Vision Transformer (ViT)-based image super-resolution (SR) methods have achieved impressive performance and surpassed CNN-based SR methods by utilizing MultiHead Self-Attention (MHSA) to model long-range dependencies. However, the quadratic complexity of MHSA and the inefficiency of non-parallelized window partition seriously affect the inference speed, hindering these SR methods from being applied to application scenarios requiring speed and quality. To address this issue, we propose an Asymmetric Large-kernel Attention Network (ALAN) utilizing a stage-to-block design paradigm inspired by ViT. In the ALAN, the core block named Asymmetric Large Kernel Convolution Block (ALKCB) adopts a similar structure to the Swin Transformer Layer but replaces the MHSA with our proposed Asymmetric Depth-Wise Convolution Attention (ADWCA) to enhance both the SR quality and inference speed. The proposed ADWCA, with linear complexity, uses large kernel depth-wise dilation convolution and Hadamard product as the attention map. The structural re-parameterization technique to strengthen the kernel skeletons with asymmetric convolution is also explored. Experimental results demonstrate that ALAN achieves state-of-the-art performance with faster inference speed than ViTbased models and smaller parameters than CNN-based models. Specifically, the tiny size of ALAN (ALAN-T) is 3× smaller than ShuffleMixer with similar performance, and ALAN is 4× faster than SwinIR-S with 0.1dB gain in PSNR.
26 October 2023 in IEEE Transactions on Vehicular Technology
To cope with the increasing demands of computing tasks in vehicles, task offloading to computing nodes (vehicles and edge servers) in a vehicular edge computing (VEC) scenario has gained significant attention. However, the high dynamic of task arrival and variant wireless network conditions could severely impact the quality of computation offloading for latency-aware and high-reliability services. Consequently, only offloading tasks to a single computing node may fail to satisfy the reliability requirements due to potential transmission errors or constraints on delay. To mitigate the limitations of task offloading to only one single computing node in VEC, task replication, in which a task can be sent to multiple computing nodes simultaneously is utilized. Nevertheless, unrestricted task replication may exacerbate the depletion of network and computing resources in VEC when a burst of computing tasks is generated. To address the above challenges, we first formulate the problem of power allocation with task replication in NOMA-enabled VEC. Our purpose is to maximize task transmission reliability based on the Finite Block Length (FBL) theory while following the delay constraint. To achieve this goal, we formulate the problem of task replication and power allocation into a constrained stochastic optimization problem and apply the Lyapunov optimization framework to adjust task replication decisions and power allocation for each vehicle. Simulation results demonstrate that the proposed algorithm can obtain better service reliability than other alternative algorithms.
Deep learning (DL) frameworks serve as the backbone for a wide range of artificial intelligence applications. However, bugs within DL frameworks can cascade into critical issues in higher-level applications, jeopardizing reliability and security. While numerous techniques have been proposed to detect bugs in DL frameworks, research exploring common API patterns across frameworks and the potential risks they entail remains limited. Notably, many DL frameworks expose similar APIs with overlapping input parameters and functionalities, rendering them vulnerable to shared bugs, where a flaw in one API may extend to analogous APIs in other frameworks. To address this challenge, we propose MirrorFuzz, an automated API fuzzing solution to discover shared bugs in DL frameworks. MirrorFuzz operates in three stages: First, MirrorFuzz collects historical bug data for each API within a DL framework to identify potentially buggy APIs. Second, it matches each buggy API in a specific framework with similar APIs within and across other DL frameworks. Third, it employs large language models (LLMs) to synthesize code for the API under test, leveraging the historical bug data of similar APIs to trigger analogous bugs across APIs. We implement MirrorFuzz and evaluate it on four popular DL frameworks (TensorFlow, PyTorch, OneFlow, and Jittor). Extensive evaluation demonstrates that MirrorFuzz improves code coverage by 39.92% and 98.20% compared to state-of-the-art methods on TensorFlow and PyTorch, respectively. Moreover, MirrorFuzz discovers 315 bugs, 262 of which are newly found, and 80 bugs are fixed, with 52 of these bugs assigned CNVD IDs.
Traditional Convolutional Neural Network (ConvNet, CNN)-based image super-resolution (SR) methods have lower computation costs, making them more friendly for real-world scenarios. However, they suffer from lower performance. On the contrary, Vision Transformer (ViT)-based SR methods have achieved impressive performance recently, but these methods often suffer from high computation costs and model storage overhead, making them hard to meet the requirements in practical application scenarios. In practical scenarios, an SR model should reconstruct an image with high quality and fast inference. To handle this issue, we propose a novel CNN-based Efficient Residual ConvNet enhanced with structural Re-parameterization (RepECN) for a better trade-off between performance and efficiency. A stage-to-block hierarchical architecture design paradigm inspired by ViT is utilized to keep the state-of-the-art performance, while the efficiency is ensured by abandoning the time-consuming Multi-Head Self-Attention (MHSA) and by re-designing the block-level modules based on CNN. Specifically, RepECN consists of three structural modules: a shallow feature extraction module, a deep feature extraction, and an image reconstruction module. The deep feature extraction module comprises multiple ConvNet Stages (CNS), each containing 6 Re-Parameterization ConvNet Blocks (RepCNB), a head layer, and a residual connection. The RepCNB utilizes larger kernel convolutions rather than MHSA to enhance the capability of learning long-range dependence. In the image reconstruction module, an upsampling module consisting of nearest-neighbor interpolation and pixel attention is deployed to reduce parameters and maintain reconstruction performance, while bicubic interpolation on another branch allows the backbone network to focus on learning high-frequency information. The extensive experimental results on multiple public benchmarks show that our RepECN can achieve 2.5∼5× faster inference than the state-of-the-art ViT-based SR model with better or competitive super-resolving performance, indicating that our RepECN can reconstruct high-quality images with fast inference.