Transformer-based methods have demonstrated impressive results on single-image super-resolution (SISR) task. However, self-attention mechanism is computationally expensive when applied to the entire image. As a result, current approaches divide low-resolution input images into small patches, which are processed separately and then fused to generate high-resolution images. Nevertheless, this conventional regular patch division is too coarse and lacks interpretability, resulting in artifacts and non-similar structure interference during attention operations. To address these challenges, we propose a novel super token interaction network (SPIN). Our method employs superpixels to cluster local similar pixels to form the explicable local regions and utilizes intra-superpixel attention to enable local information interaction. It is interpretable because only similar regions complement each other and dissimilar regions are excluded. Moreover, we design a super-pixel cross-attention module to facilitate information propagation via the surrogation of superpixels. Extensive experiments demonstrate that the proposed SPIN model performs favorably against the state-of-the-art SR methods in terms of accuracy and lightweight. Code is available at https://github.com/ArcticHare105/SPIN.

Deep learning techniques have achieved promising results in code clone detection in the past decade. However, existing techniques merely focus on how to extract more dis-criminative features from source codes, while some issues, such as structural differences of functional similar codes, are not explicitly addressed. This phenomenon is common when programmers copy a code segment along with adding or removing several statements, or use a more flexible syntax structure to implement the same function. In this paper, we unify the aforementioned problems as the problem of code misalignment, and propose a novel code alignment network to tackle it. We design a bi-directional causal convolutional neural network to extract feature representations of code fragments with rich structural and semantical information. After feature extraction, our method learns to align the two code fragments in a data-driven fashion. We present two independent strategies for code alignment, namely attention-based alignment and sparse reconstruction-based alignment. Both two strategies strive to learn an alignment matrix that represents the correspondences between two code fragments. Our method outperforms state-of-the-art methods in terms of F1 score by 0.5% and 3.1 % on BigCloneBench and OJClone, respectively11Our code is available at https://github.com/ArcticHare105/Code-Alignment.

In the intelligent manufacturing process of solar photovoltaic (PV) cells, the automatic defect detection system using the Industrial Internet of Things (IIoT) smart cameras and sensors cooperated in IIoT has become a promising solution. Many works have been devoted to defect detection of PV cells in a data-driven way. However, because of the subjectivity and fuzziness of human annotation, the data contains a high quantity of noise and unpredictable uncertainties, which creates great difficulties in automatic defect detection. To address this problem, we propose a novel architecture named fuzzy convolution, which integrates fuzzy logic and convolution operations at microscopic level. Combining the proposed fuzzy convolution with the regular convolution, we build a network called Hybrid Fuzzy Convolutional Neural Network (HFCNN). Compared with convolutional neural networks (CNNs), HFCNN can address the uncertainties of PV cell data to improve the accuracy with fewer parameters, making it possible to apply our method in smart cameras. Experimental results on a public dataset show the superiority of our proposed method compared with CNNs.

In this paper, cellular neural networks with oscillating leakage coefficients and complex deviating arguments are considered. Some criteria are established for the existence of pseudo almost periodic solutions for these models by using the contraction mapping fixed point theorem and inequality analysis technique. The results of this paper are new and complement the previously known ones.

In this paper, neutral type shunting inhibitory cellular neural networks with D operator are considered. Based on Lyapunov functional method and differential inequality technique, some new criteria are derived to guarantee the existence and global exponential stability of pseudo almost periodic solutions of considered systems. In addition, an example and its numerical simulations are provided to show the validity and the advantages of the obtained results.