Search

Large-scale pre-trained Vision-Language Models (VLMs) have gained prominence in various visual and multimodal tasks, yet the deployment of VLMs on downstream application platforms remains challenging due to their prohibitive requirements of training samples and computing resources. Fine-tuning and quantization of VLMs can substantially reduce the sample and computation costs, which are in urgent need. There are two prevailing paradigms in quantization, Quantization-Aware Training (QAT) can effectively quantize large-scale VLMs but incur a huge training cost, while low-bit Post-Training Quantization (PTQ) suffers from a notable performance drop. We propose a method that balances fine-tuning and quantization named ``Prompt for Quantization'' (P4Q), in which we design a lightweight architecture to leverage contrastive loss supervision to enhance the recognition performance of a PTQ model. Our method can effectively reduce the gap between image features and text features caused by low-bit quantization, based on learnable prompts to reorganize textual representations and a low-bit adapter to realign the distributions of image and text features. We also introduce a distillation loss based on cosine similarity predictions to distill the quantized model using a full-precision teacher. Extensive experimental results demonstrate that our P4Q method outperforms prior arts, even achieving comparable results to its full-precision counterparts. For instance, our 8-bit P4Q can theoretically compress the CLIP-ViT/B-32 by 4 $\times$ while achieving 66.94\% Top-1 accuracy, outperforming the learnable prompt fine-tuned full-precision model by 2.24\% with negligible additional parameters on the ImageNet dataset.

This paper is concerned with unmanned aerial vehicle (UAV) video coding and transmission in scenarios such as emergency rescue and environmental monitoring. Unlike existing methods of modeling UAV video source coding and channel transmission separately, we investigate the joint source-channel optimization issue for video coding and transmission. Particularly, we design eight-dimensional delay-power-rate-distortion models in terms of source coding and channel transmission and characterize the correlation between video coding and transmission, with which a joint source-channel optimization problem is formulated. Its objective is to minimize end-to-end distortion and UAV power consumption by optimizing fine-grained parameters related to UAV video coding and transmission. This problem is confirmed to be a challenging sequential-decision and non-convex optimization problem. We therefore decompose it into a family of repeated optimization problems by Lyapunov optimization and design an approximate convex optimization scheme with provable performance guarantees to tackle these problems. Based on the theoretical transformation, we propose a Lyapunov repeated iteration (LyaRI) algorithm. Both objective and subjective experiments are conducted to comprehensively evaluate the performance of LyaRI. The results indicate that, compared to its counterparts, LyaRI achieves better video quality and stability performance, with a 47.74% reduction in the variance of the obtained encoding bitrate.

This letter proposes a pilot-aided joint time synchronization and channel estimation (JTSCE) algorithm for orthogonal time frequency space (OTFS) systems. Unlike existing algorithms, JTSCE employs a maximum length sequence (MLS) rather than an isolated signal as the pilot. Distinctively, JTSCE explores MLS's autocorrelation properties to estimate timing offset and channel delay taps. After obtaining delay taps, closed-form expressions of Doppler and channel gain for each propagation path are derived. Simulation results indicate that, compared to its counterpart, JTSCE achieves better bit error rate performance, close to that with perfect time synchronization and channel state information.

Near-space information networks (NSINs) composed of high-altitude platforms (HAPs) and high- and low-altitude unmanned aerial vehicles (UAVs) are a new regime for providing quick, robust, and cost-efficient sensing and communication services. Precipitated by innovations and breakthroughs in manufacturing, materials, communications, electronics, and control techniques, NSINs have been envisioned as an essential component of the emerging sixth-generation of mobile communication systems. This article reveals some critical issues needing to be tackled in NSINs through conducting experiments and discusses the latest advances in NSINs in the research areas of channel modeling, networking, and transmission from a forward-looking, comparative, and technical evolutionary perspective. In this article, we highlight the characteristics of NSINs and present the promising use cases of NSINs. The impact of airborne platforms' unstable movements on the phase delays of onboard antenna arrays with diverse structures is mathematically analyzed. The recent advances in HAP channel modeling are elaborated on, along with the significant differences between HAP and UAV channel modeling. A comprehensive review of the networking techniques of NSINs in network deployment, handoff management, and network management aspects is provided. Besides, the promising techniques and communication protocols of the physical (PHY) layer, medium access control (MAC) layer, network layer, and transport layer of NSINs for achieving efficient transmission over NSINs are reviewed, and we have conducted experiments with practical NSINs to verify the performance of some techniques. Finally, we outline some open issues and promising directions for NSINs deserved for future study and discuss the corresponding challenges.

Small CNN-based models usually require transferring knowledge from a large model before they are deployed in computationally resource-limited edge devices. Masked image modeling (MIM) methods achieve great success in various visual tasks but remain largely unexplored in knowledge distillation for heterogeneous deep models. The reason is mainly due to the significant discrepancy between the Transformer-based large model and the CNN-based small network. In this paper, we develop the first Heterogeneous Generative Knowledge Distillation (H-GKD) based on MIM, which can efficiently transfer knowledge from large Transformer models to small CNN-based models in a generative self-supervised fashion. Our method builds a bridge between Transformer-based models and CNNs by training a UNet-style student with sparse convolution, which can effectively mimic the visual representation inferred by a teacher over masked modeling. Our method is a simple yet effective learning paradigm to learn the visual representation and distribution of data from heterogeneous teacher models, which can be pre-trained using advanced generative methods. Extensive experiments show that it adapts well to various models and sizes, consistently achieving state-of-the-art performance in image classification, object detection, and semantic segmentation tasks. For example, in the Imagenet 1K dataset, H-GKD improves the accuracy of Resnet50 (sparse) from 76.98% to 80.01%.

In this paper, we focus on developing knowledge distillation (KD) for compact 3D detectors. We observe that off-the-shelf KD methods manifest their efficacy only when the teacher model and student counterpart share similar intermediate feature representations. This might explain why they are less effective in building extreme-compact 3D detectors where significant representation disparity arises due primarily to the intrinsic sparsity and irregularity in 3D point clouds. This paper presents a novel representation disparity-aware distillation (RDD) method to address the representation disparity issue and reduce performance gap between compact students and over-parameterized teachers. This is accomplished by building our RDD from an innovative perspective of information bottleneck (IB), which can effectively minimize the disparity of proposal region pairs from student and teacher in features and logits. Extensive experiments are performed to demonstrate the superiority of our RDD over existing KD methods. For example, our RDD increases mAP of CP-Voxel-S to 57.1% on nuScenes dataset, which even surpasses teacher performance while taking up only 42% FLOPs., Comment: Accepted by ICCV2023. arXiv admin note: text overlap with arXiv:2205.15156 by other authors

This paper is concerned with the issue of improving video subscribers' quality of experience (QoE) by deploying a multi-unmanned aerial vehicle (UAV) network. Different from existing works, we characterize subscribers' QoE by video bitrates, latency, and frame freezing and propose to improve their QoE by energy-efficiently and dynamically optimizing the multi-UAV network in terms of serving UAV selection, UAV trajectory, and UAV transmit power. The dynamic multi-UAV network optimization problem is formulated as a challenging sequential-decision problem with the goal of maximizing subscribers' QoE while minimizing the total network power consumption, subject to some physical resource constraints. We propose a novel network optimization algorithm to solve this challenging problem, in which a Lyapunov technique is first explored to decompose the sequential-decision problem into several repeatedly optimized sub-problems to avoid the curse of dimensionality. To solve the sub-problems, iterative and approximate optimization mechanisms with provable performance guarantees are then developed. Finally, we design extensive simulations to verify the effectiveness of the proposed algorithm. Simulation results show that the proposed algorithm can effectively improve the QoE of subscribers and is 66.75\% more energy-efficient than benchmarks.

Real-time object detection plays a vital role in various computer vision applications. However, deploying real-time object detectors on resource-constrained platforms poses challenges due to high computational and memory requirements. This paper describes a low-bit quantization method to build a highly efficient one-stage detector, dubbed as Q-YOLO, which can effectively address the performance degradation problem caused by activation distribution imbalance in traditional quantized YOLO models. Q-YOLO introduces a fully end-to-end Post-Training Quantization (PTQ) pipeline with a well-designed Unilateral Histogram-based (UH) activation quantization scheme, which determines the maximum truncation values through histogram analysis by minimizing the Mean Squared Error (MSE) quantization errors. Extensive experiments on the COCO dataset demonstrate the effectiveness of Q-YOLO, outperforming other PTQ methods while achieving a more favorable balance between accuracy and computational cost. This research contributes to advancing the efficient deployment of object detection models on resource-limited edge devices, enabling real-time detection with reduced computational and memory overhead.

Neural architecture search (NAS) proves to be among the effective approaches for many tasks by generating an application-adaptive neural architecture, which is still challenged by high computational cost and memory consumption. At the same time, 1-bit convolutional neural networks (CNNs) with binary weights and activations show their potential for resource-limited embedded devices. One natural approach is to use 1-bit CNNs to reduce the computation and memory cost of NAS by taking advantage of the strengths of each in a unified framework, while searching the 1-bit CNNs is more challenging due to the more complicated processes involved. In this paper, we introduce Discrepant Child-Parent Neural Architecture Search (DCP-NAS) to efficiently search 1-bit CNNs, based on a new framework of searching the 1-bit model (Child) under the supervision of a real-valued model (Parent). Particularly, we first utilize a Parent model to calculate a tangent direction, based on which the tangent propagation method is introduced to search the optimized 1-bit Child. We further observe a coupling relationship between the weights and architecture parameters existing in such differentiable frameworks. To address the issue, we propose a decoupled optimization method to search an optimized architecture. Extensive experiments demonstrate that our DCP-NAS achieves much better results than prior arts on both CIFAR-10 and ImageNet datasets. In particular, the backbones achieved by our DCP-NAS achieve strong generalization performance on person re-identification and object detection., Comment: Accepted by International Journal of Computer Vision

Vision transformers (ViTs) quantization offers a promising prospect to facilitate deploying large pre-trained networks on resource-limited devices. Fully-binarized ViTs (Bi-ViT) that pushes the quantization of ViTs to its limit remain largely unexplored and a very challenging task yet, due to their unacceptable performance. Through extensive empirical analyses, we identify the severe drop in ViT binarization is caused by attention distortion in self-attention, which technically stems from the gradient vanishing and ranking disorder. To address these issues, we first introduce a learnable scaling factor to reactivate the vanished gradients and illustrate its effectiveness through theoretical and experimental analyses. We then propose a ranking-aware distillation method to rectify the disordered ranking in a teacher-student framework. Bi-ViT achieves significant improvements over popular DeiT and Swin backbones in terms of Top-1 accuracy and FLOPs. For example, with DeiT-Tiny and Swin-Tiny, our method significantly outperforms baselines by 22.1% and 21.4% respectively, while 61.5x and 56.1x theoretical acceleration in terms of FLOPs compared with real-valued counterparts on ImageNet.

The integration of a near-space information network (NSIN) with the reconfigurable intelligent surface (RIS) is envisioned to significantly enhance the communication performance of future wireless communication systems by proactively altering wireless channels. This paper investigates the problem of deploying a RIS-integrated NSIN to provide energy-efficient, ultra-reliable and low-latency communications (URLLC) services. We mathematically formulate this problem as a resource optimization problem, aiming to maximize the effective throughput and minimize the system power consumption, subject to URLLC and physical resource constraints. The formulated problem is challenging in terms of accurate channel estimation, RIS phase alignment, theoretical analysis, and effective solution. We propose a joint resource allocation algorithm to handle these challenges. In this algorithm, we develop an accurate channel estimation approach by exploring message passing and optimize phase shifts of RIS reflecting elements to further increase the channel gain. Besides, we derive an analysis-friend expression of decoding error probability and decompose the problem into two-layered optimization problems by analyzing the monotonicity, which makes the formulated problem analytically tractable. Extensive simulations have been conducted to verify the performance of the proposed algorithm. Simulation results show that the proposed algorithm can achieve outstanding channel estimation performance and is more energy-efficient than diverse benchmark algorithms.

Despite advancements in generic object detection, there remains a performance gap in detecting small objects compared to normal-scale objects. We for the first time observe that existing bounding box regression methods tend to produce distorted gradients for small objects and result in less accurate localization. To address this issue, we present a novel Confidence-driven Bounding Box Localization (C-BBL) method to rectify the gradients. C-BBL quantizes continuous labels into grids and formulates two-hot ground truth labels. In prediction, the bounding box head generates a confidence distribution over the grids. Unlike the bounding box regression paradigms in conventional detectors, we introduce a classification-based localization objective through cross entropy between ground truth and predicted confidence distribution, generating confidence-driven gradients. Additionally, C-BBL describes a uncertainty loss based on distribution entropy in labels and predictions to further reduce the uncertainty in small object localization. The method is evaluated on multiple detectors using three object detection benchmarks and consistently improves baseline detectors, achieving state-of-the-art performance. We also demonstrate the generalizability of C-BBL to different label systems and effectiveness for high resolution detection, which validates its prospect as a general solution.

The large pre-trained vision transformers (ViTs) have demonstrated remarkable performance on various visual tasks, but suffer from expensive computational and memory cost problems when deployed on resource-constrained devices. Among the powerful compression approaches, quantization extremely reduces the computation and memory consumption by low-bit parameters and bit-wise operations. However, low-bit ViTs remain largely unexplored and usually suffer from a significant performance drop compared with the real-valued counterparts. In this work, through extensive empirical analysis, we first identify the bottleneck for severe performance drop comes from the information distortion of the low-bit quantized self-attention map. We then develop an information rectification module (IRM) and a distribution guided distillation (DGD) scheme for fully quantized vision transformers (Q-ViT) to effectively eliminate such distortion, leading to a fully quantized ViTs. We evaluate our methods on popular DeiT and Swin backbones. Extensive experimental results show that our method achieves a much better performance than the prior arts. For example, our Q-ViT can theoretically accelerates the ViT-S by 6.14x and achieves about 80.9% Top-1 accuracy, even surpassing the full-precision counterpart by 1.0% on ImageNet dataset. Our codes and models are attached on https://github.com/YanjingLi0202/Q-ViT, Comment: Accepted by NeurIPS2022

Consider a processor having access only to meta-data consisting of the timings of data packets and acknowledgment (ACK) packets from all nodes in a network. The meta-data report the source node of each packet, but not the destination nodes or the contents of the packets. The goal of the processor is to infer the network topology based solely on such information. Prior work leveraged causality metrics to identify which links are active. If the data timings and ACK timings of two nodes -- say node 1 and node 2, respectively -- are causally related, this may be taken as evidence that node 1 is communicating to node 2 (which sends back ACK packets to node 1). This paper starts with the observation that packet losses can weaken the causality relationship between data and ACK timing streams. To obviate this problem, a new Expectation Maximization (EM)-based algorithm is introduced -- EM-causality discovery algorithm (EM-CDA) -- which treats packet losses as latent variables. EM-CDA iterates between the estimation of packet losses and the evaluation of causality metrics. The method is validated through extensive experiments in wireless sensor networks on the NS-3 simulation platform., Comment: submitted

Knowledge distillation (KD) has been proven to be useful for training compact object detection models. However, we observe that KD is often effective when the teacher model and student counterpart share similar proposal information. This explains why existing KD methods are less effective for 1-bit detectors, caused by a significant information discrepancy between the real-valued teacher and the 1-bit student. This paper presents an Information Discrepancy-aware strategy (IDa-Det) to distill 1-bit detectors that can effectively eliminate information discrepancies and significantly reduce the performance gap between a 1-bit detector and its real-valued counterpart. We formulate the distillation process as a bi-level optimization formulation. At the inner level, we select the representative proposals with maximum information discrepancy. We then introduce a novel entropy distillation loss to reduce the disparity based on the selected proposals. Extensive experiments demonstrate IDa-Det's superiority over state-of-the-art 1-bit detectors and KD methods on both PASCAL VOC and COCO datasets. IDa-Det achieves a 76.9% mAP for a 1-bit Faster-RCNN with ResNet-18 backbone. Our code is open-sourced on https://github.com/SteveTsui/IDa-Det.

In the upcoming 6G era, existing terrestrial networks have evolved toward space-air-ground integrated networks (SAGIN), providing ultra-high data rates, seamless network coverage, and ubiquitous intelligence for communications of applications and services. However, conventional communications in SAGIN still face data confidentiality issues. Fortunately, the concept of Quantum Key Distribution (QKD) over SAGIN is able to provide information-theoretic security for secure communications in SAGIN with quantum cryptography. Therefore, in this paper, we propose the quantum-secured SAGIN which is feasible to achieve proven secure communications using quantum mechanics to protect data channels between space, air, and ground nodes. Moreover, we propose a universal QKD service provisioning framework to minimize the cost of QKD services under the uncertainty and dynamics of communications in quantum-secured SAGIN. In this framework, fiber-based QKD services are deployed in passive optical networks with the advantages of low loss and high stability. Moreover, the widely covered and flexible satellite- and UAV-based QKD services are provisioned as a supplement during the real-time data transmission phase. Finally, to examine the effectiveness of the proposed concept and framework, a case study of quantum-secured SAGIN in the Metaverse is conducted where uncertain and dynamic factors of the secure communications in Metaverse applications are effectively resolved in the proposed framework.

The development of emerging applications, such as autonomous transportation systems, are expected to result in an explosive growth in mobile data traffic. As the available spectrum resource becomes more and more scarce, there is a growing need for a paradigm shift from Shannon's Classical Information Theory (CIT) to semantic communication (SemCom). Specifically, the former adopts a "transmit-before-understanding" approach while the latter leverages artificial intelligence (AI) techniques to "understand-before-transmit", thereby alleviating bandwidth pressure by reducing the amount of data to be exchanged without negating the semantic effectiveness of the transmitted symbols. However, the semantic extraction (SE) procedure incurs costly computation and storage overheads. In this article, we introduce an edge-driven training, maintenance, and execution of SE. We further investigate how edge intelligence can be enhanced with SemCom through improving the generalization capabilities of intelligent agents at lower computation overheads and reducing the communication overhead of information exchange. Finally, we present a case study involving semantic-aware resource optimization for the wireless powered Internet of Things (IoT).

Dubbed "the successor to the mobile Internet", the concept of the Metaverse has recently exploded in popularity. While there exists lite versions of the Metaverse today, we are still far from realizing the vision of a seamless, shardless, and interoperable Metaverse given the stringent sensing, communication, and computation requirements. Moreover, the birth of the Metaverse comes amid growing privacy concerns among users. In this article, we begin by providing a preliminary definition of the Metaverse. We discuss the architecture of the Metaverse and mainly focus on motivating the convergence of edge intelligence and the infrastructure layer of the Metaverse. We present major edge-based technological developments and their integration to support the Metaverse engine. Then, we present our research attempts through a case study of virtual city development in the Metaverse. Finally, we discuss the open research issues., Comment: 9 pages, 5 figures

Internet of unmanned aerial vehicle (I-UAV) networks promise to accomplish sensing and transmission tasks quickly, robustly, and cost-efficiently via effective cooperation among UAVs. To achieve the promising benefits, the crucial I-UAV networking issue should be tackled. This article argues that I-UAV networking can be classified into three categories, quality-of-service (QoS) driven networking, quality-of-experience (QoE) driven networking, and situation aware networking. Each category of networking poses emerging challenges which have severe effects on the safe and efficient accomplishment of I-UAV missions. This article elaborately analyzes these challenges and expounds on the corresponding intelligent approaches to tackle the I-UAV networking issue. Besides, considering the uplifting effect of extending the scalability of I-UAV networks through cooperating with high altitude platforms (HAPs), this article gives an overview of the integrated HAP and I-UAV networks and presents the corresponding networking challenges and intelligent approaches.

In this paper, we investigate an unmanned aerial vehicle (UAV)-assisted Internet-of-Things (IoT) system in a sophisticated three-dimensional (3D) environment, where the UAV's trajectory is optimized to efficiently collect data from multiple IoT ground nodes. Unlike existing approaches focusing only on a simplified two-dimensional scenario and the availability of perfect channel state information (CSI), this paper considers a practical 3D urban environment with imperfect CSI, where the UAV's trajectory is designed to minimize data collection completion time subject to practical throughput and flight movement constraints. Specifically, inspired from the state-of-the-art deep reinforcement learning approaches, we leverage the twin-delayed deep deterministic policy gradient (TD3) to design the UAV's trajectory and present a TD3-based trajectory design for completion time minimization (TD3-TDCTM) algorithm. In particular, we set an additional information, i.e., the merged pheromone, to represent the state information of UAV and environment as a reference of reward which facilitates the algorithm design. By taking the service statuses of IoT nodes, the UAV's position, and the merged pheromone as input, the proposed algorithm can continuously and adaptively learn how to adjust the UAV's movement strategy. By interacting with the external environment in the corresponding Markov decision process, the proposed algorithm can achieve a near-optimal navigation strategy. Our simulation results show the superiority of the proposed TD3-TDCTM algorithm over three conventional non-learning based baseline methods., Comment: Accepted by IEEE Internet of Things Journal. The codes and some other materials about this work may be available at https://gaozhen16.github.io/

This paper investigates the problem of providing ultra-reliable and energy-efficient virtual reality (VR) experiences for wireless mobile users. To ensure reliable ultra-high-definition (UHD) video frame delivery to mobile users and enhance their immersive visual experiences, a coordinated multipoint (CoMP) transmission technique and millimeter wave (mmWave) communications are exploited. Owing to user movement and time-varying wireless channels, the wireless VR experience enhancement problem is formulated as a sequence-dependent and mixed-integer problem with a goal of maximizing users' feeling of presence (FoP) in the virtual world, subject to power consumption constraints on access points (APs) and users' head-mounted displays (HMDs). The problem, however, is hard to be directly solved due to the lack of users' accurate tracking information and the sequence-dependent and mixed-integer characteristics. To overcome this challenge, we develop a parallel echo state network (ESN) learning method to predict users' tracking information by training fresh and historical tracking samples separately collected by APs. With the learnt results, we propose a deep reinforcement learning (DRL) based optimization algorithm to solve the formulated problem. In this algorithm, we implement deep neural networks (DNNs) as a scalable solution to produce integer decision variables and solving a continuous power control problem to criticize the integer decision variables. Finally, the performance of the proposed algorithm is compared with various benchmark algorithms, and the impact of different design parameters is also discussed. Simulation results demonstrate that the proposed algorithm is more 4.14% energy-efficient than the benchmark algorithms.

In this paper, we investigate a private and cache-enabled unmanned aerial vehicle (UAV) network for content provision. Aiming at delivering fresh, fair, and energy-efficient content files to terrestrial users, we formulate a joint UAV caching, UAV trajectory, and UAV transmit power optimization problem. This problem is confirmed to be a sequential decision problem with mixed-integer non-convex constraints, which is intractable directly. To this end, we propose a novel algorithm based on the techniques of subproblem decomposition and convex approximation. Particularly, we first propose to decompose the sequential decision problem into multiple repeated optimization subproblems via a Lyapunov technique. Next, an iterative optimization scheme incorporating a successive convex approximation (SCA) technique is explored to tackle the challenging mixed-integer non-convex subproblems. Besides, we analyze the convergence and computational complexity of the proposed algorithm and derive the theoretical value of the expected peak age of information (PAoI) to estimate the content freshness. Simulation results demonstrate that the proposed algorithm can achieve the expected PAoI close to the theoretical value and is more 22.11% and 70.51% energy-efficient and fairer than benchmark algorithms.

This letter is concerned with power control for a ultra-reliable and low-latency communications (URLLC) enabled unmanned aerial vehicle (UAV) system incorporated with deep neural network (DNN) based channel estimation. Particularly, we formulate the power control problem for the UAV system as an optimization problem to accommodate the URLLC requirement of uplink control and non-payload signal delivery while ensuring the downlink high-speed payload transmission. This problem is challenging to be solved due to the requirement of analytically tractable channel models and the non-convex characteristic as well. To address the challenges, we propose a novel power control algorithm, which constructs analytically tractable channel models based on DNN estimation results and explores a semidefinite relaxation (SDR) scheme to tackle the non-convexity. Simulation results demonstrate the accuracy of the DNN estimation and verify the effectiveness of the proposed algorithm.

Links in most real networks often change over time. Such temporality of links encodes the ordering and causality of interactions between nodes and has a profound effect on network dynamics and function. Empirical evidences have shown that the temporal nature of links in many real-world networks is not random. Nonetheless, it is challenging to predict temporal link patterns while considering the entanglement between topological and temporal link patterns. Here we propose an entropy-rate based framework, based on combined topological-temporal regularities, for quantifying the predictability of any temporal network. We apply our framework on various model networks, demonstrating that it indeed captures the intrinsic topological-temporal regularities whereas previous methods considered only temporal aspects. We also apply our framework on 18 real networks of different types and determine their predictability. Interestingly, we find that for most real temporal networks, despite the greater complexity of predictability brought by the increase in dimension the combined topological-temporal predictability is higher than the temporal predictability. Our results demonstrate the necessity of incorporating both temporal and topological aspects of networks in order to improve predictions of dynamical processes.

In this paper, a full-duplex unmanned aerial vehicle (FD-UAV) relay is employed to increase the communication capacity of millimeter-wave (mmWave) networks. Large antenna arrays are equipped at the source node (SN), destination node (DN), and FD-UAV relay to overcome the high path loss of mmWave channels and to help mitigate the self-interference at the FD-UAV relay. Specifically, we formulate a problem for maximization of the achievable rate from the SN to the DN, where the UAV position, analog beamforming, and power control are jointly optimized. Since the problem is highly non-convex and involves high-dimensional, highly coupled variable vectors, we first obtain the conditional optimal position of the FD-UAV relay for maximization of an approximate upper bound on the achievable rate in closed form, under the assumption of a line-of-sight (LoS) environment and ideal beamforming. Then, the UAV is deployed to the position which is closest to the conditional optimal position and yields LoS paths for both air-to-ground links. Subsequently, we propose an alternating interference suppression (AIS) algorithm for the joint design of the beamforming vectors and the power control variables. In each iteration, the beamforming vectors are optimized for maximization of the beamforming gains of the target signals and the successive reduction of the interference, where the optimal power control variables are obtained in closed form. Our simulation results confirm the superiority of the proposed positioning, beamforming, and power control method compared to three benchmark schemes. Furthermore, our results show that the proposed solution closely approaches a performance upper bound for mmWave FD-UAV systems., Comment: This paper has been accepted by IEEE Journal on Selected Areas in Communications, special issue on Multiple Antenna Technologies for Beyond 5G

Most of the recent advances in crowd counting have evolved from hand-designed density estimation networks, where multi-scale features are leveraged to address the scale variation problem, but at the expense of demanding design efforts. In this work, we automate the design of counting models with Neural Architecture Search (NAS) and introduce an end-to-end searched encoder-decoder architecture, Automatic Multi-Scale Network (AMSNet). Specifically, we utilize a counting-specific two-level search space. The encoder and decoder in AMSNet are composed of different cells discovered from micro-level search, while the multi-path architecture is explored through macro-level search. To solve the pixel-level isolation issue in MSE loss, AMSNet is optimized with an auto-searched Scale Pyramid Pooling Loss (SPPLoss) that supervises the multi-scale structural information. Extensive experiments on four datasets show AMSNet produces state-of-the-art results that outperform hand-designed models, fully demonstrating the efficacy of NAS-Count., Comment: Accepted to European Conference on Computer Vision(ECCV) 2020

This paper is concerned with slicing a radio access network (RAN) for simultaneously serving two typical 5G and beyond use cases, i.e., enhanced mobile broadband (eMBB) and ultra-reliable and low latency communications (URLLC). Although many researches have been conducted to tackle this issue, few of them have considered the impact of bursty URLLC. The bursty characteristic of URLLC traffic may significantly increase the difficulty of RAN slicing on the aspect of ensuring a ultra-low packet blocking probability. To reduce the packet blocking probability, we re-visit the structure of physical resource blocks (PRBs) orchestrated for bursty URLLC traffic in the time-frequency plane based on our theoretical results. Meanwhile, we formulate the problem of slicing a RAN enabling coordinated multi-point (CoMP) transmissions for multicast eMBB and bursty URLLC service multiplexing as a multi-timescale optimization problem. The goal of this problem is to maximize multicast eMBB and bursty URLLC slice utilities, subject to physical resource constraints. To mitigate this thorny multi-timescale problem, we transform it into multiple single timescale problems by exploring the fundamental principle of a sample average approximation (SAA) technique. Next, an iterative algorithm with provable performance guarantees is developed to obtain solutions to these single timescale problems and aggregate the obtained solutions into those of the multi-timescale problem. We also design a prototype for the CoMP-enabled RAN slicing system incorporating with multicast eMBB and bursty URLLC traffic and compare the proposed iterative algorithm with the state-of-the-art algorithm to verify the effectiveness of the algorithm.

Future wireless networks are envisioned to serve massive Internet of things (mIoT) via some radio access technologies, where the random access channel (RACH) procedure should be exploited for IoT devices to access the networks. However, the theoretical analysis of the RACH procedure for massive IoT devices is challenging. To address this challenge, we first correlate the RACH request of an IoT device with the status of its maintained queue and analyze the evolution of the queue status. Based on the analysis result, we then derive the closed-form expression of the random access (RA) success probability, which is a significant indicator characterizing the RACH procedure of the device. Besides, considering the agreement on converging different services onto a shared infrastructure, we investigate the RAN slicing for mIoT and bursty ultra-reliable and low latency communications (URLLC) service multiplexing. Specifically, we formulate the RAN slicing problem as an optimization one to maximize the total RA success probabilities of all IoT devices and provide URLLC services for URLLC devices in an energy-efficient way. A slice resource optimization (SRO) algorithm exploiting relaxation and approximation with provable tightness and error bound is then proposed to mitigate the optimization problem. Simulation results demonstrate that the proposed SRO algorithm can effectively implement the service multiplexing of mIoT and bursty URLLC traffic.

This paper is concerned with the resource allocation in a multi-unmanned aerial vehicle (UAV)-aided network for providing enhanced mobile broadband (eMBB) services for user equipments. Different from most of the existing network resource allocation approaches, we investigate a joint non-orthogonal user association, subchannel allocation and power control problem. The objective of the problem is to maximize the network energy efficiency under the constraints on user equipments' quality of service, UAVs' network capacity and power consumption. We formulate the energy efficiency maximization problem as a challenging mixed-integer non-convex programming problem. To alleviate this problem, we first decompose the original problem into two subproblems, namely, an integer non-linear user association and subchannel allocation subproblem and a non-convex power control subproblem. We then design a two-stage approximation strategy to handle the non-linearity of the user association and subchannel allocation subproblem and exploit a successive convex approximation approach to tackle the non-convexity of the power control subproblem. Based on the derived results, we develop an iterative algorithm with provable convergence to mitigate the original problem. Simulation results show that our proposed framework can improve energy efficiency compared with several benchmark algorithms.

Unmanned aerial vehicle (UAV) networks are convinced as a significant part of 5G and emerging 6G wireless networks. UAV slicing is a promising proposal of converging different services onto a common UAV network without deploying individual network solution for each type of service. This paper is concerned with UAV slicing for providing energy-efficient and fair service coverage for enhanced mobile broadband (eMBB) users (UEs). Aiming at physically configuring UAV slices, the UAV slicing problem is formulated as a time-dependent mixed-integer-non-convex programming problem with a goal of maximizing all UEs' data rates while minimizing UAVs' total transmit power. To mitigate this challenging problem, we first decompose the original problem into two time-dependent subproblems using a Lyapunov approach. We then derive the procedure of tackling the non-convexity and the mixed-integer property of the subproblems by exploring a successive convex approximate (SCA) method and an alternative optimization scheme, respectively. Based on the derived results, we develop an algorithm with provable performance guarantees to mitigate the two subproblems repeatedly., Comment: The original title is: Repeatedly Energy-Efficient and Fair Service Coverage: UAV Slicing

Future wireless networks are convinced to provide flexible and cost-efficient services via exploiting network slicing techniques. However, it is challenging to configure network slicing systems for bursty ultra-reliable and low latency communications (URLLC) service provision due to its stringent requirements on low packet blocking probability and low codeword error decoding probability. In this paper, we propose to orchestrate network resources for a network slicing system to guarantee a more reliable bursty URLLC service provision. We re-cut physical resource blocks (PRBs) and derive the minimum upper bound of bandwidth for URLLC transmission with a low packet blocking probability. We correlate coordinated multipoint (CoMP) beamforming with channel uses and derive the minimum upper bound of channel uses for URLLC transmission with a low codeword error decoding probability. Considering the agreement on converging diverse services onto shared infrastructures, we further investigate the network slicing for URLLC and enhanced mobile broadband (eMBB) service multiplexing. Particularly, we formulate the service multiplexing as an optimization problem to maximize the long-term total slice utility. The mitigation of this problem is challenging due to the requirements of future channel information and tackling a two timescale issue. To address the challenges, we develop a joint resource optimization algorithm based on a sample average approximate (SAA) technique and a distributed optimization method with provable performance guarantees.

This paper investigates the application of non-orthogonal multiple access in millimeter-Wave communications (mmWave-NOMA). Particularly, we consider downlink transmission with a hybrid beamforming structure. A user grouping algorithm is first proposed according to the channel correlations of the users. Whereafter, a joint hybrid beamforming and power allocation problem is formulated to maximize the achievable sum rate, subject to a minimum rate constraint for each user. To solve this non-convex problem with high-dimensional variables, we first obtain the solution of power allocation under arbitrary fixed hybrid beamforming, which is divided into intra-group power allocation and inter-group power allocation. Then, given arbitrary fixed analog beamforming, we utilize the approximate zero-forcing method to design the digital beamforming to minimize the inter-group interference. Finally, the analog beamforming problem with the constant-modulus constraint is solved with a proposed boundary-compressed particle swarm optimization algorithm. Simulation results show that the proposed joint approach, including user grouping, hybrid beamforming and power allocation, outperforms the state-of-the-art schemes and the conventional mmWave orthogonal multiple access system in terms of achievable sum rate and energy efficiency., Comment: Accepted for publication in IEEE Trans. Wireless Comm

Crowd counting has recently attracted increasing interest in computer vision but remains a challenging problem. In this paper, we propose a trellis encoder-decoder network (TEDnet) for crowd counting, which focuses on generating high-quality density estimation maps. The major contributions are four-fold. First, we develop a new trellis architecture that incorporates multiple decoding paths to hierarchically aggregate features at different encoding stages, which can handle large variations of objects. Second, we design dense skip connections interleaved across paths to facilitate sufficient multi-scale feature fusions and to absorb the supervision information. Third, we propose a new combinatorial loss to enforce local coherence and spatial correlation in density maps. By distributedly imposing this combinatorial loss on intermediate outputs, gradient vanishing can be largely alleviated for better back-propagation and faster convergence. Finally, our TEDnet achieves new state-of-the art performance on four benchmarks, with an improvement up to 14% in terms of MAE., Comment: CVPR 2019, Accepted

Being able to track an anonymous object, a model-free tracker is comprehensively applicable regardless of the target type. However, designing such a generalized framework is challenged by the lack of object-oriented prior information. As one solution, a real-time model-free object tracking approach is designed in this work relying on Convolutional Neural Networks (CNNs). To overcome the object-centric information scarcity, both appearance and motion features are deeply integrated by the proposed AMNet, which is an end-to-end offline trained two-stream network. Between the two parallel streams, the ANet investigates appearance features with a multi-scale Siamese atrous CNN, enabling the tracking-by-matching strategy. The MNet achieves deep motion detection to localize anonymous moving objects by processing generic motion features. The final tracking result at each frame is generated by fusing the output response maps from both sub-networks. The proposed AMNet reports leading performance on both OTB and VOT benchmark datasets with favorable real-time processing speed., Comment: 10 pages, 10 figures. Accepted in IEEE Winter Conference on Applications of Computer Vision (WACV), 2019

The advancement of deep convolutional neural networks (DCNNs) has driven significant improvement in the accuracy of recognition systems for many computer vision tasks. However, their practical applications are often restricted in resource-constrained environments. In this paper, we introduce projection convolutional neural networks (PCNNs) with a discrete back propagation via projection (DBPP) to improve the performance of binarized neural networks (BNNs). The contributions of our paper include: 1) for the first time, the projection function is exploited to efficiently solve the discrete back propagation problem, which leads to a new highly compressed CNNs (termed PCNNs); 2) by exploiting multiple projections, we learn a set of diverse quantized kernels that compress the full-precision kernels in a more efficient way than those proposed previously; 3) PCNNs achieve the best classification performance compared to other state-of-the-art BNNs on the ImageNet and CIFAR datasets.

In this paper, we investigate the combination of non-orthogonal multiple access and millimeter-Wave communications (mmWave-NOMA). A downlink cellular system is considered, where an analog phased array is equipped at both the base station and users. A joint Tx-Rx beamforming and power allocation problem is formulated to maximize the achievable sum rate (ASR) subject to a minimum rate constraint for each user. As the problem is non-convex, we propose a sub-optimal solution with three stages. In the first stage, the optimal power allocation with a closed form is obtained for an arbitrary fixed Tx-Rx beamforming. In the second stage, the optimal Rx beamforming with a closed form is designed for an arbitrary fixed Tx beamforming. In the third stage, the original problem is reduced to a Tx beamforming problem by using the previous results, and a boundary-compressed particle swarm optimization (BC-PSO) algorithm is proposed to obtain a sub-optimal solution. Extensive performance evaluations are conducted to verify the rational of the proposed solution, and the results show that the proposed sub-optimal solution can achieve a near-upper-bound performance in terms of ASR, which is significantly improved compared with those of the state-of-the-art schemes and the conventional mmWave orthogonal multiple access (mmWave-OMA) system., Comment: 11Pages, 10 figures

Crowd counting usually addressed by density estimation becomes an increasingly important topic in computer vision due to its widespread applications in video surveillance, urban planning, and intelligence gathering. However, it is essentially a challenging task because of the greatly varied sizes of objects, coupled with severe occlusions and vague appearance of extremely small individuals. Existing methods heavily rely on multi-column learning architectures to extract multi-scale features, which however suffer from heavy computational cost, especially undesired for crowd counting. In this paper, we propose the single-column counting network (SCNet) for efficient crowd counting without relying on multi-column networks. SCNet consists of residual fusion modules (RFMs) for multi-scale feature extraction, a pyramid pooling module (PPM) for information fusion, and a sub-pixel convolutional module (SPCM) followed by a bilinear upsampling layer for resolution recovery. Those proposed modules enable our SCNet to fully capture multi-scale features in a compact single-column architecture and estimate high-resolution density map in an efficient way. In addition, we provide a principled paradigm for density map generation and data augmentation for training, which shows further improved performance. Extensive experiments on three benchmark datasets show that our SCNet delivers new state-of-the-art performance and surpasses previous methods by large margins, which demonstrates the great effectiveness of SCNet as a single-column network for crowd counting.

Compression artifacts reduction (CAR) is a challenging problem in the field of remote sensing. Most recent deep learning based methods have demonstrated superior performance over the previous hand-crafted methods. In this paper, we propose an end-to-end one-two-one (OTO) network, to combine different deep models, i.e., summation and difference models, to solve the CAR problem. Particularly, the difference model motivated by the Laplacian pyramid is designed to obtain the high frequency information, while the summation model aggregates the low frequency information. We provide an in-depth investigation into our OTO architecture based on the Taylor expansion, which shows that these two kinds of information can be fused in a nonlinear scheme to gain more capacity of handling complicated image compression artifacts, especially the blocking effect in compression. Extensive experiments are conducted to demonstrate the superior performance of the OTO networks, as compared to the state-of-the-arts on remote sensing datasets and other benchmark datasets.

In this paper, we investigate the combination of two key enabling technologies for the fifth generation (5G) wireless mobile communication, namely millimeter-wave (mmWave) communications and non-orthogonal multiple access (NOMA). In particular, we consider a typical 2-user uplink mmWave-NOMA system, where the base station (BS) equips an analog beamforming structure with a single RF chain and serves 2 NOMA users. An optimization problem is formulated to maximize the achievable sum rate of the 2 users while ensuring a minimal rate constraint for each user. The problem turns to be a joint power control and beamforming problem, i.e., we need to find the beamforming vectors to steer to the two users simultaneously subject to an analog beamforming structure, and meanwhile control appropriate power on them. As direct search for the optimal solution of the non-convex problem is too complicated, we propose to decompose the original problem into two sub-problems that are relatively easy to solve: one is a power control and beam gain allocation problem, and the other is an analog beamforming problem under a constant-modulus constraint. The rational of the proposed solution is verified by extensive simulations, and the performance evaluation results show that the proposed sub-optimal solution achieve a close-to-bound uplink sum-rate performance., Comment: 12 pages, 9 figures. arXiv admin note: substantial text overlap with arXiv:1711.01380

In this paper, we present YoTube-a novel network fusion framework for searching action proposals in untrimmed videos, where each action proposal corresponds to a spatialtemporal video tube that potentially locates one human action. Our method consists of a recurrent YoTube detector and a static YoTube detector, where the recurrent YoTube explores the regression capability of RNN for candidate bounding boxes predictions using learnt temporal dynamics and the static YoTube produces the bounding boxes using rich appearance cues in a single frame. Both networks are trained using rgb and optical flow in order to fully exploit the rich appearance, motion and temporal context, and their outputs are fused to produce accurate and robust proposal boxes. Action proposals are finally constructed by linking these boxes using dynamic programming with a novel trimming method to handle the untrimmed video effectively and efficiently. Extensive experiments on the challenging UCF-101 and UCF-Sports datasets show that our proposed technique obtains superior performance compared with the state-of-the-art.