Search

Abstract Generative AI tools exemplified by ChatGPT are becoming a new reality. This study is motivated by the premise that “AI generated content may exhibit a distinctive behavior that can be separated from scientific articles”. In this study, we show how articles can be generated using means of prompt engineering for various diseases and conditions. We then show how we tested this premise in two phases and prove its validity. Subsequently, we introduce xFakeSci, a novel learning algorithm, that is capable of distinguishing ChatGPT-generated articles from publications produced by scientists. The algorithm is trained using network models driven from both sources. To mitigate overfitting issues, we incorporated a calibration step that is built upon data-driven heuristics, including proximity and ratios. Specifically, from a total of a 3952 fake articles for three different medical conditions, the algorithm was trained using only 100 articles, but calibrated using folds of 100 articles. As for the classification step, it was performed using 300 articles per condition. The actual label steps took place against an equal mix of 50 generated articles and 50 authentic PubMed abstracts. The testing also spanned publication periods from 2010 to 2024 and encompassed research on three distinct diseases: cancer, depression, and Alzheimer’s. Further, we evaluated the accuracy of the xFakeSci algorithm against some of the classical data mining algorithms (e.g., Support Vector Machines, Regression, and Naive Bayes). The xFakeSci algorithm achieved F1 scores ranging from 80 to 94%, outperforming common data mining algorithms, which scored F1 values between 38 and 52%. We attribute the noticeable difference to the introduction of calibration and a proximity distance heuristic, which underscores this promising performance. Indeed, the prediction of fake science generated by ChatGPT presents a considerable challenge. Nonetheless, the introduction of the xFakeSci algorithm is a significant step on the way to combating fake science.

Summary: As the influence of transformer-based approaches in general and generative artificial intelligence (AI) in particular continues to expand across various domains, concerns regarding authenticity and explainability are on the rise. Here, we share our perspective on the necessity of implementing effective detection, verification, and explainability mechanisms to counteract the potential harms arising from the proliferation of AI-generated inauthentic content and science. We recognize the transformative potential of generative AI, exemplified by ChatGPT, in the scientific landscape. However, we also emphasize the urgency of addressing associated challenges, particularly in light of the risks posed by disinformation, misinformation, and unreproducible science. This perspective serves as a response to the call for concerted efforts to safeguard the authenticity of information in the age of AI. By prioritizing detection, fact-checking, and explainability policies, we aim to foster a climate of trust, uphold ethical standards, and harness the full potential of AI for the betterment of science and society.

Entity alignment refers to discovering identical entity pairs in 2 knowledge graphs, which is a significant task in knowledge fusion. Early automated entity alignment techniques are based mainly on similarity calculation and comparing symbolic features, i.e., entity names, between entities. Nevertheless, such methods’ performance would reduce significantly when the difference between knowledge graphs is enormous because of relying on predefined comparison rules. Recently, embedding-based methods calculate the similarity between entity pairs through vector embeddings and thus can deal with different knowledge graphs. However, embedding-based methods mostly require humans to annotate data, which is laborious. Therefore, we learn from each other to propose an unsupervised entity alignment framework in this work, which can generate initial alignment seeds automatically by considering symbolic similarities. It can effectively avoid the waste of human resources and is suitable for handling multiple types of knowledge graphs. In addition, we investigate improving the quality and quantity of initial alignment by integrating multiple symbolic similarity features of entities and dealing with the situation of entity information missing better. Experimental results on 3 real datasets demonstrate its state-of-the-art performance.

Unsupervised domain adaptation involves knowledge transfer from a labeled source to unlabeled target domains to assist target learning tasks. A critical aspect of unsupervised domain adaptation is the learning of more transferable and distinct feature representations from different domains. Although previous investigations, using, for example, CNN-based and auto-encoder-based methods, have produced remarkable results in domain adaptation, there are still two main problems that occur with these methods. The first is a training problem for deep neural networks; some optimization methods are ineffective when applied to unsupervised deep networks for domain adaptation tasks. The second problem that arises is that redundancy of image data results in performance degradation in feature learning for domain adaptation. To address these problems, in this paper, we propose an unsupervised domain adaptation method with a stacked convolutional sparse autoencoder, which is based on performing layer projection from the original data to obtain higher-level representations for unsupervised domain adaptation. More specifically, in a convolutional neural network, lower layers generate more discriminative features whose kernels are learned via a sparse autoencoder. A reconstruction independent component analysis optimization algorithm was introduced to perform individual component analysis on the input data. Experiments undertaken demonstrated superior classification performance of up to 89.3% in terms of accuracy compared to several state-of-the-art domain adaptation methods, such as SSRLDA and TLMRA.

With the rapid increase of internet information, personalized recommendation systems are an effective way to alleviate the information overload problem, which has attracted extensive attention in recent years. The traditional collaborative filtering utilizes matrix factorization methods to learn hidden feature representations of users and/or items. With deep learning achieved good performance in representation learning, the autoencoder model is widely applied in recommendation systems for the advantages of fast convergence and no label requirement. However, the previous recommendation systems may take the reconstruction output of an autoencoder as the prediction of missing values directly, which may deteriorate their performance and cause unsatisfactory results of recommendation. In addition, the parameters of an autoencoder need to be pre-trained ahead, which greatly increases the time complexity. To address these problems, in this paper, we propose a Hybrid Collaborative Recommendation method via Dual-Autoencoder (HCRDa). More specifically, firstly, a novel dual-autoencoder is utilized to simultaneously learn the feature representations of users and items in our HCRDa, which obviously reduces time complexity. Secondly, embedding matrix factorization into the training process of the autoencoder further improves the quality of hidden features for users and items. Finally, additional attributes of users and items are utilized to alleviate the cold start problem and to make hybrid recommendations. Comprehensive experiments on several real-world data sets demonstrate the effectiveness of our proposed method in comparison with several state-of-the-art methods.

Bionic polarization navigation has a broad variety of application in diverse fields for high reliability and strong robustness to interference, fundamental to which is the use of a polarization compass based on polarized light cues. Nevertheless, dramatical reduction of the orientation accuracy resulted from the noise in a measured angle of polarization (AoP) and the tilted angles of a polarization compass during operation gives imperative influence on navigation precision. Herein, we investigate how to improve the navigation accuracy effectively by the proposed comprehensive heading error processing technique for a polarization compass, where a novel denoising scheme is designed to eliminate the noise in AoP images directly by integrating the strength of iterative variance-stabilizing transformation (IVST) and adaptive soft interval thresholding (SIT) so as to compensate the following tilt-induced error accurately. Subsequently, a promising compensation approach inspired by efficient extreme learning machine (EELM) is introduced to correct the tilt-induced error caused by realistic execution. The AoP image denoising advance and the tilt-induced error modeling advance combine to produce remarkable performance gains on the heading error. Experimental results and comparisons with prior arts reveal that the proposed comprehensive heading error processing technique is highly appealing in terms of improving the orientation accuracy for a polarization compass with superiority to state-of-the-art alternatives.

As the backbone of many real-world complex systems, networks interact with others in nontrivial ways from time to time. It is a challenging problem to detect subgraphs that have dependencies on each other across multiple networks. Instead of devising a method for a specific scenario, we propose a generic framework to discover subgraphs in multiple interdependent networks, which generalizes the classical subgraph detection problem in a single network and can be applied to more practical applications. Specifically, we propose the Graph Block-structured Gradient Hard Thresholding Pursuit (GB-GHTP) framework to optimize interdependent networks with block-structured constraints, which enjoys 1) a theoretical guarantee and 2) a nearly linear time complexity on the network size. It is demonstrated how our framework can be applied to three practical applications: 1) evolving anomalous subgraph detection in dynamic networks, 2) anomalous subgraph detection in networks of networks, and 3) connected dense subgraph detection in dual networks. We evaluate our framework on large-scale datasets with comprehensive experiments, which validate our framework's effectiveness and efficiency.

In recent years, online feature selection has been a research topic on streaming feature mining, as it can reduce the dimensionality of the streaming features by removing the irrelevant and redundant features in real time. There are many representative research efforts on the online feature selection with streaming features, i.e., alpha - investing, online streaming feature selection (OSFS), and scalable and accurate online approach (SAOLA) for feature selection. In these studies, alpha-investing has limited prediction accuracy and a large number of selected features. SAOLA sometimes offers outstanding efficiency in running time and prediction accuracy but possesses a large number of selected features. OSFS offers high prediction accuracy in many datasets, but its running time increases exponentially with an increasing number of features with low redundancy and high relevance. To address the limitations of the above-mentioned works, we propose an online learning algorithm named OSFAS, which samples streaming features in real-time by a self-adaption sliding-window and discards the irrelevant and redundant features by conditional independence. The OSFAS obtains an approximate Markov blanket with high prediction accuracy, meanwhile reducing the number of selected features. The efficiency of the proposed OSFASW algorithm was validated in a performance test on widely used datasets, e.g., NIPS2003 and causality workbench. Through the extensive experimental results, we demonstrate that OSFAS significantly improves the prediction accuracy and requires a smaller number of selected features than alpha - investing, OSFS, and SAOLA.

Keyphrase extraction is the task of automatically extracting descriptive phrases or concepts that represent the main topics in a document. Finding good keyphrases in a document can quickly summarize knowledge for information retrieval and decision making. Existing keyphrase extraction methods cannot be customized to each specific document, and cannot capture flexible semantic relations. In this paper, a keyphrase extraction algorithm using maximum sequential pattern mining with one-off and general gaps condition, called Ke-MSMING, is presented. Ke_MSMING first searches all keyphrase candidates from a document using sequential patterns mining and the topic model, and then adopts supervised machine learning to classify each keyphrase candidate as a keyphrase or not. Finally, Ke_MSMING selects top-N keyphrases as the final keyphrases. Ke_MSMING not only uses baseline features and pattern features but also uses centrality features obtained from the cooccurrence semantic network, and the cooccurrence networks can yield powerful semantic relations for keyphrase extraction. Experimental results on two datasets demonstrate that Ke_MSMING has better performance than other state-of-the-art keyphrase extraction approaches.

In pattern matching, a gap constraint is a more flexible wildcard than traditional wildcards “?”and “*”. Pattern matching with gap constraints is more difficult to handle and fulfills user's enquiries more easily. Pattern matching with gap constraints has therefore been carried out in numerous research works, such as music information retrieval, searching protein sites, and sequence pattern mining. Strict pattern matching under a nonoverlapping condition, as a type of pattern matching with gap constraints, is a key issue of sequence pattern mining with gap constraints since it can be used to compute the frequency of a pattern. Exact matching limits the flexibility of the match to some extent since it requires each character to be matched exactly. We therefore address approximate strict pattern matching under the nonoverlapping constraints (ASPNO) and propose an effective algorithm, named NETtree for ASPNO (NETASPNO), which first transforms the problem into a Nettree data structure, an extensive tree structure. To find the nonoverlapping occurrences effectively, we propose the concept of number of roots paths with distance constraints (NRPDC) which indicates the number of path from a node to the roots with distanced and can be used to delete useless parent-child relationships and useless nodes. We iteratively recalculate the NRPDCs of each node on the subnettree with the rightmost root. Then we can get a path from the rightmost leaf to its rightmost root without using the backtracking strategy. NETASPNO therefore iteratively gets the rightmost root-leaf-path and prunes the path on the Nettree. Extensive experimental results demonstrate that NETASPNO has better performance than the other competitive algorithms.

While most of the existing topic models perform a full analysis on a set of documents to discover all topics, it is noticed recently that in many situations users are interested in fine-grained topics related to some specific aspects only. As a result, targeted analysis (or focused analysis) has been proposed to address this problem. Given a corpus of documents from a broad area, targeted analysis discovers only topics related with user-interested aspects that are expressed by a set of user-provided query keywords. Existing approaches for targeted analysis suffer from problems such as topic loss and topic suppression because of their inherent assumptions and strategies. Moreover, existing approaches are not designed to address computation efficiency, while targeted analysis is supposed to provide responses to user queries as soon as possible. In this paper, we propose a core BiTerms-based Topic Model (BiTTM). By modelling topics from core biterms that are potentially relevant to the target query, on one hand, BiTTM captures the context information across documents to alleviate the problem of topic loss or suppression; on the other hand, our proposed model enables the efficient modelling of topics related to specific aspects. Our experiments on nine real-world datasets demonstrate BiTTM outperforms existing approaches in terms of both effectiveness and efficiency.

The spread of the Coronavirus pandemic has been accompanied by an infodemic. The false information that is embedded in the infodemic affects people’s ability to have access to safety information and follow proper procedures to mitigate the risks. This research aims to target the falsehood part of the infodemic, which prominently proliferates in news articles and false medical publications. Here, we present NeoNet, a novel supervised machine learning algorithm that analyzes the content of a document (news article, a medical publication) and assigns a label to it. The algorithm was trained by Term Frequency Inverse Document Frequency (TF-IDF) bigram features, which contribute a network training model. The algorithm was tested on two different real-world datasets from the CBC news network and COVID-19 publications. In five different fold comparisons, the algorithm predicted a label of an article with a precision of 97–99%. When compared with prominent algorithms such as Neural Networks, SVM, and Random Forests NeoNet surpassed them. The analysis highlighted the promise of NeoNet in detecting disputed online contents, which may contribute negatively to the COVID-19 pandemic.

Starting in July 2016, the Ministry of Science and Technology of China, along with several other national agencies, sponsors a 54-month 45-million RMB (Chinese Yuan) project on knowledge engineering with Big Data (www.bigke.org) for 15 top research and development institutions to study the fundamental theory and the applications of BigKE, a big-data knowledge engineering framework that handles fragmented knowledge modeling and online learning from multiple information sources, nonlinear fusion on fragmented knowledge, and automated demand-driven knowledge navigation. The project seeks to provide petabytescale data and knowledge services in identified application domains. In this paper, we discuss our BigKE framework, and present a novel application scenario for BigKE services.

Since sparse representation (SR) was first introduced into robust face recognition, the argument has lasted for several years about whether sparsity can improve robust face recognition or not. Some work argued that the robust sparse representation (RSR) model has a similar recognition rate as non‐sparse solution, while it needs a much higher computational cost due to the larger feature dimensionality in the pixel space. In this study, the authors reveal that the standard RSR model, which expands the dictionary with the identity matrix to reconstruct corruption or occlusion in face images, is essentially a non‐sparse solution with a relatively large residual. The reason why the RSR model underperforms may be its inappropriately expanded bases rather than the sparsity itself. Thereby, this study proposes to design a dictionary with an expanded noise bases set which can precisely reconstructs any corruption or occlusion in face images in a subspace. Experimental results show that the algorithm can greatly improve recognition rates for robust face recognition. In addition, the algorithm can be simply performed in a subspace with a small feature dimensionality, thus efficient enough for real systems. This study makes us come to the conclusion that solving the approximation problem in raw pixel space is not necessary for robust face recognition, while solving in a subspace with a much smaller feature dimensionality is enough when the dictionary is well expanded. Finally, this study also confirms that the sparsity plays an important role in SR based classification.

Anti-community detection in networks can discover negative relations among objects. However, a few researches pay attention to detecting anti-community structure and they do not consider the node degree and most of them require high computational cost. Block models are promising methods for exploring modular regularities, but their results are highly dependent on the observed structure. In this paper, we first propose a Degree-based Block Model (DBM) for anti-community structure. DBM takes the node degree into consideration and evolves a new objective function Q(C) for evaluation. And then, a Local Expansion Optimization Algorithm (LEOA), which preferentially considers the nodes with high degree, is proposed for anti-community detection. LEOA consists of three stages: structural center detection, local anti-community expansion and group membership adjustment. Based on the formulation of DBM, we develop a synthetic benchmark DBM-Net for evaluating comparison algorithms in detecting known anti-community structures. Experiments on DBM-Net with up to 100000 nodes and 17 real-world networks demonstrate the effectiveness and efficiency of LEOA for anti-community detection in networks.

Online feature selection is a challenging topic in data mining. It aims to reduce the dimensionality of streaming features by removing irrelevant and redundant features in real time. Existing works, such as Alpha-investing and Online Streaming Feature Selection (OSFS), have been proposed to serve this purpose, but they have drawbacks, including low prediction accuracy and high running time if the streaming features exhibit characteristics such as low redundancy and high relevance. In this paper, we propose a novel algorithm about online streaming feature selection, named ConInd that uses a three-layer filtering strategy to process streaming features with the aim of overcoming such drawbacks. Through three-layer filtering, i.e., null-conditional independence, single-conditional independence, and multi-conditional independence, we can obtain an approximate Markov blanket with high accuracy and low running time. To validate the efficiency, we implemented the proposed algorithm and tested its performance on a prevalent dataset, i.e., NIPS 2003 and Causality Workbench. Through extensive experimental results, we demonstrated that ConInd offers significant performance improvements in prediction accuracy and running time compared to Alpha-investing and OSFS. ConInd offers 5.62% higher average prediction accuracy than Alpha-investing, with a 53.56% lower average running time compared to that for OSFS when the dataset is lowly redundant and highly relevant. In addition, the ratio of the average number of features for ConInd is 242% less than that for Alpha-investing.

Reviewer assignment is critical to peer review systems, such as peer-reviewed research conferences or peer-reviewed funding applications, and its effectiveness is a deep concern of all academics. However, there are some problems in existing peer review systems during reviewer assignment. For example, some of the reviewers are much more stringent than others, leading to an unfair final decision, i.e., some submissions (i.e., papers or applications) with better quality are rejected. In this paper, we propose a context-aware reviewer assignment for trust enhanced peer review. More specifically, in our approach, we first consider the research area specific expertise of reviewers, and the institution relevance and co-authorship between reviewers and authors, so that reviewers with the right expertise are assigned to the corresponding submissions without potential conflict of interest. In addition, we propose a novel cross-assignment paradigm, and reviewers are cross-assigned in order to avoid assigning a group of stringent reviewers or a group of lenient reviewers to the same submission. More importantly, on top of them, we propose an academic CONtext-aware expertise relevanCe oriEnted Reviewer cross-assignmenT approach (CONCERT), which aims to effectively estimate the "true" ratings of submissions based on the ratings from all reviewers, even though no prior knowledge exists about the distribution of stringent reviewers and lenient reviewers. The experiments illustrate that compared with existing approaches, our proposed CONCERT approach can less likely assign more than one stringent reviewers or lenient reviewers to a submission simultaneously and significantly reduce the influence of ratings from stringent reviewers and lenient reviewers, leading to trust enhanced peer review and selection, no matter what kind of distributions of stringent reviewers and lenient reviewers are.

Computer intrusions are taking place everywhere, and have become a major concern for information security. Most intrusions to a computer system may result from illegitimate or irregular calls to the operating system, so analyzing the system-call sequences becomes an important and fundamental technique to detect potential intrusions. This paper proposes two models based on data mining technology, respectively called frequency patterns () and tree patterns () for intrusion detection. employs a typical method of sequential mining based on frequency analysis, and uses a short sequence model to find out quickly frequent sequential patterns in the training system-call sequences. makes use of the technique of tree pattern mining, and can get a quality profile from the training system-call sequences of a given system. Experimental results show that has good performances in training and detecting intrusions from short system-call sequences, and can achieve a high detection precision in handling long sequences.