. To tackle this complex and challenging mixed integer programming optimization problem, we introduce a quantum-assisted method, named the Hybrid quantum-classical Benders’ Decomposition (HyBD) algorithm, which leverages the strengths of both quantum and classical computing. Experiments on the commercial quantum annealing machine demonstrate the effectiveness and robustness of the proposed HyBD method, with up to 64.3% improvement of iteration number and 82.8% improvement of average computation time over the classical Benders’ Decomposition algorithm on classical CPUs even at small scales, which demonstrates the quantum advantage.

Incentivizing participation and coordinating decisions across hierarchical agents remain critical challenges in blockchain-enabled federated learning (FL) over edge clouds, especially when a coded FL is performed over a client-edge-cloud hierarchical system. This paper proposes a novel hybrid incentive framework that integrates multidimensional contract theory with reinforcement learning (RL)-based Stackelberg game modeling for such a system. Specifically, we design personalized contracts between edge servers and clients, addressing their heterogeneous data volume, privacy sensitivity, and computational capacity under incomplete information. Simultaneously, we model the task publisher's reward allocation to edge servers as a one-leader multi-follower Stackelberg game, where each follower acts based on local observations. A decentralized RL algorithm is proposed to learn optimal reward strategies without revealing other agents' private information, such as local data volume/quality. Simulations demonstrate that our method can converge to equilibrium and achieve effectiveness under incomplete information compared to baseline incentive schemes.

Swapping and purification are the two fundamental building blocks for multi-hop quantum networks. However, their interplay and its impact on end-to-end fidelity and cost are not yet fully explored. Existing scheduling algorithms address this problem under certain simplified assumptions and models that may not fully capture the complexities of real scenarios. In this work, we first consider more general assumptions that account for operation failures and extend a tree-based modeling approach for joint swapping and purification. Then, for the first time, we analytically prove the previous conjecture that the optimal strategy under Binary system is always to purify the entanglements before any swapping. This sheds light on the protocol and device design for entanglement distribution in quantum networks. We then further propose a tree-based algorithm, which can efficiently schedule swapping and purification along a path for both Binary and Werner systems. Extensive simulations have been conducted to evaluate the proposed method against the existing solutions, and the results show that our method uses fewer entanglements to establish qualified end-to-end entanglements and thus achieves higher network throughput.

Recent advancements have explored text-to-image diffusion models for synthesizing out-of-distribution (OOD) samples, substantially enhancing the performance of OOD detection. However, existing approaches typically rely on perturbing text-conditioned embeddings, resulting in semantic instability and insufficient shift diversity, which limit generalization to realistic OOD. To address these challenges, we propose GOOD, a novel and flexible framework that directly guides diffusion sampling trajectories towards OOD regions using off-the-shelf in-distribution (ID) classifiers. GOOD incorporates dual-level guidance: (1) Image-level guidance based on the gradient of log partition to reduce input likelihood, drives samples toward low-density regions in pixel space. (2) Feature-level guidance, derived from k-NN distance in the classifier’s latent space, promotes sampling in feature-sparse regions. Hence, this dual-guidance design enables more controllable and diverse OOD sample generation. Additionally, we introduce a unified OOD score that adaptively combines image and feature discrepancies, enhancing detection robustness. We perform thorough quantitative and qualitative analyses to evaluate the effectiveness of GOOD, demonstrating that training with samples generated by GOOD can notably enhance OOD detection performance.

Reconstructing speech envelopes from EEG signals is essential for exploring neural mechanisms underlying speech perception. Yet, EEG variability across subjects and physiological artifacts complicate accurate reconstruction. To address this problem, we introduce Subject Disentangling Neural Network (SDN-Net), which disentangles subject identity information from reconstructed speech envelopes to enhance cross-subject reconstruction accuracy. SDN-Net integrates three key components: MLA-Codec, MPN-MI, and CTA-MTDNN. The MLA-Codec, a fully convolutional neural network, decodes EEG signals into speech envelopes. The CTA-MTDNN module, a multi-scale time-delay neural network with channel and temporal attention, extracts subject identity features from EEG signals. Lastly, the MPN-MI module, a mutual information estimator with a multilayer perceptron, supervises the removal of subject identity information from the reconstructed speech envelope. Experiments on the Auditory EEG Decoding Dataset demonstrate that SDN-Net achieves superior performance in inner- and cross-subject speech envelope reconstruction compared to recent state-of-the-art methods.

The scarcity of speaker-annotated far-field speech presents a significant challenge in developing high-performance far-field speaker verification (SV) systems. While data augmentation using large-scale near-field speech has been a common strategy to address this limitation, the mismatch in acoustic environments between near-field and far-field speech significantly hinders the improvement of far-field SV effectiveness. In this paper, we propose an adaptive speech augmentation approach leveraging NaturalSpeech3, a pre-trained foundation text-to-speech (TTS) model, to convert near-field speech into far-field speech by incorporating far-field acoustic ambient noise for data augmentation. Specifically, we utilize FACodec from NaturalSpeech3 to decompose the speech waveform into distinct embedding subspaces —content, prosody, speaker, and residual (acoustic details) embeddings—and reconstruct the speech waveform from these disentangled representations. In our method, the prosody, content, and residual embeddings of far-field speech are combined with speaker embeddings from near-field speech to generate augmented pseudo far-field speech that maintains the speaker identity from the out-domain near-field speech while preserving the acoustic environment of the in-domain far-field speech. This approach not only serves as an effective strategy for augmenting training data for far-field speaker verification but also extends to cross-data augmentation for enrollment and test speech in evaluation trials. In augmentation of enrollment and test utterances, the method mitigates performance degradation caused by discrepancies in text content or environmental noise between enrollment and test data. This data augmentation method, which preserves the acoustic environment of the in-domain far-field data, qualifies as an adaptive augmentation method. Experimental results on FFSVC demonstrate that the adaptive data augmentation method significantly outperforms traditional approaches, such as random noise addition and reverberation, as well as other competitive data augmentation strategies.

Numerous optimization scenarios such as industrial production planning, network communication routing, and logistic scheduling can be modeled as large-scale integer linear programming problems. However, due to the NP-Hardness of these problems, it is very challenging to optimally solve these problems in a short time on classical computers. Quantum computers have emerged as a new computing platform to provide new computing paradigms to tackle these problems. However, the scalability and efficiency of current quantum computers pose significant challenges in practical implementations of quantum optimization algorithms. In this paper, we propose a novel hybrid quantum-classical approach, termed Hybrid quantum-classical Dantzig-Wolfe Decomposition (HyDWD), aimed at solving these problems. In this framework, the subproblems can be solved in parallel on quantum computers. Our results demonstrate the benefits of integrating parallel quantum computing with the proposed hybrid quantum-classical framework via Dantzig-Wolfe decomposition, paving the way for advancements in optimization and decision-making processes.

Adversarial examples are notably characterized by their strong transferability, allowing attackers to craft these examples on their models and subsequently deploy them against other models. This poses significant threats to existing deep learning systems and raises substantial security concerns. While several methods have been developed to enhance transferability, ensemble attacks stand out for their effectiveness. However, previous approaches of ensemble attacks typically rely on simple averaging of logits, probabilities, or losses for ensembling, without delving into the underlying reasons for the improved transferability. In this work, we propose a new approach that makes full use of the information of each surrogate model by regularizing the optimization direction to concurrently attack all surrogate models. This is achieved by promoting cosine similarity between their gradients. Extensive experiments conducted on the ImageNet dataset demonstrate the superior efficacy of our method in enhancing adversarial transferability. Notably, our approach outperforms leading state-of-the-art attackers across 18 discriminative classifiers and adversarially trained models, underscoring its potential in this domain.

Topology is one of the most critical properties of networks. Quantum networks, as a new type of network, have fundamentally different principles for establishing connections compared to classical networks, leading to distinct challenges in topology design. Finding the optimal topology for quantum networks to meet traffic demands is a crucial yet not fully understood problem. In this paper, we explore the topology design problem for quantum networks, considering both resource allocation and entanglement distribution. We propose and investigate both flow-based and path-based formulations, along with their associated solutions, aimed at minimizing the topology cost. For the path-based formulation, we also provide the first theoretical analysis of the cost associated with swapping strategies over a quantum path. Extensive simulations demonstrate that our enhanced path-based formulation is both efficient and effective.

In a blockchain-based federated learning (FL) framework, clients can contribute private data or computing resources to the overall FL training or mining task. To overcome the impractical assumption that participants will voluntarily join training or mining, it is crucial to design an incentive mechanism that motivates participants to achieve optimal training and mining outcomes. In this paper, we investigate the incentive mechanism design for a semi-asynchronous blockchain-based FL system. We model the resource pricing mechanism among clients and task publishers as a Stackelberg game, and prove the existence and uniqueness of a Nash equilibrium in such a game. We then propose an iterative algorithm based on the Alternating Direction Method of Multipliers (ADMM) to achieve the optimal strategies for each participant. Finally, our simulation results verify the convergence and efficiency of our proposed scheme.

Hierarchical federated learning has emerged as a pragmatic approach to addressing scalability, robustness, and privacy concerns within distributed machine learning, particularly in the context of edge computing. This hierarchical method involves grouping clients at the edge, where the constitution of client groups significantly impacts overall learning performance, influenced by both the benefits obtained and costs incurred during group operations (such as group formation and group training). This is especially true for edge and mobile devices, which are more sensitive to computation and communication overheads. The formation of groups is critical for group-based hierarchical federated learning but often neglected by researchers, especially in the realm of edge systems. In this paper, we present a comprehensive exploration of a group-based federated edge learning framework utilizing the hierarchical cloud-edge-client architecture and employing probabilistic group sampling. Our theoretical analysis of its convergence rate, considering the characteristics of client groups, reveals the pivotal role played by group heterogeneity in achieving convergence. Building on this insight, we introduce new methods for group formation and group sampling, aiming to mitigate data heterogeneity within groups and enhance the convergence and overall performance of federated learning. Our proposed methods are validated through extensive experiments, demonstrating their superiority over current algorithms in terms of prediction accuracy and training cost.

Deciphering the intricate neural mechanisms behind human speech perception mandates the seamless alignment of auditory signals with electroencephalogram (EEG) data. Yet, the unpredictable fluctuations and inherent noise within EEG, intertwined with the delicate dance between sound and neural response, cast a daunting challenge before researchers. Within the realm of this challenge, our study presents a groundbreaking iteration of the Conformer architecture, meticulously fine-tuned for the exacting task of audio-EEG matching.Where the conventional Conformer pivots on a self-attention mechanism, our enhanced version echoes a more nuanced tune, introducing what we’ve termed "local attention." This deliberate design choice emphasizes nearby inputs, capturing those elusive local patterns and relationships pivotal to understanding the spatial interplay between auditory and EEG features. Such a precision-focused approach guides the model to spotlight vital data sectors during its training voyage, unearthing features rich in meaning and contextual relevance.Our empirical symphony resonates with the brilliance of this enhanced Conformer. It consistently surpasses traditional methods, echoing superior prowess in audio-EEG matching accuracy, and in turn, amplifying the resonance of our innovative approach.

Electroencephalography (EEG) plays a vital role in detecting how brain responses to different stimulus. In this paper, we propose a novel Shallow-Deep Attention-based Network (SDANet) to classify the correct auditory stimulus evoking the EEG signal. It adopts the Attention-based Correlation Module (ACM) to discover the connection between auditory speech and EEG from global aspect, and the Shallow-Deep Similarity Classification Module (SDSCM) to decide the classification result via the embeddings learned from the shallow and deep layers. Moreover, various training strategies and data augmentation are used to boost the model robustness. Experiments are conducted on the dataset provided by Auditory EEG challenge (ICASSP Signal Processing Grand Challenge 2023). Results show that the proposed model has a significant gain over the baseline on the match-mismatch track.

Hierarchical federated learning has been studied as a more practical approach to federated learning in terms of scalability, robustness, and privacy protection, particularly in edge computing. To achieve these advantages, operations are typically conducted in a grouped manner at the edge, which means that the formation of client groups can affect the learning performance, such as the benefits gained and costs incurred by group operations. This is especially true for edge and mobile devices, which are more sensitive to computation and communication overheads. The formation of groups is critical for group-based federated edge learning but has not been studied in detail, and even been overlooked by researchers. In this paper, we consider a group-based federated edge learning framework that leverages the hierarchical cloud-edge-client architecture and probabilistic group sampling. We first theoretically analyze the convergence rate with respect to the characteristics of the client groups, and find that group heterogeneity plays an important role in the convergence. Then, on the basis of this key observation, we propose new group formation and group sampling methods to reduce data heterogeneity within groups and to boost the convergence and performance of federated learning. Finally, our extensive experiments show that our methods outperform current algorithms in terms of prediction accuracy and training cost.

Motor imagery (MI) based on Electroencephalogram (EEG) analysis is a common used paradigm in BCI. Previous works to classify MI have obtained promising classification results. However, there still exit some challenges: 1) Although the deep learning (DL) models are the mainstream methods to solve MI classification, with their depths increasing, the accuracy gets saturated then degrades rapidly. 2) The complex and changeable relationship of the time sequence in the EEG signals makes it difficult to model. In this paper, we propose REEG-BTCNet, a novel framework that achieves outstanding accuracy with stronger robustness for EEG-based motor imagery classification. The REEG-BTCNet consists of residual compact convolution (RCV) module and bi-directional temporal convolution (BTCN) module. Specifically, the RCV module consists of convolution with residual connection to learn high-level task specific EEG feature. The BTCN module consists of a window splitting module and various bi-directional temporal convolutions blocks to model the temporal information from the MI-EEG signals. The proposed model outperforms the current state-of-the-art techniques in the BCI Competition IV-2a dataset with an accuracy of 86.15% for the subject-dependent modes. For reproducibility, the code for this research and the trained models will be released on GitHub.

Electroencephalogram (EEG) emotion recognition has become a research focus in the field of human-computer interaction (HCI). However, the process of EEG signal collection requires lots of expertise, which makes the amount of labeled EEG data very limited. It constrains the performance of supervised methods which require large amounts of annotated data in some sense. Self-supervised learning paradigm, which aims to train models that do not require any labeled samples can make full use of a large amount of unlabeled EEG samples. But a drawback is that they fall short of learning class discriminative sample representations since no labeled information is utilized during training. To solve the above problem, we propose a semi-supervised model, named consistency regularization enhanced graph attention network (CR-GAT) for EEG emotion recognition. The CR-GAT mainly consists of three modules, namely the feature extraction and fusion (FEF) module, the feature graph building and augment (GBA) module as well as the consistency regularization (CR) module. Specifically, t he F EFm odule is to extract task-specific EEG features and highlight the most valuable features from the EEG signals. The GBA module is to build a sample-related graph representation of the EEG feature set. The CR module, which draws support samples from labeled samples and anchor samples from the entire sample set, intends to minimize the difference between the predicted class distributions from different graphs constructed by multi-views of the sample set to push samples that belong to the same class to be grouped together. We conduct our experiment on three real-world datasets, the experimental results show the method surpasses most of competitive models.

Major depressive disorder (MDD) has been widely studied because it is one of the most common and severe mental health issues. EEG-based depression identification is receiving a lot of attention in research on mental health care. This paper presents a new deep learning method for depression identification that employs deep learning algorithms with functional connectivity graphs in different states. The features commonly used to analyze the depression EEG, differential entropy (DE), and power spectral density (PSD) are extracted. Then, the adjacency matrix of the EEG signals is constructed using the Pearson correlation coefficient (PCC), phase locking value (PLV), and phase lag index (PLI) matrices between EEG signal pairs. For the classifier, an ovel multi-network architecture consisting of spatial-temporal graph convolution networks with brain graphs based on three functional connectivity measurement methods is employed to improve the learning ability of spatio-temporal features. Compared with other methods, our method has the best accuracy of 99.03%. The combination of the functional connectivity matrices of brain networks associated with multi-stream ST-GCN can predict the occurrence of depression and assist in diagnosing depression in an early stage.

Federated learning (FL) has been emerging as a new distributed machine learning paradigm recently. Although FL can protect the data privacy of participants by keeping their training data on local devices, there are recent works raising new privacy concerns especially when workers or the parameter server of FL are untrustworthy or malicious. One effective way to solve the problem is using hierarchical federated learning (HFL) where a few middle-layer aggregators (or called group leaders) are used to aggregate local model updates from workers and send group model updates to the parameter server. In this paper, we consider the participant selection problem of HFL in an edge cloud with multiple FL models, where each model needs to select one parameter server, a few group leaders and a certain amount of workers from edge servers to jointly perform HFL. We first formulate this problem as a non-linear integer programming, aiming to minimize the total learning cost of all models while satisfying the constrained edge resources. We then design a three-stage algorithm by decoupling the original problem into three sub-problems and solving them iteratively. Simulations with real-world datasets and FL models confirm that our proposed algorithm can efficiently reduce the average total learning cost in edge cloud compared with existing methods.