TO-UGDA: target-oriented unsupervised graph domain adaptation.

Graph domain adaptation (GDA) aims to address the challenge of limited label data in the target graph domain. Existing methods such as UDAGCN, GRADE, DEAL, and COCO for different-level (node-level, graph-level) adaptation tasks exhibit variations in domain feature extraction, and most of them solely rely on representation alignment to transfer label information from a labeled source domain to an unlabeled target domain. However, this approach can be influenced by irrelevant information and usually ignores the conditional shift of the downstream predictor. To effectively address this issue, we introduce a target-oriented unsupervised graph domain adaptive framework for graph adaptation called TO-UGDA. Particularly, domain-invariant feature representations are extracted using graph information bottleneck. The discrepancy between two domains is minimized using an adversarial alignment strategy to obtain a unified feature distribution. Additionally, the meta pseudo-label is introduced to enhance downstream adaptation and improve the model's generalizability. Through extensive experimentation on real-world graph datasets, it is proved that the proposed framework achieves excellent performance across various node-level and graph-level adaptation tasks.

Few-shot learning for joint model in underwater acoustic target recognition.

In underwater acoustic target recognition, there is a lack of massive high-quality labeled samples to train robust deep neural networks, and it is difficult to collect and annotate a large amount of base class data in advance unlike the image recognition field. Therefore, conventional few-shot learning methods are difficult to apply in underwater acoustic target recognition. In this report, following advanced self-supervised learning frameworks, a learning framework for underwater acoustic target recognition model with few samples is proposed. Meanwhile, a semi-supervised fine-tuning method is proposed to improve the fine-tuning performance by mining and labeling partial unlabeled samples based on the similarity of deep features. A set of small sample datasets with different amounts of labeled data are constructed, and the performance baselines of four underwater acoustic target recognition models are established based on these datasets. Compared with the baselines, using the proposed framework effectively improves the recognition effect of four models. Especially for the joint model, the recognition accuracy has increased by 2.04% to 12.14% compared with the baselines. The model performance on only 10 percent of the labeled data can exceed that on the full dataset, effectively reducing the dependence of model on the number of labeled samples. The problem of lack of labeled samples in underwater acoustic target recognition is alleviated.

Feature augmentation based on information fusion rectification for few-shot image classification.

In the issue of few-shot image classification, due to lack of sufficient data, directly training the model will lead to overfitting. In order to alleviate this problem, more and more methods focus on non-parametric data augmentation, which uses the information of known data to construct non-parametric normal distribution to expand samples in the support set. However, there are some differences between base class data and new ones, and the distribution of different samples belonging to same class is also different. The sample features generated by the current methods may have some deviations. A new few-shot image classification algorithm is proposed on the basis of information fusion rectification (IFR), which adequately uses the relationship between the data (including the relationship between base class data and new ones, and the relationship between support set and query set in the new class data), to rectify the distribution of support set in the new class data. In the proposed algorithm, feature of support set is expanded through sampling from the rectified normal distribution, so as to augment the data. Compared with other image augmentation algorithms, the experimental results on three few-shot datasets show that the accuracy of the proposed IFR algorithm is improved by 1.84-4.66% on 5-way 1-shot task and 0.99-1.43% on 5-way 5-shot task.

MaskID: An effective deep-learning-based algorithm for dense rebar counting.

As a dense instance segmentation problem, rebar counting in a complex environment such as rebar yard and rebar transpotation has received significant attention in both academic and industrial contexts. Traditional counting approaches, such as manual counting and machine vision-based algorithms, are often inefficient or inaccurate since rebars with varied sizes and shapes are stacked overlapping, rebar image is not clear for complex light condition such as dawn, night and strong light, and other environmental noises exist in rebar image; thus, they no longer fulfil the requirements of modern automation. This paper proposes MaskID, an innovative counting method based on deep learning and heuristic strategies. First, an improved version of the Mask region-based convolutional neural network (Mask R-CNN) was designed to obtain the segmentation results through splitting and rescaling so as to capture more detail in a large-scale rebar image. Then, a series of intelligent denoising strategies corresponding to aspect ratio of recognized box, overlapping recognized objects, object distribution and environmental noise, were applied to improve the segmentation results. The performance of the proposed method was evaluated on open-competition and test-platform datasets. The F1-score was found to be over 0.99 on all datasets. The experimental results demonstrate that the proposed method is effective for dense rebar counting and significantly outperforms existing state-of-the-art methods.

Spreading predictability in complex networks.

Many state-of-the-art researches focus on predicting infection scale or threshold in infectious diseases or rumor and give the vaccination strategies correspondingly. In these works, most of them assume that the infection probability and initially infected individuals are known at the very beginning. Generally, infectious diseases or rumor has been spreading for some time when it is noticed. How to predict which individuals will be infected in the future only by knowing the current snapshot becomes a key issue in infectious diseases or rumor control. In this report, a prediction model based on snapshot is presented to predict the potentially infected individuals in the future, not just the macro scale of infection. Experimental results on synthetic and real networks demonstrate that the infected individuals predicted by the model have good consistency with the actual infected ones based on simulations.

Deep convolution stack for waveform in underwater acoustic target recognition.

In underwater acoustic target recognition, deep learning methods have been proved to be effective on recognizing original signal waveform. Previous methods often utilize large convolutional kernels to extract features at the beginning of neural networks. It leads to a lack of depth and structural imbalance of networks. The power of nonlinear transformation brought by deep network has not been fully utilized. Deep convolution stack is a kind of network frame with flexible and balanced structure and it has not been explored well in underwater acoustic target recognition, even though such frame has been proven to be effective in other deep learning fields. In this paper, a multiscale residual unit (MSRU) is proposed to construct deep convolution stack network. Based on MSRU, a multiscale residual deep neural network (MSRDN) is presented to classify underwater acoustic target. Dataset acquired in a real-world scenario is used to verify the proposed unit and model. By adding MSRU into Generative Adversarial Networks, the validity of MSRU is proved. Finally, MSRDN achieves the best recognition accuracy of 83.15%, improved by 6.99% from the structure related networks which take the original signal waveform as input and 4.48% from the networks which take the time-frequency representation as input.

Evaluating the effect of Chinese control measures on COVID-19 via temporal reproduction number estimation.

Control measures are necessary to contain the spread of serious infectious diseases such as COVID-19, especially in its early stage. We propose to use temporal reproduction number an extension of effective reproduction number, to evaluate the efficacy of control measures, and establish a Monte-Carlo method to estimate the temporal reproduction number without complete information about symptom onsets. The province-level analysis indicates that the effective reproduction numbers of the majority of provinces in mainland China got down to < 1 just by one week from the setting of control measures, and the temporal reproduction number of the week [15 Feb, 21 Feb] is only about 0.18. It is therefore likely that Chinese control measures on COVID-19 are effective and efficient, though more research needs to be performed.

Influential Nodes Identification in Complex Networks via Information Entropy.

Identifying a set of influential nodes is an important topic in complex networks which plays a crucial role in many applications, such as market advertising, rumor controlling, and predicting valuable scientific publications. In regard to this, researchers have developed algorithms from simple degree methods to all kinds of sophisticated approaches. However, a more robust and practical algorithm is required for the task. In this paper, we propose the EnRenew algorithm aimed to identify a set of influential nodes via information entropy. Firstly, the information entropy of each node is calculated as initial spreading ability. Then, select the node with the largest information entropy and renovate its l-length reachable nodes' spreading ability by an attenuation factor, repeat this process until specific number of influential nodes are selected. Compared with the best state-of-the-art benchmark methods, the performance of proposed algorithm improved by 21.1%, 7.0%, 30.0%, 5.0%, 2.5%, and 9.0% in final affected scale on CEnew, Email, Hamster, Router, Condmat, and Amazon network, respectively, under the Susceptible-Infected-Recovered (SIR) simulation model. The proposed algorithm measures the importance of nodes based on information entropy and selects a group of important nodes through dynamic update strategy. The impressive results on the SIR simulation model shed light on new method of node mining in complex networks for information spreading and epidemic prevention.

Deep Anomaly Detection for CNC Machine Cutting Tool Using Spindle Current Signals.

In recent years, industrial production has become more and more automated. Machine cutting tool as an important part of industrial production have a large impact on the production efficiency and costs of products. In a real manufacturing process, tool breakage often occurs in an instant without warning, which results a extremely unbalanced ratio of the tool breakage samples to the normal ones. In this case, the traditional supervised learning model can not fit the sample of tool breakage well, which results to inaccurate prediction of tool breakage. In this paper, we use the high precision Hall sensor to collect spindle current data of computer numerical control (CNC). Combining the anomaly detection and deep learning methods, we propose a simple and novel method called CNN-AD to solve the class-imbalance problem in tool breakage prediction. Compared with other prediction algorithms, the proposed method can converge faster and has better accuracy.

Identifying critical nodes in temporal networks by network embedding.

Critical nodes in temporal networks play more significant role than other nodes on the structure and function of networks. The research on identifying critical nodes in temporal networks has attracted much attention since the real-world systems can be illustrated more accurately by temporal networks than static networks. Considering the topological information of networks, the algorithm MLI based on network embedding and machine learning are proposed in this paper. we convert the critical node identification problem in temporal networks into regression problem by the algorithm. The effectiveness of proposed methods is evaluated by SIR model and compared with well-known existing metrics such as temporal versions of betweenness, closeness, k-shell, degree deviation and dynamics-sensitive centralities in one synthetic and five real temporal networks. Experimental results show that the proposed method outperform these well-known methods in identifying critical nodes under spreading dynamic.

Identifying influential spreaders in complex networks by propagation probability dynamics.

Numerous well-known processes of complex systems such as spreading and cascading are mainly affected by a small number of critical nodes. Identifying influential nodes that lead to broad spreading in complex networks is of great theoretical and practical importance. Since the identification of vital nodes is closely related to propagation dynamics, a novel method DynamicRank that employs the probability model to measure the ranking scores of nodes is suggested. The influence of a node can be denoted by the sum of probability scores of its i order neighboring nodes. This simple yet effective method provides a new idea to understand the identification of vital nodes in propagation dynamics. Experimental studies on both Susceptible-Infected-Recovered and Susceptible-Infected-Susceptible models in real networks demonstrate that it outperforms existing methods such as Coreness, H-index, LocalRank, Betweenness, and Spreading Probability in terms of the Kendall τ coefficient. The linear time complexity enables it to be applied to real large-scale networks with tens of thousands of nodes and edges in a short time.

Identifying critical edges in complex networks.

The critical edges in complex networks are extraordinary edges which play more significant role than other edges on the structure and function of networks. The research on identifying critical edges in complex networks has attracted much attention because of its theoretical significance as well as wide range of applications. Considering the topological structure of networks and the ability to disseminate information, an edge ranking algorithm BCCMOD based on cliques and paths in networks is proposed in this report. The effectiveness of the proposed method is evaluated by SIR model, susceptibility index S and the size of giant component σ and compared with well-known existing metrics such as Jaccard coefficient, Bridgeness index, Betweenness centrality and Reachability index in nine real networks. Experimental results show that the proposed method outperforms these well-known methods in identifying critical edges both in network connectivity and spreading dynamic.

Correction: A Novel Top-k Strategy for Influence Maximization in Complex Networks with Community Structure.

[This corrects the article DOI: 10.1371/journal.pone.0145283.].

Identifying a set of influential spreaders in complex networks.

Identifying a set of influential spreaders in complex networks plays a crucial role in effective information spreading. A simple strategy is to choose top-r ranked nodes as spreaders according to influence ranking method such as PageRank, ClusterRank and k-shell decomposition. Besides, some heuristic methods such as hill-climbing, SPIN, degree discount and independent set based are also proposed. However, these approaches suffer from a possibility that some spreaders are so close together that they overlap sphere of influence or time consuming. In this report, we present a simply yet effectively iterative method named VoteRank to identify a set of decentralized spreaders with the best spreading ability. In this approach, all nodes vote in a spreader in each turn, and the voting ability of neighbors of elected spreader will be decreased in subsequent turn. Experimental results on four real networks show that under Susceptible-Infected-Recovered (SIR) and Susceptible-Infected (SI) models, VoteRank outperforms the traditional benchmark methods on both spreading rate and final affected scale. What's more, VoteRank has superior computational efficiency.

A Novel Top-k Strategy for Influence Maximization in Complex Networks with Community Structure.

In complex networks, it is of great theoretical and practical significance to identify a set of critical spreaders which help to control the spreading process. Some classic methods are proposed to identify multiple spreaders. However, they sometimes have limitations for the networks with community structure because many chosen spreaders may be clustered in a community. In this paper, we suggest a novel method to identify multiple spreaders from communities in a balanced way. The network is first divided into a great many super nodes and then k spreaders are selected from these super nodes. Experimental results on real and synthetic networks with community structure show that our method outperforms the classic methods for degree centrality, k-core and ClusterRank in most cases.

Predicting the evolution of spreading on complex networks.

Due to the wide applications, spreading processes on complex networks have been intensively studied. However, one of the most fundamental problems has not yet been well addressed: predicting the evolution of spreading based on a given snapshot of the propagation on networks. With this problem solved, one can accelerate or slow down the spreading in advance if the predicted propagation result is narrower or wider than expected. In this paper, we propose an iterative algorithm to estimate the infection probability of the spreading process and then apply it to a mean-field approach to predict the spreading coverage. The validation of the method is performed in both artificial and real networks. The results show that our method is accurate in both infection probability estimation and spreading coverage prediction.

Evolution characteristics of the network core in the Facebook.

Statistical properties of the static networks have been extensively studied. However, online social networks are evolving dynamically, understanding the evolving characteristics of the core is one of major concerns in online social networks. In this paper, we empirically investigate the evolving characteristics of the Facebook core. Firstly, we separate the Facebook-link(FL) and Facebook-wall(FW) datasets into 28 snapshots in terms of timestamps. By employing the k-core decomposition method to identify the core of each snapshot, we find that the core sizes of the FL and FW networks approximately contain about 672 and 373 nodes regardless of the exponential growth of the network sizes. Secondly, we analyze evolving topological properties of the core, including the k-core value, assortative coefficient, clustering coefficient and the average shortest path length. Empirical results show that nodes in the core are getting more interconnected in the evolving process. Thirdly, we investigate the life span of nodes belonging to the core. More than 50% nodes stay in the core for more than one year, and 19% nodes always stay in the core from the first snapshot. Finally, we analyze the connections between the core and the whole network, and find that nodes belonging to the core prefer to connect nodes with high k-core values, rather than the high degrees ones. This work could provide new insights into the online social network analysis.

Optimizing online social networks for information propagation.

Online users nowadays are facing serious information overload problem. In recent years, recommender systems have been widely studied to help people find relevant information. Adaptive social recommendation is one of these systems in which the connections in the online social networks are optimized for the information propagation so that users can receive interesting news or stories from their leaders. Validation of such adaptive social recommendation methods in the literature assumes uniform distribution of users' activity frequency. In this paper, our empirical analysis shows that the distribution of online users' activity is actually heterogenous. Accordingly, we propose a more realistic multi-agent model in which users' activity frequency are drawn from a power-law distribution. We find that previous social recommendation methods lead to serious delay of information propagation since many users are connected to inactive leaders. To solve this problem, we design a new similarity measure which takes into account users' activity frequencies. With this similarity measure, the average delay is significantly shortened and the recommendation accuracy is largely improved.

Ranking reputation and quality in online rating systems.

How to design an accurate and robust ranking algorithm is a fundamental problem with wide applications in many real systems. It is especially significant in online rating systems due to the existence of some spammers. In the literature, many well-performed iterative ranking methods have been proposed. These methods can effectively recognize the unreliable users and reduce their weight in judging the quality of objects, and finally lead to a more accurate evaluation of the online products. In this paper, we design an iterative ranking method with high performance in both accuracy and robustness. More specifically, a reputation redistribution process is introduced to enhance the influence of highly reputed users and two penalty factors enable the algorithm resistance to malicious behaviors. Validation of our method is performed in both artificial and real user-object bipartite networks.