The end-cloud synergy with neural networks offers significant promise for enhancing AI-IoT systems. Numerous studies have been conducted to effectively address the dynamic and resource-constrained nature of IoT; however, challenges persist in adaptability to constraints, horizontal integration between technologies, and system-level optimization. In this article, we propose an end-cloud collaborative system (EECS) with an on-demand offloading mechanism for AI-IoT. We enhance three critical areas previously narrowly explored: 1) model pruning to fit into the end device’s constraints; 2) dynamic data transmission between end and cloud; and 3) an effective offloading policy. We pay particular attention to the constraints imposed by multiply-and-add (MAC) operations and bandwidth (BW) at the end, with a focus on image classification tasks. Our contributions are threefold. First, we introduce a two-stage trainable pruning method that can automatically adjust the end model to end constraints and optimize from a system-wide perspective. Second, we propose an adaptive mechanism that accommodates fluctuating BW conditions based on our trainable pruning, integrating static pruning with dynamic inference. Third, we develop a smart offloading policy that enhances decision-making, thereby elevating system-level efficiency. Finally, our EECS shows substantial improvements, achieving$2.6\times $

Snow is a highly complicated atmospheric phenomenon that usually contains snowflake, snow streak, and veiling effect (similar to the haze or the mist). In this literature, we propose a single image desnowing algorithm to address the diversity of snow particles in shape and size. First, to better represent the complex snow shape, we apply the dual-tree wavelet transform and propose a complex wavelet loss in the network. Second, we propose a hierarchical decomposition paradigm in our network for better under-standing the different sizes of snow particles. Last, we propose a novel feature called the contradict channel (CC) for the snow scenes. We find that the regions containing the snow particles tend to have higher intensity in the CC than that in the snow-free regions. We leverage this discriminative feature to construct the contradict channel loss for improving the performance of snow removal. Moreover, due to the limitation of existing snow datasets, to simulate the snow scenarios comprehensively, we propose a large-scale dataset called Comprehensive Snow Dataset (CSD). Experimental results show that the proposed method can favorably outperform existing methods in three synthetic datasets and real-world datasets. The code and dataset are released in https://github.com/weitingchen83/ICCV2021-Single-Image-Desnowing-HDCWNet.

This paper presents a systematic approach for solving wireless max-min utility fairness optimization problems in multiuser wireless networks with general monotonic constraints. These problems are often challenging to solve due to their nonconvexity. By establishing a connection between this class of optimization problems and the class of conditional eigenvalue problems that can be addressed by a generalized nonlinear Perron-Frobenius theory, we show how these problems can be solved optimally using an iterative algorithm that converges geometrically fast. The mathematical development in this paper unifies previous work and allows us to handle a broader class of competitive utility functions with general nonlinear monotonic constraints. Several representative applications illustrate the effectiveness of the proposed framework, including the max-min quality-of-service subject to robust interference temperature constraints in cognitive radio networks, the min-max weighted mean-square error subject to signal-to-interference-and-noise ratio constraints in multiuser downlink systems, the max-min throughput subject to nonlinear power constraints in energyefficient wireless networks, the max-min sigmoid utility in multimedia wireless networks, and the min-max outage probability subject to outage constraints in heterogeneous wireless networks. Numerical results are presented to demonstrate the fast-convergence behavior of the algorithms to the optimal fixed-point solution characterized by our generalized nonlinear Perron-Frobenius theoretic framework.

This paper presents a unifying and systematic framework to solve wireless max-min utility fairness optimization problems in multiuser wireless networks with generalized monotonic constraints. These problems are often challenging to solve due to their nonlinearity and non-convexity. Our framework leverages a general result in nonlinear Perron-Frobenius theory to characterize the global optimal solution of these problems analytically, and to design scalable and fast-convergent algorithms for the computation of the optimal solution. This work advances the state-of-the-art in handling wireless utility optimization problems with nonlinear monotonic constraints, which existing methodologies cannot handle, and also unifies previous works in this area. Several representative applications are considered to illustrate the effectiveness of the proposed framework, including max-min quality of service subject to robust interference temperature constraints in cognitive radio networks, min-max outage subject to outage constraints in heterogeneous networks, and min-max weighted MSE subject to SINR constraints in multiuser downlink system.