Yik-Chung Wu

Yingxian Chen, Zhengzhe Liu, Baoheng Zhang et al.

Weakly supervised detection of anomalies in surveillance videos is a challenging task. Going beyond existing works that have deficient capabilities to localize anomalies in long videos, we propose a novel glance and focus network to effectively integrate spatial-temporal information for accurate anomaly detection. In addition, we empirically found that existing approaches that use feature magnitudes to represent the degree of anomalies typically ignore the effects of scene variations, and hence result in sub-optimal performance due to the inconsistency of feature magnitudes across scenes. To address this issue, we propose the Feature Amplification Mechanism and a Magnitude Contrastive Loss to enhance the discriminativeness of feature magnitudes for detecting anomalies. Experimental results on two large-scale benchmarks UCF-Crime and XD-Violence manifest that our method outperforms state-of-the-art approaches.

Jian Du, Shaodan Ma, Yik-Chung Wu et al.

Phasor measurement units (PMUs) have the advantage of providing direct measurements of power states. However, as the number of PMUs in a power system is limited, the traditional supervisory control and data acquisition (SCADA) system cannot be replaced by the PMU-based system overnight. Therefore, hybrid power state estimation taking advantage of both systems is important. As experiments show that sampling phase errors among PMUs are inevitable in practical deployment, this paper proposes a distributed power state estimation algorithm under PMU phase errors. The proposed distributed algorithm only involves local computations and limited information exchange between neighboring areas, thus alleviating the heavy communication burden compared to the centralized approach. Simulation results show that the performance of the proposed algorithm is very close to that of centralized optimal hybrid state estimates without sampling phase error.

Siyuan Li, Shikai Fang, Lei Cheng et al.

Functional tensor decomposition can analyze multi-dimensional data with real-valued indices, paving the path for applications in machine learning and signal processing. A limitation of existing approaches is the assumption that the tensor rank-a critical parameter governing model complexity-is known. However, determining the optimal rank is a non-deterministic polynomial-time hard (NP-hard) task and there is a limited understanding regarding the expressive power of functional low-rank tensor models for continuous signals. We propose a rank-revealing functional Bayesian tensor completion (RR-FBTC) method. Modeling the latent functions through carefully designed multioutput Gaussian processes, RR-FBTC handles tensors with real-valued indices while enabling automatic tensor rank determination during the inference process. We establish the universal approximation property of the model for continuous multi-dimensional signals, demonstrating its expressive power in a concise format. To learn this model, we employ the variational inference framework and derive an efficient algorithm with closed-form updates. Experiments on both synthetic and real-world datasets demonstrate the effectiveness and superiority of the RR-FBTC over state-of-the-art approaches. The code is available at https://github.com/OceanSTARLab/RR-FBTC.

Wei-Bin Kou, Shuai Wang, Guangxu Zhu et al.

While federated learning (FL) improves the generalization of end-to-end autonomous driving by model aggregation, the conventional single-hop FL (SFL) suffers from slow convergence rate due to long-range communications among vehicles and cloud server. Hierarchical federated learning (HFL) overcomes such drawbacks via introduction of mid-point edge servers. However, the orchestration between constrained communication resources and HFL performance becomes an urgent problem. This paper proposes an optimization-based Communication Resource Constrained Hierarchical Federated Learning (CRCHFL) framework to minimize the generalization error of the autonomous driving model using hybrid data and model aggregation. The effectiveness of the proposed CRCHFL is evaluated in the Car Learning to Act (CARLA) simulation platform. Results show that the proposed CRCHFL both accelerates the convergence rate and enhances the generalization of federated learning autonomous driving model. Moreover, under the same communication resource budget, it outperforms the HFL by 10.33% and the SFL by 12.44%.

Zongze Li, Shuai Wang, Qingfeng Lin et al.

Reconfigurable intelligent surfaces (RISs) have a revolutionary capability to customize the radio propagation environment for wireless networks. To fully exploit the advantages of RISs in wireless systems, the phases of the reflecting elements must be jointly designed with conventional communication resources, such as beamformers, transmit power, and computation time. However, due to the unique constraints on the phase shift, and massive numbers of reflecting units and users in large-scale networks, the resulting optimization problems are challenging to solve. This paper provides a review of current optimization methods and artificial intelligence-based methods for handling the constraints imposed by RIS and compares them in terms of solution quality and computational complexity. Future challenges in phase shift optimization involving RISs are also described and potential solutions are discussed.

Yangge Chen, Lei Cheng, Yik-Chung Wu

Recently, there is a revival of interest in low-rank matrix completion-based unsupervised learning through the lens of dual-graph regularization, which has significantly improved the performance of multidisciplinary machine learning tasks such as recommendation systems, genotype imputation and image inpainting. While the dual-graph regularization contributes a major part of the success, computational costly hyper-parameter tunning is usually involved. To circumvent such a drawback and improve the completion performance, we propose a novel Bayesian learning algorithm that automatically learns the hyper-parameters associated with dual-graph regularization, and at the same time, guarantees the low-rankness of matrix completion. Notably, a novel prior is devised to promote the low-rankness of the matrix and encode the dual-graph information simultaneously, which is more challenging than the single-graph counterpart. A nontrivial conditional conjugacy between the proposed priors and likelihood function is then explored such that an efficient algorithm is derived under variational inference framework. Extensive experiments using synthetic and real-world datasets demonstrate the state-of-the-art performance of the proposed learning algorithm for various data analysis tasks.

Yunqi Wang, Yang Li, Qingjiang Shi et al.

In order to achieve high data rate and ubiquitous connectivity in future wireless networks, a key task is to efficiently manage the radio resource by judicious beamforming and power allocation. Unfortunately, the iterative nature of the commonly applied optimization-based algorithms cannot meet the low latency requirements due to the high computational complexity. For real-time implementations, deep learning-based approaches, especially the graph neural networks (GNNs), have been demonstrated with good scalability and generalization performance due to the permutation equivariance (PE) property. However, the current architectures are only equipped with the node-update mechanism, which prohibits the applications to a more general setup, where the unknown variables are also defined on the graph edges. To fill this gap, we propose an edge-update mechanism, which enables GNNs to handle both node and edge variables and prove its PE property with respect to both transmitters and receivers. Simulation results on typical radio resource management problems demonstrate that the proposed method achieves higher sum rate but with much shorter computation time than state-of-the-art methods and generalizes well on different numbers of base stations and users, different noise variances, interference levels, and transmit power budgets.

Zunhai Su, Rui Yang, Chao Zhang et al.

The rapid advancement toward long-context reasoning and multi-modal intelligence has made the memory footprint of the Key-Value (KV) cache a dominant memory bottleneck for efficient deployment. While the established per-channel quantization effectively accommodates intrinsic channel-wise outliers in Key tensors, its efficacy diminishes under extreme compression. In this work, we revisit the inherent limitations of the per-channel quantization paradigm from both empirical and theoretical perspectives. Our analysis identifies Token Norm Imbalance (TNI) as the primary bottleneck to quantization fidelity. We demonstrate that TNI systematically amplifies errors when shared quantization parameters are required to span token groups exhibiting substantial norm disparities. Instead of relying on intricate quantization pipelines (e.g., TurboQuant), we propose OScaR (Omni-Scaled Canalized Rotation), an accurate and lightweight KV cache compression framework for X-LLMs (i.e., text-only, multi-modal, and omni-modal LLMs). Advancing the per-channel paradigm, OScaR employs Canalized Rotation followed by Omni-Token Scaling to mitigate TNI-induced sequence-dimensional variance both effectively and efficiently, further supported by our optimized system design and CUDA kernels. Extensive evaluations across X-LLMs show that OScaR consistently outperforms existing methods and achieves near-lossless performance under INT2 quantization, establishing it as a robust, low-complexity, and universal framework that defines a new Pareto front. Compared with the BF16 FlashDecoding-v2 baseline, our OScaR implementation achieves a notable up to 3.0x speedup in decoding, reduces memory footprint by 5.3x, and increases throughput by 4.1x. The code for OScaR is publicly available at https://github.com/ZunhaiSu/OScaR-KV-Quant.

Le Xu, Lei Cheng, Ngai Wong et al.

Tensor train (TT) representation has achieved tremendous success in visual data completion tasks, especially when it is combined with tensor folding. However, folding an image or video tensor breaks the original data structure, leading to local information loss as nearby pixels may be assigned into different dimensions and become far away from each other. In this paper, to fully preserve the local information of the original visual data, we explore not folding the data tensor, and at the same time adopt graph information to regularize local similarity between nearby entries. To overcome the high computational complexity introduced by the graph-based regularization in the TT completion problem, we propose to break the original problem into multiple sub-problems with respect to each TT core fiber, instead of each TT core as in traditional methods. Furthermore, to avoid heavy parameter tuning, a sparsity promoting probabilistic model is built based on the generalized inverse Gaussian (GIG) prior, and an inference algorithm is derived under the mean-field approximation. Experiments on both synthetic data and real-world visual data show the superiority of the proposed methods.

Yunqi Wang, Yang Li, Qingjiang Shi et al.

Cooperative beamforming design has been recognized as an effective approach in modern wireless networks to meet the dramatically increasing demand of various wireless data traffics. It is formulated as an optimization problem in conventional approaches and solved iteratively in an instance-by-instance manner. Recently, learning-based methods have emerged with real-time implementation by approximating the mapping function from the problem instances to the corresponding solutions. Among various neural network architectures, graph neural networks (GNNs) can effectively utilize the graph topology in wireless networks to achieve better generalization ability on unseen problem sizes. However, the current GNNs are only equipped with the node-update mechanism, which restricts it from modeling more complicated problems such as the cooperative beamforming design, where the beamformers are on the graph edges of wireless networks. To fill this gap, we propose an edge-graph-neural-network (Edge-GNN) by incorporating an edge-update mechanism into the GNN, which learns the cooperative beamforming on the graph edges. Simulation results show that the proposed Edge-GNN achieves higher sum rate with much shorter computation time than state-of-the-art approaches, and generalizes well to different numbers of base stations and user equipments.

Wei-Bin Kou, Qingfeng Lin, Ming Tang et al.

Street Scene Semantic Understanding (denoted as TriSU) is a complex task for autonomous driving (AD). However, inference model trained from data in a particular geographical region faces poor generalization when applied in other regions due to inter-city data domain-shift. Hierarchical Federated Learning (HFL) offers a potential solution for improving TriSU model generalization by collaborative privacy-preserving training over distributed datasets from different cities. Unfortunately, it suffers from slow convergence because data from different cities are with disparate statistical properties. Going beyond existing HFL methods, we propose a Gaussian heterogeneous HFL algorithm (FedGau) to address inter-city data heterogeneity so that convergence can be accelerated. In the proposed FedGau algorithm, both single RGB image and RGB dataset are modelled as Gaussian distributions for aggregation weight design. This approach not only differentiates each RGB image by respective statistical distribution, but also exploits the statistics of dataset from each city in addition to the conventionally considered data volume. With the proposed approach, the convergence is accelerated by 35.5\%-40.6\% compared to existing state-of-the-art (SOTA) HFL methods. On the other hand, to reduce the involved communication resource, we further introduce a novel performance-aware adaptive resource scheduling (AdapRS) policy. Unlike the traditional static resource scheduling policy that exchanges a fixed number of models between two adjacent aggregations, AdapRS adjusts the number of model aggregation at different levels of HFL so that unnecessary communications are minimized. Extensive experiments demonstrate that AdapRS saves 29.65\% communication overhead compared to conventional static resource scheduling policy while maintaining almost the same performance.

Xueke Tong, Hancheng Zhu, Lei Cheng et al.

To efficiently express tensor data using the Tucker format, a critical task is to minimize the multilinear rank such that the model would not be over-flexible and lead to overfitting. Due to the lack of rank minimization tools in tensor, existing works connect Tucker multilinear rank minimization to trace norm minimization of matrices unfolded from the tensor data. While these formulations try to exploit the common aim of identifying the low-dimensional structure of the tensor and matrix, this paper reveals that existing trace norm-based formulations in Tucker completion are inefficient in multilinear rank minimization. We further propose a new interpretation of Tucker format such that trace norm minimization is applied to the factor matrices of the equivalent representation, rather than some matrices unfolded from tensor data. Based on the newly established problem formulation, a fixed point iteration algorithm is proposed, and its convergence is proved. Numerical results are presented to show that the proposed algorithm exhibits significant improved performance in terms of multilinear rank learning and consequently tensor signal recovery accuracy, compared to existing trace norm based Tucker completion methods.

Chengyang Li, Shuai Wang, Kejiang Ye et al.

This paper considers multi-agent embodied question answering (MA-EQA), which aims to query robot teams on what they have seen over a long horizon. In contrast to existing edge resource management methods that emphasize sensing, communication, or computation performance metrics, MA-EQA emphasizes the memory qualities. To cope with this paradigm shift, we propose a quality of memory (QoM) model based on generative adversarial exam (GAE), which leverages forward simulation to assess memory retrieval and uses the resulting exam scores to compute QoM values. Then we propose memory centric power allocation (MCPA), which maximizes the QoM function under communication resource constraints. Through asymptotic analysis, it is found that the transmit powers are proportional to the GAE error probability, thus prioritizing towards high-QoM robots. Extensive experiments demonstrate that MCPA achieves significant improvements over extensive benchmarks in terms of diverse metrics in various scenarios.

Wei Zuo, Zeyi Ren, Chengyang Li et al.

Existing motion planning methods often struggle with rapid-motion obstacles due to an insufficient understanding of environmental changes. To address this, we propose integrating motion planners with Doppler LiDARs, which provide not only ranging measurements but also instantaneous point velocities. However, this integration is nontrivial due to the requirements of high accuracy and high frequency. To this end, we introduce Doppler Planning Network (DPNet), which tracks and reacts to rapid obstacles via Doppler model-based learning. We first propose a Doppler Kalman neural network (D-KalmanNet) to track obstacle states under a partially observable Gaussian state space model. We then leverage the predicted motions of obstacles to construct a Doppler-tuned model predictive control (DT-MPC) framework for ego-motion planning, enabling runtime auto-tuning of controller parameters. These two modules allow DPNet to learn fast environmental changes from minimal data while remaining lightweight, achieving high frequency and high accuracy in both tracking and planning. Experiments on high-fidelity simulator and real-world datasets demonstrate the superiority of DPNet over extensive benchmark schemes.

Yikun Wang, Yang Li, Yik-Chung Wu et al.

While deep learning (DL)-based methods have achieved remarkable success in continuous wireless resource allocation, efficient solutions for problems involving discrete variables remain challenging. This is primarily due to the zero-gradient issue in backpropagation, the difficulty of enforcing intricate constraints with discrete variables, and the inability in generating solutions with non-same-parameter-same-decision (non-SPSD) property. To address these challenges, this paper proposes a general DL framework by introducing the support set to represent the discrete variables. We model the elements of the support set as random variables and learn their joint probability distribution. By factorizing the joint probability as the product of conditional probabilities, each conditional probability is sequentially learned. This probabilistic modeling directly tackles all the aforementioned challenges of DL for handling discrete variables. By operating on probability distributions instead of hard binary decisions, the framework naturally avoids the zero-gradient issue. During the learning of the conditional probabilities, discrete constraints can be seamlessly enforced by masking out infeasible solutions. Moreover, with a dynamic context embedding that captures the evolving discrete solutions, the non-SPSD property is inherently provided by the proposed framework. We apply the proposed framework to two representative mixed-discrete wireless resource allocation problems: (a) joint user association and beamforming in cell-free systems, and (b) joint antenna positioning and beamforming in movable antenna-aided systems. Simulation results demonstrate that the proposed DL framework consistently outperforms existing baselines in terms of both system performance and computational efficiency.

Runming Yang, Taiqiang Wu, Jiahao Wang et al.

Knowledge distillation (KD) has been a predominant method for compressing Large Language Models (LLMs). In this paper, we first revisit KD and Low-Rank Adaption (LoRA) and demonstrate that they follow the same paradigm. Inspired by this observation, we propose a parameter-efficient knowledge distillation method, LLM-NEO, which integrates LoRA into KD to improve the efficiency of knowledge transfer. After that, we summarize some valuable guidelines for the hyperparameters in LLM-NEO. Experimental results on compressing Llama 2 and Llama 3.2 show that LLM-NEO outperforms various baselines. Further analysis demonstrates the robustness of the proposed LLM-NEO on variants of LoRA. The code and trained models are available at [Github](https://github.com/yang3121099/LLM-Neo).

Yikun Wang, Yujie Wan, Wei Zuo et al.

Low-altitude Gaussian splatting (LAGS) facilitates 3D scene reconstruction by aggregating aerial images from distributed drones. However, as LAGS prioritizes maximizing reconstruction quality over communication throughput, existing low-altitude resource allocation schemes become inefficient. This inefficiency stems from their failure to account for image diversity introduced by varying viewpoints. To fill this gap, we propose a groupwise heterogeneous graph neural network (GW-HGNN) for LAGS resource allocation. GW-HGNN explicitly models the non-uniform contribution of different image groups to the reconstruction process, thus automatically balancing data fidelity and transmission cost. The key insight of GW-HGNN is to transform LAGS losses and communication constraints into graph learning costs for dual-level message passing. Experiments on real-world LAGS datasets demonstrate that GW-HGNN significantly outperforms state-of-the-art benchmarks across key rendering metrics, including PSNR, SSIM, and LPIPS. Furthermore, GW-HGNN reduces computational latency by approximately 100x compared to the widely-used MOSEK solver, achieving millisecond-level inference suitable for real-time deployment.

Wei-Bin Kou, Guangxu Zhu, Bingyang Cheng et al.

Federated Learning (FL) enables collaborative training of autonomous driving (AD) models across distributed vehicles while preserving data privacy. However, FL encounters critical challenges such as poor generalization and slow convergence due to non-independent and identically distributed (non-IID) data from diverse driving environments. To overcome these obstacles, we introduce Federated Deep Supervision and Regularization (FedDSR), a paradigm that incorporates multi-access intermediate layer supervision and regularization within federated AD system. Specifically, FedDSR comprises following integral strategies: (I) to select multiple intermediate layers based on predefined architecture-agnostic standards. (II) to compute mutual information (MI) and negative entropy (NE) on those selected layers to serve as intermediate loss and regularizer. These terms are integrated into the output-layer loss to form a unified optimization objective, enabling comprehensive optimization across the network hierarchy. (III) to aggregate models from vehicles trained based on aforementioned rules of (I) and (II) to generate the global model on central server. By guiding and penalizing the learning of feature representations at intermediate stages, FedDSR enhances the model generalization and accelerates model convergence for federated AD. We then take the semantic segmentation task as an example to assess FedDSR and apply FedDSR to multiple model architectures and FL algorithms. Extensive experiments demonstrate that FedDSR achieves up to 8.93% improvement in mIoU and 28.57% reduction in training rounds, compared to other FL baselines, making it highly suitable for practical deployment in federated AD ecosystems.

Taiqiang Wu, Runming Yang, Jiayi Li et al.

Large language models (LLMs) consistently benefit from further fine-tuning on various tasks. However, we observe that directly tuning the Instruct (i.e., instruction-tuned) models often leads to marginal improvements and even performance degeneration. Notably, paired Base models, the foundation for these Instruct variants, contain highly similar weight values (i.e., less than 2% on average for Llama 3.1 8B). The Base model tends to be a good learner yet a weak backbone without post-training. Therefore, we propose a novel Shadow-FT framework to tune the Instruct models by leveraging the corresponding Base models. The key insight is to fine-tune the Base model, and then \textit{directly} graft the learned weight updates to the Instruct model. Our proposed Shadow-FT introduces no additional parameters, is easy to implement, and significantly improves performance. We conduct extensive experiments on tuning mainstream LLMs, such as Qwen 3 and Llama 3 series, and evaluate them across 19 benchmarks covering coding, reasoning, and mathematical tasks. Experimental results demonstrate that Shadow-FT consistently outperforms conventional full-parameter and parameter-efficient tuning approaches. Further analyses indicate that Shadow-FT can be applied to multimodal large language models (MLLMs) and combined with direct preference optimization~(DPO). Codes and weights are available at \href{https://github.com/wutaiqiang/Shadow-FT}{Github}.

Chen Zhang, Yuxin Cheng, Chenchen Ding et al.

Zeroth-order (ZO) optimization enables memory-efficient training of neural networks by estimating gradients via forward passes only, eliminating the need for backpropagation. However, the stochastic nature of gradient estimation significantly obscures the training dynamics, in contrast to the well-characterized behavior of first-order methods under Neural Tangent Kernel (NTK) theory. To address this, we introduce the Neural Zeroth-order Kernel (NZK) to describe model evolution in function space under ZO updates. For linear models, we prove that the expected NZK remains constant throughout training and depends explicitly on the first and second moments of the random perturbation directions. This invariance yields a closed-form expression for model evolution under squared loss. We further extend the analysis to linearized neural networks. Interpreting ZO updates as kernel gradient descent via NZK provides a novel perspective for potentially accelerating convergence. Extensive experiments across synthetic and real-world datasets (including MNIST, CIFAR-10, and Tiny ImageNet) validate our theoretical results and demonstrate acceleration when using a single shared random vector.

Hao Zhang, Qingfeng Lin, Yang Li et al.

Activity detection is an important task in the next generation grant-free multiple access. While there are a number of existing algorithms designed for this purpose, they mostly require precise information about the network, such as large-scale fading coefficients, small-scale fading channel statistics, noise variance at the access points, and user activity probability. Acquiring these information would take a significant overhead and their estimated values might not be accurate. This problem is even more severe in cell-free networks as there are many of these parameters to be acquired. Therefore, this paper sets out to investigate the activity detection problem without the above-mentioned information. In order to handle so many unknown parameters, this paper employs the Bayesian approach, where the unknown variables are endowed with prior distributions which effectively act as regularizations. Together with the likelihood function, a maximum a posteriori (MAP) estimator and a variational inference algorithm are derived. Extensive simulations demonstrate that the proposed methods, even without the knowledge of these system parameters, perform better than existing state-of-the-art methods, such as covariance-based and approximate message passing methods.

Chen Zhang, Steven Tin Sui Luo, Jason Chun Lok Li et al.

We investigate the learning of implicit neural representation (INR) using an overparameterized multilayer perceptron (MLP) via a novel nonparametric teaching perspective. The latter offers an efficient example selection framework for teaching nonparametrically defined (viz. non-closed-form) target functions, such as image functions defined by 2D grids of pixels. To address the costly training of INRs, we propose a paradigm called Implicit Neural Teaching (INT) that treats INR learning as a nonparametric teaching problem, where the given signal being fitted serves as the target function. The teacher then selects signal fragments for iterative training of the MLP to achieve fast convergence. By establishing a connection between MLP evolution through parameter-based gradient descent and that of function evolution through functional gradient descent in nonparametric teaching, we show for the first time that teaching an overparameterized MLP is consistent with teaching a nonparametric learner. This new discovery readily permits a convenient drop-in of nonparametric teaching algorithms to broadly enhance INR training efficiency, demonstrating 30%+ training time savings across various input modalities.

Wei-Bin Kou, Qingfeng Lin, Ming Tang et al.

To improve the generalization of the autonomous driving (AD) perception model, vehicles need to update the model over time based on the continuously collected data. As time progresses, the amount of data fitted by the AD model expands, which helps to improve the AD model generalization substantially. However, such ever-expanding data is a double-edged sword for the AD model. Specifically, as the fitted data volume grows to exceed the the AD model's fitting capacities, the AD model is prone to under-fitting. To address this issue, we propose to use a pretrained Large Vision Models (LVMs) as backbone coupled with downstream perception head to understand AD semantic information. This design can not only surmount the aforementioned under-fitting problem due to LVMs' powerful fitting capabilities, but also enhance the perception generalization thanks to LVMs' vast and diverse training data. On the other hand, to mitigate vehicles' computational burden of training the perception head while running LVM backbone, we introduce a Posterior Optimization Trajectory (POT)-Guided optimization scheme (POTGui) to accelerate the convergence. Concretely, we propose a POT Generator (POTGen) to generate posterior (future) optimization direction in advance to guide the current optimization iteration, through which the model can generally converge within 10 epochs. Extensive experiments demonstrate that the proposed method improves the performance by over 66.48\% and converges faster over 6 times, compared to the existing state-of-the-art approach.

Wei-Bin Kou, Guangxu Zhu, Rongguang Ye et al.

Learning-based street scene semantic understanding in autonomous driving (AD) has advanced significantly recently, but the performance of the AD model is heavily dependent on the quantity and quality of the annotated training data. However, traditional manual labeling involves high cost to annotate the vast amount of required data for training robust model. To mitigate this cost of manual labeling, we propose a Label Anything Model (denoted as LAM), serving as an interpretable, high-fidelity, and prompt-free data annotator. Specifically, we firstly incorporate a pretrained Vision Transformer (ViT) to extract the latent features. On top of ViT, we propose a semantic class adapter (SCA) and an optimization-oriented unrolling algorithm (OptOU), both with a quite small number of trainable parameters. SCA is proposed to fuse ViT-extracted features to consolidate the basis of the subsequent automatic annotation. OptOU consists of multiple cascading layers and each layer contains an optimization formulation to align its output with the ground truth as closely as possible, though which OptOU acts as being interpretable rather than learning-based blackbox nature. In addition, training SCA and OptOU requires only a single pre-annotated RGB seed image, owing to their small volume of learnable parameters. Extensive experiments clearly demonstrate that the proposed LAM can generate high-fidelity annotations (almost 100% in mIoU) for multiple real-world datasets (i.e., Camvid, Cityscapes, and Apolloscapes) and CARLA simulation dataset.

Chen Zhang, Weixin Bu, Zeyi Ren et al.

Inferring properties of graph-structured data, e.g., the solubility of molecules, essentially involves learning the implicit mapping from graphs to their properties. This learning process is often costly for graph property learners like Graph Convolutional Networks (GCNs). To address this, we propose a paradigm called Graph Neural Teaching (GraNT) that reinterprets the learning process through a novel nonparametric teaching perspective. Specifically, the latter offers a theoretical framework for teaching implicitly defined (i.e., nonparametric) mappings via example selection. Such an implicit mapping is realized by a dense set of graph-property pairs, with the GraNT teacher selecting a subset of them to promote faster convergence in GCN training. By analytically examining the impact of graph structure on parameter-based gradient descent during training, and recasting the evolution of GCNs--shaped by parameter updates--through functional gradient descent in nonparametric teaching, we show for the first time that teaching graph property learners (i.e., GCNs) is consistent with teaching structure-aware nonparametric learners. These new findings readily commit GraNT to enhancing learning efficiency of the graph property learner, showing significant reductions in training time for graph-level regression (-36.62%), graph-level classification (-38.19%), node-level regression (-30.97%) and node-level classification (-47.30%), all while maintaining its generalization performance.

Zeyi Ren, Qingfeng Lin, Jingreng Lei et al.

In the realm of activity detection for massive machine-type communications, intelligent reflecting surfaces (IRS) have shown significant potential in enhancing coverage for devices lacking direct connections to the base station (BS). However, traditional activity detection methods are typically designed for a single type of channel model, which does not reflect the complexities of real-world scenarios, particularly in systems incorporating IRS. To address this challenge, this paper introduces a novel approach that combines model-driven deep unfolding with a mixture of experts (MoE) framework. By automatically selecting one of three expert designs and applying it to the unfolded projected gradient method, our approach eliminates the need for prior knowledge of channel types between devices and the BS. Simulation results demonstrate that the proposed MoE-augmented deep unfolding method surpasses the traditional covariance-based method and black-box neural network design, delivering superior detection performance under mixed channel fading conditions.

Zhenrong Liu, Yang Li, Yi Gong et al.

In continual learning, networks confront a trade-off between stability and plasticity when trained on a sequence of tasks. To bolster plasticity without sacrificing stability, we propose a novel training algorithm called LRFR. This approach optimizes network parameters in the null space of the past tasks' feature representation matrix to guarantee the stability. Concurrently, we judiciously select only a subset of neurons in each layer of the network while training individual tasks to learn the past tasks' feature representation matrix in low-rank. This increases the null space dimension when designing network parameters for subsequent tasks, thereby enhancing the plasticity. Using CIFAR-100 and TinyImageNet as benchmark datasets for continual learning, the proposed approach consistently outperforms state-of-the-art methods.

Chen Zhang, Weixin Bu, Wendong Xu et al.

Transformers have achieved remarkable success in time series modeling, yet their internal mechanisms remain opaque. This work demystifies the Transformer encoder by establishing its fundamental equivalence to a Graph Convolutional Network (GCN). We show that in the forward pass, the attention distribution matrix serves as a dynamic adjacency matrix, and its composition with subsequent transformations performs computations analogous to graph convolution. Moreover, we demonstrate that in the backward pass, the update dynamics of value and feed-forward projections mirror those of GCN parameters. Building on this unified theoretical reinterpretation, we propose \textbf{Fighter} (Flexible Graph Convolutional Transformer), a streamlined architecture that removes redundant linear projections and incorporates multi-hop graph aggregation. This perspective yields an explicit and interpretable representation of temporal dependencies across different scales, naturally expressed as graph edges. Experiments on standard forecasting benchmarks confirm that Fighter achieves competitive performance while providing clearer mechanistic interpretability of its predictions.

Yingxian Chen, Jiahui Liu, Ruidi Fan et al.

Understanding abnormal events in videos is a vital and challenging task that has garnered significant attention in a wide range of applications. Although current video understanding Multi-modal Large Language Models (MLLMs) are capable of analyzing general videos, they often struggle to handle anomalies due to the spatial and temporal sparsity of abnormal events, where the redundant information always leads to suboptimal outcomes. To address these challenges, exploiting the representation and generalization capabilities of Vison Language Models (VLMs) and Large Language Models (LLMs), we propose VA-GPT, a novel MLLM designed for summarizing and localizing abnormal events in various videos. Our approach efficiently aligns effective tokens between visual encoders and LLMs through two key proposed modules: Spatial Effective Token Selection (SETS) and Temporal Effective Token Generation (TETG). These modules enable our model to effectively capture and analyze both spatial and temporal information associated with abnormal events, resulting in more accurate responses and interactions. Furthermore, we construct an instruction-following dataset specifically for fine-tuning video-anomaly-aware MLLMs, and introduce a cross-domain evaluation benchmark based on XD-Violence dataset. Our proposed method outperforms existing state-of-the-art methods on various benchmarks.

Bingyang Cheng, Zhongtao Chen, Yichen Jin et al.

Tensor CANDECOMP/PARAFAC decomposition (CPD) is a fundamental model for tensor reconstruction. Although the Bayesian framework allows for principled uncertainty quantification and automatic hyperparameter learning, existing methods do not scale well for large tensors because of high-dimensional matrix inversions. To this end, we introduce CP-GAMP, a scalable Bayesian CPD algorithm. This algorithm leverages generalized approximate message passing (GAMP) to avoid matrix inversions and incorporates an expectation-maximization routine to jointly infer the tensor rank and noise power. Through multiple experiments, for synthetic 100x100x100 rank 20 tensors with only 20% elements observed, the proposed algorithm reduces runtime by 82.7% compared to the state-of-the-art variational Bayesian CPD method, while maintaining comparable reconstruction accuracy.

Zeyi Ren, Jingreng Lei, Yichen Jin et al.

The development of edge computing places critical demands on energy-efficient model deployment for multiple-input multiple-output (MIMO) detection tasks. Deploying deep unfolding models such as PGD-Nets and ADMM-Nets into resource-constrained edge devices using quantization methods is challenging. Existing quantization methods based on quantization aware training (QAT) suffer from performance degradation due to their reliance on parametric distribution assumption of activations and static quantization step sizes. To address these challenges, this paper proposes a novel kernel-based adaptive quantization (KAQ) framework for deep unfolding networks. By utilizing a joint kernel density estimation (KDE) and maximum mean discrepancy (MMD) approach to align activation distributions between full-precision and quantized models, the need for prior distribution assumptions is eliminated. Additionally, a dynamic step size updating method is introduced to adjust the quantization step size based on the channel conditions of wireless networks. Extensive simulations demonstrate that the accuracy of proposed KAQ framework outperforms traditional methods and successfully reduces the model's inference latency.

Wei-Bin Kou, Guangxu Zhu, Bingyang Cheng et al.

Street Scene Semantic Understanding (denoted as S3U) is a crucial but complex task for autonomous driving (AD) vehicles. Their inference models typically face poor generalization due to domain-shift. Federated Learning (FL) has emerged as a promising paradigm for enhancing the generalization of AD models through privacy-preserving distributed learning. However, these FL AD models face significant temporal catastrophic forgetting when deployed in dynamically evolving environments, where continuous adaptation causes abrupt erosion of historical knowledge. This paper proposes Federated Exponential Moving Average (FedEMA), a novel framework that addresses this challenge through two integral innovations: (I) Server-side model's historical fitting capability preservation via fusing current FL round's aggregation model and a proposed previous FL round's exponential moving average (EMA) model; (II) Vehicle-side negative entropy regularization to prevent FL models' possible overfitting to EMA-introduced temporal patterns. Above two strategies empower FedEMA a dual-objective optimization that balances model generalization and adaptability. In addition, we conduct theoretical convergence analysis for the proposed FedEMA. Extensive experiments both on Cityscapes dataset and Camvid dataset demonstrate FedEMA's superiority over existing approaches, showing 7.12% higher mean Intersection-over-Union (mIoU).

Shuai Wang, Dachuan Li, Rui Wang et al.

Wireless federated learning (FL) is an emerging machine learning paradigm that trains a global parametric model from distributed datasets via wireless communications. This paper proposes a unit-modulus wireless FL (UMWFL) framework, which simultaneously uploads local model parameters and computes global model parameters via optimized phase shifting. The proposed framework avoids sophisticated baseband signal processing, leading to both low communication delays and implementation costs. A training loss bound is derived and a penalty alternating minimization (PAM) algorithm is proposed to minimize the nonconvex nonsmooth loss bound. Experimental results in the Car Learning to Act (CARLA) platform show that the proposed UMWFL framework with PAM algorithm achieves smaller training losses and testing errors than those of the benchmark scheme.

Yunqi Wang, Furui Liu, Zhitang Chen et al.

Domain generalization aims to learn knowledge invariant across different distributions while semantically meaningful for downstream tasks from multiple source domains, to improve the model's generalization ability on unseen target domains. The fundamental objective is to understand the underlying "invariance" behind these observational distributions and such invariance has been shown to have a close connection to causality. While many existing approaches make use of the property that causal features are invariant across domains, we consider the causal invariance of the average causal effect of the features to the labels. This invariance regularizes our training approach in which interventions are performed on features to enforce stability of the causal prediction by the classifier across domains. Our work thus sheds some light on the domain generalization problem by introducing invariance of the mechanisms into the learning process. Experiments on several benchmark datasets demonstrate the performance of the proposed method against SOTAs.

Shuai Wang, Yuncong Hong, Rui Wang et al.

Edge federated learning (FL) is an emerging paradigm that trains a global parametric model from distributed datasets based on wireless communications. This paper proposes a unit-modulus over-the-air computation (UMAirComp) framework to facilitate efficient edge federated learning, which simultaneously uploads local model parameters and updates global model parameters via analog beamforming. The proposed framework avoids sophisticated baseband signal processing, leading to low communication delays and implementation costs. Training loss bounds of UMAirComp FL systems are derived and two low-complexity large-scale optimization algorithms, termed penalty alternating minimization (PAM) and accelerated gradient projection (AGP), are proposed to minimize the nonconvex nonsmooth loss bound. Simulation results show that the proposed UMAirComp framework with PAM algorithm achieves a smaller mean square error of model parameters' estimation, training loss, and test error compared with other benchmark schemes. Moreover, the proposed UMAirComp framework with AGP algorithm achieves satisfactory performance while reduces the computational complexity by orders of magnitude compared with existing optimization algorithms. Finally, we demonstrate the implementation of UMAirComp in a vehicle-to-everything autonomous driving simulation platform. It is found that autonomous driving tasks are more sensitive to model parameter errors than other tasks since the neural networks for autonomous driving contain sparser model parameters.

Dan Liu, Shuai Wang, Zhigang Wen et al.

Edge learning (EL), which uses edge computing as a platform to execute machine learning algorithms, is able to fully exploit the massive sensing data generated by Internet of Things (IoT). However, due to the limited transmit power at IoT devices, collecting the sensing data in EL systems is a challenging task. To address this challenge, this paper proposes to integrate unmanned ground vehicle (UGV) with EL. With such a scheme, the UGV could improve the communication quality by approaching various IoT devices. However, different devices may transmit different data for different machine learning jobs and a fundamental question is how to jointly plan the UGV path, the devices' energy consumption, and the number of samples for different jobs? This paper further proposes a graph-based path planning model, a network energy consumption model and a sample size planning model that characterizes F-measure as a function of the minority class sample size. With these models, the joint path, energy and sample size planning (JPESP) problem is formulated as a large-scale mixed integer nonlinear programming (MINLP) problem, which is nontrivial to solve due to the high-dimensional discontinuous variables related to UGV movement. To this end, it is proved that each IoT device should be served only once along the path, thus the problem dimension is significantly reduced. Furthermore, to handle the discontinuous variables, a tabu search (TS) based algorithm is derived, which converges in expectation to the optimal solution to the JPESP problem. Simulation results under different task scenarios show that our optimization schemes outperform the fixed EL and the full path EL schemes.

Lei Cheng, Zhongtao Chen, Qingjiang Shi et al.

Tensor rank learning for canonical polyadic decomposition (CPD) has long been deemed as an essential yet challenging problem. In particular, since the tensor rank controls the complexity of the CPD model, its inaccurate learning would cause overfitting to noise or underfitting to the signal sources, and even destroy the interpretability of model parameters. However, the optimal determination of a tensor rank is known to be a non-deterministic polynomial-time hard (NP-hard) task. Rather than exhaustively searching for the best tensor rank via trial-and-error experiments, Bayesian inference under the Gaussian-gamma prior was introduced in the context of probabilistic CPD modeling, and it was shown to be an effective strategy for automatic tensor rank determination. This triggered flourishing research on other structured tensor CPDs with automatic tensor rank learning. On the other side of the coin, these research works also reveal that the Gaussian-gamma model does not perform well for high-rank tensors and/or low signal-to-noise ratios (SNRs). To overcome these drawbacks, in this paper, we introduce a more advanced generalized hyperbolic (GH) prior to the probabilistic CPD model, which not only includes the Gaussian-gamma model as a special case, but also is more flexible to adapt to different levels of sparsity. Based on this novel probabilistic model, an algorithm is developed under the framework of variational inference, where each update is obtained in a closed-form. Extensive numerical results, using synthetic data and real-world datasets, demonstrate the significantly improved performance of the proposed method in learning both low as well as high tensor ranks even for low SNR cases.

Shuai Wang, Rui Wang, Qi Hao et al.

While machine-type communication (MTC) devices generate massive data, they often cannot process this data due to limited energy and computation power. To this end, edge intelligence has been proposed, which collects distributed data and performs machine learning at the edge. However, this paradigm needs to maximize the learning performance instead of the communication throughput, for which the celebrated water-filling and max-min fairness algorithms become inefficient since they allocate resources merely according to the quality of wireless channels. This paper proposes a learning centric power allocation (LCPA) method, which allocates radio resources based on an empirical classification error model. To get insights into LCPA, an asymptotic optimal solution is derived. The solution shows that the transmit powers are inversely proportional to the channel gain, and scale exponentially with the learning parameters. Experimental results show that the proposed LCPA algorithm significantly outperforms other power allocation algorithms.

Shuai Wang, Yik-Chung Wu, Minghua Xia et al.

While machine-type communication (MTC) devices generate considerable amounts of data, they often cannot process the data due to limited energy and computational power. To empower MTC with intelligence, edge machine learning has been proposed. However, power allocation in this paradigm requires maximizing the learning performance instead of the communication throughput, for which the celebrated water-filling and max-min fairness algorithms become inefficient. To this end, this paper proposes learning centric power allocation (LCPA), which provides a new perspective on radio resource allocation in learning driven scenarios. By employing 1) an empirical classification error model that is supported by learning theory and 2) an uncertainty sampling method that accounts for different distributions at users, LCPA is formulated as a nonconvex nonsmooth optimization problem, and is solved using a majorization minimization (MM) framework. To get deeper insights into LCPA, asymptotic analysis shows that the transmit powers are inversely proportional to the channel gains, and scale exponentially with the learning parameters. This is in contrast to traditional power allocations where quality of wireless channels is the only consideration. Last but not least, a large-scale optimization algorithm termed mirror-prox LCPA is further proposed to enable LCPA in large-scale settings. Extensive numerical results demonstrate that the proposed LCPA algorithms outperform traditional power allocation algorithms, and the large-scale optimization algorithm reduces the computation time by orders of magnitude compared with MM-based LCPA but still achieves competing learning performance.

Jian Du, Shaodan Ma, Yik-Chung Wu et al.

Gaussian belief propagation (BP) has been widely used for distributed inference in large-scale networks such as the smart grid, sensor networks, and social networks, where local measurements/observations are scattered over a wide geographical area. One particular case is when two neighboring agents share a common observation. For example, to estimate voltage in the direct current (DC) power flow model, the current measurement over a power line is proportional to the voltage difference between two neighboring buses. When applying the Gaussian BP algorithm to this type of problem, the convergence condition remains an open issue. In this paper, we analyze the convergence properties of Gaussian BP for this pairwise linear Gaussian model. We show analytically that the updating information matrix converges at a geometric rate to a unique positive definite matrix with arbitrary positive semidefinite initial value and further provide the necessary and sufficient convergence condition for the belief mean vector to the optimal estimate.

Jian Du, Shaodan Ma, Yik-Chung Wu et al.

Gaussian belief propagation (BP) has been widely used for distributed estimation in large-scale networks such as the smart grid, communication networks, and social networks, where local measurements/observations are scattered over a wide geographical area. However, the convergence of Gaus- sian BP is still an open issue. In this paper, we consider the convergence of Gaussian BP, focusing in particular on the convergence of the information matrix. We show analytically that the exchanged message information matrix converges for arbitrary positive semidefinite initial value, and its dis- tance to the unique positive definite limit matrix decreases exponentially fast.

Jian Du, Shaodan Ma, Yik-Chung Wu et al.

This paper considers inference over distributed linear Gaussian models using factor graphs and Gaussian belief propagation (BP). The distributed inference algorithm involves only local computation of the information matrix and of the mean vector, and message passing between neighbors. Under broad conditions, it is shown that the message information matrix converges to a unique positive definite limit matrix for arbitrary positive semidefinite initialization, and it approaches an arbitrarily small neighborhood of this limit matrix at a doubly exponential rate. A necessary and sufficient convergence condition for the belief mean vector to converge to the optimal centralized estimator is provided under the assumption that the message information matrix is initialized as a positive semidefinite matrix. Further, it is shown that Gaussian BP always converges when the underlying factor graph is given by the union of a forest and a single loop. The proposed convergence condition in the setup of distributed linear Gaussian models is shown to be strictly weaker than other existing convergence conditions and requirements, including the Gaussian Markov random field based walk-summability condition, and applicable to a large class of scenarios.