Efficiency and security are the core challenges in the intelligent communication field. Lightweight neural networks have accelerated the information interpretation and communication efficiency, thereby fostering the rapid development of the Internet of Things. Enhancing the recognition capability of lightweight neural networks remains challenging. Knowledge distillation, a technique that transfers knowledge from a complex model to a smaller one, is often used to improve the recognition performance of lightweight networks. However, practical issues such as transmission constraints and user privacy make the original data required for knowledge distillation difficult to access directly. To tackle this issue, this paper proposes a Data-Free Cloud-Edge Knowledge Distillation (DF-CEKD) model, which uses a complex network in the cloud to provide training guidance for lightweight networks deployed on mobile devices. Specifically, DF-CEKD employs a novel Deep Inversion Diffusion Generation (DIDG) module to provide proxy data as input for the distillation process, thereby transferring the feature learning capability from the cloud network to the edge network. Meanwhile, a Multi-Layer Feature Joint Supervision Distillation (MLF-JSD) module is designed to further enhance the feature selection guidance provided by the teacher network in the cloud for training the lightweight student network. The simulation results demonstrate that the proposed DF-CEKD reduces the number of parameters to 1/20 and the floating-point operations to 1/12 when distilling from WRN40-2 to WRN16-1, resulting in only a 0.27% decrease in accuracy.

Constructing effective proxy data is one of the core challenges in data-free knowledge distillation. The existing models ignore the influence of the category entanglement of the generated data on the distillation. To alleviate this issue, imitating the human learning process, a new category-aware curriculum learning mechanism is proposed in this paper to perform data-free knowledge distillation, called CCL-D. The main ideology of this category-aware curriculum learning mechanism is to provide a new learning mode for data generation and network training, which enables the model to realize the knowledge distillation process from easy to difficult through automated curriculum learning. In this novel learning mechanism, a category-aware monitoring module is proposed to constrain the category attribute of generated data. Based on this monitoring module, the curriculum learning process for data generation and network training is designed and applied. Initially, the generator is guided to obtain new data with clear category features. The utilization of data with apparent category features is easy for student network training, and it enables the student network to learn clear and significant category features at the early training stage. Subsequently, the generator is guided to generate data with category entanglement. Utilizing these new data with category entanglement problems can improve the recognition ability of the student network to interclass interference and enhance network robustness. The effectiveness of the CCL-D is verified on the six benchmark experimental datasets (MNIST, CIFAR-10, CIFAR-100, SVHN, Caltech-101, Tiny-Imagenet).

Global average pooling (GAP) plays an important role in traditional channel attention. However, there is the disadvantage of insufficient information to use the result of GAP as the channel scalar. At the same time, the existing spatial attention models focus on the areas of interest using average pooling or convolutional networks, but there is a loss of feature information and neglect of the structural feature. In this paper, dual wavelet attention is proposed, which can effectively alleviate the aforementioned problems and enhance the representation ability of CNNs. Firstly, the equivalence between the sum of the low-frequency subband coefficients of 2D DWT (Haar) and GAP is proved. On this basis, the statistical characteristics of low-frequency and high-frequency subbands are effectively combined to obtain the channel scalars, which can better measure the importance of each channel. In addition, 2D DWT can effectively capture the approximate and detailed structural features. Thus, wavelet spatial attention is proposed, which can effectively focus on the key spatial structural features. Different from traditional spatial attention, it can better curve the structural and spatial attention for different channels. The experiments are verified on four natural image data sets and three remote sensing scene classification data sets, which shows the effectiveness and versatility of the proposed methods. The code of this paper will be available athttps://github.com/yutinyang/DWAN.

Model quantization can reduce the model size and computational latency, it has been successfully applied for many applications of mobile phones, embedded devices, and smart chips. Mixed-precision quantization models can match different bit precision according to the sensitivity of different layers to achieve great performance. However, it is difficult to quickly determine the quantization bit precision of each layer in deep neural networks under some constraints (for example, hardware resources, energy consumption, model size, and computational latency). In this article, a novel sequential single-path search (SSPS) method for mixed-precision model quantization is proposed, in which some given constraints are introduced to guide the searching process. A single-path search cell is proposed to combine a fully differentiable supernet, which can be optimized by gradient-based algorithms. Moreover, we sequentially determine the candidate precisions according to the selection certainties to exponentially reduce the search space and speed up the convergence of the searching process. Experiments show that our method can efficiently search the mixed-precision models for different architectures (for example, ResNet-20, 18, 34, 50, and MobileNet-V2) and datasets (for example, CIFAR-10, ImageNet, and COCO) under given constraints, and our experimental results verify that SSPS significantly outperforms their uniform-precision counterparts.

Data-free knowledge distillation further broadens the applications of the distillation model. Nevertheless, the problem of providing diverse data with rich expression patterns needs to be further explored. In this paper, a novel dynastic data-free knowledge distillation ($D^{3}K$) model is proposed to alleviate this problem. In this model, a dynastic supernet generator (D-SG) with a flexible network structure is proposed to generate diverse data. The D-SG can adaptively alter architectural configurations and activate different subnet generators in different sequential iteration spaces. The variable network structure increases the complexity and capacity of the generator, and strengthens its ability to generate diversified data. In addition, a novel additive constraint based on the differentiable dhash (D-Dhash) is designed to guide the structure parameter selection of the D-SG. This constraint forces the D-SG to constantly jump out of the fixed generation mode and generate diverse data in semantics and instance. The effectiveness of the proposed model is verified on the experimental benchmark datasets (MNIST, CIFAR-10, CIFAR-100, and SVHN).

Hunks generated by textual-differencing tools are often used for understanding code changes.However, in the side-by-side view, the match between the deleted and added lines of a hunk is sometimes inconsistent with actual changes to the corresponding syntax entities.This mismatch usually occurs in syntax coupled hunks, i.e. hunks that contain changes to multiple syntax entities.It makes the hunks incomprehensible and misleading.A hybrid differencing algorithm is proposed to alleviate this problem.It applies tree-differencing to syntax coupled hunks to generate edits.It then maps edits back to the source code to generate adjusted hunks.Based on the current implementation, we conducted a case study on 10 open source projects.The results showed that 15% of commits contain syntax coupled hunks.And, we evaluated the results of the algorithm on 1,500 randomly drawn samples, and the correct matching rate was as high as 97%, demonstrating the effectiveness of the algorithm in reducing mismatches.

The training of deep neural networks (DNNs) requires intensive resources both for computation and for storage performance. Thus, DNNs cannot be efficiently applied to mobile phones and embedded devices, which seriously limits their applicability in industry applications. To address this issue, we propose a novel encoding scheme of using {−1, +1} to decompose quantized neural networks (QNNs) into multibranch binary networks, which can be efficiently implemented by bitwise operations (xnor and bitcount) to achieve model compression, computational acceleration and resource saving. Based on our method, users can easily achieve different encoding precisions arbitrarily according to their requirements and hardware resources. The proposed mechanism is very suitable for the use of FPGA and ASIC in terms of data storage and computation, which provides a feasible idea for smart chips. We validate the effectiveness of our method on both large-scale image classification tasks (e.g., ImageNet) and object detection tasks. In particular, our method with lowbit encoding can still achieve almost the same performance as its full-precision counterparts.

Exploiting rich spatial and spectral features contributes to improve the classification accuracy of hyperspectral images (HSIs). In this paper, based on the mechanism of the population receptive field (pRF) in human visual cortex, we further utilize the spatial correlation of pixels in images and propose pixel directed acyclic graph recurrent neural network (Pixel DAG-RNN) to extract and apply spectral-spatial features for HSIs classification. In our model, an undirected cyclic graph (UCG) is used to represent the relevance connectivity of pixels in an image patch, and four DAGs are used to approximate the spatial relationship of UCGs. In order to avoid overfitting, weight sharing and dropout are adopted. The higher classification performance of our model on HSIs classification has been verified by experiments on three benchmark data sets.

Polarimetric synthetic aperture radar (PolSAR) images are widely used in disaster detection and military reconnaissance and so on. However, their interpretation faces some challenges, e.g., deficiency of labeled data, inadequate utilization of data information and so on. In this paper, a complex-valued generative adversarial network (GAN) is proposed for the first time to address these issues. The complex number form of model complies with the physical mechanism of PolSAR data and in favor of utilizing and retaining amplitude and phase information of PolSAR data. GAN architecture and semi-supervised learning are combined to handle deficiency of la-beled data. GAN expands training data and semi-supervised learning is used to train network with generated, labeled and unlabeled data. Experimental results on two benchmark data sets show that our model outperforms existing state-of-the-art models, especially for conditions with fewer labeled data.

In this paper, several useful inequalities for uncertain variables are proved. A Borel-Cantelli lemma for uncertain measures is obtained and some convergence theorems for continuous uncertain measures are derived. Finally, these theorems are applied to compute the uncertainty distribution of Liu integral. We prove that the uncertain integral of a deterministic function with respect to a Liu process has a normal uncertainty distribution.