To classify the sentiment polarity of the aspect entity in a sentence, most existing research evaluates the semantic knowledge among a certain aspect of a sentence and corresponding context as significant clues for the task. However, available accompanying information has not been completely exploited, especially the coarse-grained category-level knowledge in contexts. Such knowledge can help to alleviate polysemy and ambivalence problems. In this paper, we propose a multi-task learning framework Co-interactive Attention Network(CoAN) to jointly learn and handle multiple granularity features at both target and category levels. In order to leverage the fine-grained and coarse-grained knowledge in contexts and get multi-granularity sentiment related sentence representations, we introduce two co-interactive attention layers to conduct accompanying semantic interactions at the word-level and the feature-level. The experimental results on three restaurant review datasets prove that CoAN is superior to the baselines by 1.41% in accuracy and 2.81% in F1-score. Furthermore, ablation studies and attention visualizations show that the multi-task framework and novel co-interactive mechanisms can distinguish and fuse multi-granularity knowledge, which benefits the two subtasks in aspect based sentiment analysis.

Single-image super-resolution (SISR) aims to reconstruct a high-resolution image from a single low-resolution (LR) image. Although convolutional neural network (CNN)-based SISR methods greatly enhance image restoration, they face critical challenges. First, SISR models using CNNs process image patches uniformly regardless of importance, causing spatial inefficiency in computation and representation. However, due to resource constraints, edge devices in the Internet of Things (IoT) cannot bear heavy computations or large memory storage. Second, most of the existing SISR methods are designed only for the widely adopted bicubic degradation and cannot handle LR images with arbitrary blur kernels, resulting in poor recovery performance. To address these issues, in this article, we propose a two-stage semantic and spatial deep super-resolution (SSDSR) model suitable for the IoT environment. The proposed SSDSR model is capable of handling a variety of blur kernels (e.g., isotropic Gaussian, motion, and disk blur) by effectively using their prior information. Moreover, the semantic feature extraction (SFE) module enables the proposed model to focus on key areas of LR images rather than treating all pixels equally, which significantly reduces the computational load. The semantic information from the SFE module and the spatial information from the spatial attention module are fused adaptively, allowing the proposed model to extract key information in LR images, thereby increasing the representation capacity of the CNN and improving image recovery. Compared with state-of-the-art SISR methods on benchmark datasets, the proposed SSDSR model demonstrates superior performance. When run in real time on an IoT edge device, our model exhibits high computational efficiency and excellent image quality.

Text matching is a critical task in natural language processing to measure semantic similarity between two texts. A significant portion of online texts are labeled with a variety of coarse topic responses. These supervised topic indicators can provide prior structured and explicable semantics for textual similarity modeling. However, most existing state-of-the-art neural network methods cannot benefit from such complementary topic signals. Therefore, we propose a novel Topic Supervision BERT-based model (TSB) for text matching. TSB provides a reference multi-task joint training framework involving two types of topic supervision, including explicit and implicit topic supervision. To constrain consistent topic correspondences between texts, we introduce a supervised auxiliary learning task to incorporate explicit pre-defined topic supervision. Furthermore, to adapt to latent topic structures for mutual benefit between text representations and multiple tasks, we integrate a topic model into a contextual text representation model BERT to mine and incorporate implicit self-learnable topic supervision. Experimental results show that TSB supplements explicit and implicit topic information through a multi-task learning approach, which significantly improves the performance of text matching on two public datasets, especially on challenging short text matches.

Graph anomaly detection, here, aims to find rare patterns that are significantly different from other nodes. Attributed graphs containing complex structure and attribute information are ubiquitous in our life scenarios such as bank account transaction graph and paper citation graph. Anomalous nodes on attributed graphs show great difference from others in the perspectives of structure and attributes, and give rise to various types of graph anomalies. In this paper, we investigate three types of graph anomalies: local, global, and structure anomalies. And, graph neural networks (GNNs) based anomaly detection methods attract considerable research interests due to the power of modeling attributed graphs. However, the convolution operation of GNNs aggregates neighbors information to represent nodes, which makes node representations more similar and cannot effectively distinguish between normal and anomalous nodes, thus result in sub-optimal results. To improve the performance of anomaly detection, we propose a novel community-aware attributed graph anomaly detection framework (ComGA). We design a tailored deep graph convolutional network (tGCN) to anomaly detection on attributed graphs. Extensive experiments on eight real-life graph datasets demonstrate the effectiveness of ComGA.

A critical issue of drug risk-benefit evaluation is to determine the frequencies of drug side effects. Randomized controlled trail is the conventional method for obtaining the frequencies of side effects, while it is laborious and slow. Therefore, it is necessary to guide the trail by computational methods. Existing methods for predicting the frequencies of drug side effects focus on modeling drug-side effect interaction graph. The inherent disadvantage of these approaches is that their performance is closely linked to the density of interactions but which is highly sparse. More importantly, for a cold start drug that does not appear in the training data, such methods cannot learn the preference embedding of the drug because there is no link to the drug in the interaction graph. In this work, we propose a new method for predicting the frequencies of drug side effects, DSGAT, by using the drug molecular graph instead of the commonly used interaction graph. This leads to the ability to learn embeddings for cold start drugs with graph attention networks. The proposed novel loss function, i.e. weighted $\varepsilon$-insensitive loss function, could alleviate the sparsity problem. Experimental results on one benchmark dataset demonstrate that DSGAT yields significant improvement for cold start drugs and outperforms the state-of-the-art performance in the warm start scenario. Source code and datasets are available at https://github.com/xxy45/DSGAT.

Network representation aims to learn low-dimensional vector representations of network nodes while preserving the inherent properties of the network. For all its popularity, majority of the existing methods focus on exploitation of diverse information, including network topology and semantic information on nodes of network, and ignore their implicit semantics. For example, we all know the saying that birds of a feather flock together. More concretely, semantic information of one node can be influenced by its neighbors' semantic information. Furthermore, even two nodes are not directly connected, they may have similar implicit semantic information (i.e., high-order semantic proximity). Thus, they should be close in the represented vector space. To this end, we propose a Deep Semantic Network Representation approach (DSNR) in the self-translation framework from sequence to sequence. To excavate the implicit semantic information of nodes and capture the high-order semantic proximity, three key components make our approach effective, i.e., aggregation of nodes neighbors' semantic information and enhancement to the semantic feature representations of nodes by a deep autoencoder, integration of nodes semantic information in node identity sequence to generate node semantic sequence, and translation from node semantic sequence to node identity sequence to capture the high-order semantic proximity in an attention-enhanced seq2seq framework. Extensive experiments based on three real-world datasets have verified the effectiveness of our proposed approach11Code is available at https://github.com/DASE4/DSNR.