Haitao Zhao

Kang Wei, Jun Li, Chuan Ma et al.

Federated learning (FL) enables distributed clients to collaboratively train a machine learning model without sharing raw data with each other. However, it suffers the leakage of private information from uploading models. In addition, as the model size grows, the training latency increases due to limited transmission bandwidth and the model performance degrades while using differential privacy (DP) protection. In this paper, we propose a gradient sparsification empowered FL framework over wireless channels, in order to improve training efficiency without sacrificing convergence performance. Specifically, we first design a random sparsification algorithm to retain a fraction of the gradient elements in each client's local training, thereby mitigating the performance degradation induced by DP and and reducing the number of transmission parameters over wireless channels. Then, we analyze the convergence bound of the proposed algorithm, by modeling a non-convex FL problem. Next, we formulate a time-sequential stochastic optimization problem for minimizing the developed convergence bound, under the constraints of transmit power, the average transmitting delay, as well as the client's DP requirement. Utilizing the Lyapunov drift-plus-penalty framework, we develop an analytical solution to the optimization problem. Extensive experiments have been implemented on three real life datasets to demonstrate the effectiveness of our proposed algorithm. We show that our proposed algorithms can fully exploit the interworking between communication and computation to outperform the baselines, i.e., random scheduling, round robin and delay-minimization algorithms.

Zhongze Wang, Haitao Zhao, Lujian Yao et al.

In image dehazing task, haze density is a key feature and affects the performance of dehazing methods. However, some of the existing methods lack a comparative image to measure densities, and others create intermediate results but lack the exploitation of their density differences, which can facilitate perception of density. To address these deficiencies, we propose a density-aware dehazing method named Density Feature Refinement Network (DFR-Net) that extracts haze density features from density differences and leverages density differences to refine density features. In DFR-Net, we first generate a proposal image that has lower overall density than the hazy input, bringing in global density differences. Additionally, the dehazing residual of the proposal image reflects the level of dehazing performance and provides local density differences that indicate localized hard dehazing or high density areas. Subsequently, we introduce a Global Branch (GB) and a Local Branch (LB) to achieve density-awareness. In GB, we use Siamese networks for feature extraction of hazy inputs and proposal images, and we propose a Global Density Feature Refinement (GDFR) module that can refine features by pushing features with different global densities further away. In LB, we explore local density features from the dehazing residuals between hazy inputs and proposal images and introduce an Intermediate Dehazing Residual Feedforward (IDRF) module to update local features and pull them closer to clear image features. Sufficient experiments demonstrate that the proposed method achieves results beyond the state-of-the-art methods on various datasets.

Jingchao Peng, Haitao Zhao, Zhengwei Hu

For both visible and infrared images have their own advantages and disadvantages, RGBT tracking has attracted more and more attention. The key points of RGBT tracking lie in feature extraction and feature fusion of visible and infrared images. Current RGBT tracking methods mostly pay attention to both individual features (features extracted from images of a single camera) and common features (features extracted and fused from an RGB camera and a thermal camera), while pay less attention to the different and dynamic contributions of individual features and common features for different sequences of registered image pairs. This paper proposes a novel RGBT tracking method, called Dynamic Fusion Network (DFNet), which adopts a two-stream structure, in which two non-shared convolution kernels are employed in each layer to extract individual features. Besides, DFNet has shared convolution kernels for each layer to extract common features. Non-shared convolution kernels and shared convolution kernels are adaptively weighted and summed according to different image pairs, so that DFNet can deal with different contributions for different sequences. DFNet has a fast speed, which is 28.658 FPS. The experimental results show that when DFNet only increases the Mult-Adds of 0.02% than the non-shared-convolution-kernel-based fusion method, Precision Rate (PR) and Success Rate (SR) reach 88.1% and 71.9% respectively.

Xiaochen Zhang, Haitao Zhao, Jun Xiong et al.

Machine learning (ML) has been widely used for efficient resource allocation (RA) in wireless networks. Although superb performance is achieved on small and simple networks, most existing ML-based approaches are confronted with difficulties when heterogeneity occurs and network size expands. In this paper, specifically focusing on power control/beamforming (PC/BF) in heterogeneous device-to-device (D2D) networks, we propose a novel unsupervised learning-based framework named heterogeneous interference graph neural network (HIGNN) to handle these challenges. First, we characterize diversified link features and interference relations with heterogeneous graphs. Then, HIGNN is proposed to empower each link to obtain its individual transmission scheme after limited information exchange with neighboring links. It is noteworthy that HIGNN is scalable to wireless networks of growing sizes with robust performance after trained on small-sized networks. Numerical results show that compared with state-of-the-art benchmarks, HIGNN achieves much higher execution efficiency while providing strong performance.

Yuru Chen, Haitao Zhao, Zhengwei Hu

Depth estimation is a traditional computer vision task, which plays a crucial role in understanding 3D scene geometry. Recently, deep-convolutional-neural-networks based methods have achieved promising results in the monocular depth estimation field. Specifically, the framework that combines the multi-scale features extracted by the dilated convolution based block (atrous spatial pyramid pooling, ASPP) has gained the significant improvement in the dense labeling task. However, the discretized and predefined dilation rates cannot capture the continuous context information that differs in diverse scenes and easily introduce the grid artifacts in depth estimation. In this paper, we propose an attention-based context aggregation network (ACAN) to tackle these difficulties. Based on the self-attention model, ACAN adaptively learns the task-specific similarities between pixels to model the context information. First, we recast the monocular depth estimation as a dense labeling multi-class classification problem. Then we propose a soft ordinal inference to transform the predicted probabilities to continuous depth values, which can reduce the discretization error (about 1% decrease in RMSE). Second, the proposed ACAN aggregates both the image-level and pixel-level context information for depth estimation, where the former expresses the statistical characteristic of the whole image and the latter extracts the long-range spatial dependencies for each pixel. Third, for further reducing the inconsistency between the RGB image and depth map, we construct an attention loss to minimize their information entropy. We evaluate on public monocular depth-estimation benchmark datasets (including NYU Depth V2, KITTI). The experiments demonstrate the superiority of our proposed ACAN and achieve the competitive results with the state of the arts.