Search

• We treat the solution selection procedure in EAs as a fuzzy classification problem to reduce the number of FEs. Selected solutions belong to the 'promising' class, while discarded solutions belong to the 'unpromising' class • We propose a fuzzy-classification-assisted selection (FCAS) strategy to decide solutions for FE. Different from the existing classification-based strategies that decide solution evaluations according to the predicted labels, we use fuzzy membership degrees, which is more reliable than only using the labels. • The proposed FCAS strategy is a general algorithm framework, where different kinds of fuzzy classification models can be applied, and the FCAS can be applied to different kinds of EAs. We integrate FCAS into two state-of-the-art algorithms on three classical test suites. The experimental results show that the number of FEs can be significantly reduced by our proposed FCAS when the same fitness values are achieved. In evolutionary optimization, solution selection is an important operator since it will be normally used to decide the optimization direction via determining new solutions. Most selection methods are objective fitness-based approaches which will lead to a waste of fitness evaluations. This is because some evaluated but unpromising solutions are discarded without contributing useful search information. We are, thus, motivated to treat the solution selection as a classification procedure, where the selected solutions and discarded solutions belong to different classes. However, another problem is that the difference between 'promising' and 'unpromising' solutions becomes fuzzy when iterations go on. Therefore, we employ fuzzy classification to predict the categories of solutions by the fuzzy membership function. And then the predicted results are used to assist solution selection to reduce the number of fitness evaluations. Finally, we propose a fuzzy-classification-assisted selection (FCAS) strategy to boost evolutionary optimization. FCAS is experimentally integrated into two state-of-the-art algorithms and studied on three test suites. The results reveal the efficiency of FCAS for boosting evolutionary optimization. [ABSTRACT FROM AUTHOR]

The belief rule-based (BRB) system has demonstrated advantages in complex system modeling and evaluation, with strong nonlinear relationship approximation capabilities. BRB parameter learning processes have been proved to be effective in improving the approximation accuracy of BRB systems. However, the running time complexity is regarded as an important challenge in BRB parameter learning efficiency. In this paper, a new approach to BRB parameter learning via extended causal strength (CAST) logic (BRBcast) is proposed in order to reduce the complexity of BRB parameter learning and maintain the approximation accuracy of BRB systems. First, the parameter numbers of traditional BRB parameter learning are analyzed to show the necessity of complexity reduction. Furthermore, the binary CAST logic is extended to fulfill the requirements of multi-state modeling and evaluation. Thereafter, an optimization model for parameter learning with CAST logic is established based on the analysis conclusion, and further applied to reduce the BRB parameter learning complexity. In BRBcast, the CAST parameters, instead of BRB parameters, are trained and translated to construct belief rule bases in BRB parameter learning, which involves less parameters than those of traditional BRB parameter learning approaches. Following this, the detailed BRBcast procedure is presented with the differential evolutionary (DE) algorithm. Finally, a numerical case and practical example on pipeline leak detection are investigated in order to verify the efficiency of BRBcast. The experimental results indicate that the proposed BRBcast exhibits superior performance, in both reducing the BRB parameter learning complexity and ensuring the approximation accuracy of BRB systems, which provides a promising avenue for constructing accurate and robust disaster emergency and rapid response systems. [ABSTRACT FROM AUTHOR]

Abstract: Proximity among query terms has been found to be useful for improving retrieval performance. However, its application to classical probabilistic information retrieval models, such as Okapi’s BM25, remains a challenging research problem. In this paper, we propose to improve the classical BM25 model by utilizing the term proximity evidence. Four novel methods, namely a window-based N-gram Counting method, Survival Analysis over different statistics, including the Poisson process, an exponential distribution and an empirical function, are proposed to model the proximity between query terms. Through extensive experiments on standard TREC collections, our proposed proximity-based BM25 model, called BM25P, is compared to strong state-of-the-art evaluation baselines, including the original unigram BM25 model, the Markov Random Field model, and the positional language model. According to the experimental results, the window-based N-gram Counting method, and Survival Analysis over an exponential distribution are the most effective among all four proposed methods, which lead to marked improvement over the baselines. This shows that the use of term proximity considerably enhances the retrieval effectiveness of the classical probabilistic models. It is therefore recommended to deploy a term proximity component in retrieval systems that employ probabilistic models. [Copyright &y& Elsevier]

Recently, attention-based neural networks (NNs) have been widely used for aspect-level sentiment classification (ASC). Most neural models focus on incorporating the aspect representation into attention, however, the position information of each aspect is not studied well. Furthermore, the existing ASC datasets are relatively small owing to the labor-intensive labeling that largely limits the performance of NNs. In this paper, we propose a position-aware hierarchical transfer (PAHT) model that models the position information from multiple levels and enhances the ASC performance by transferring hierarchical knowledge from the resource-rich sentence-level sentiment classification (SSC) dataset. We first present aspect-based positional attention in the word and the segment levels to capture more salient information toward a given aspect. To make up for the limited data for ASC, we devise three sampling strategies to select related instances from the large-scale SSC dataset for pre-training and transfer the learned knowledge into ASC from four levels: embedding, word, segment and classifier. Extensive experiments on four benchmark datasets demonstrate that the proposed model is effective in improving the performance of ASC. Particularly, our model outperforms the state-of-the-art approaches in terms of accuracy over all the datasets considered. [ABSTRACT FROM AUTHOR]

In response to fighting COVID-19 pandemic, researchers in machine learning and artificial intelligence have constructed some medical knowledge graphs (KG) based on existing COVID-19 datasets, however, these KGs contain a considerable amount of semantic relations which are incomplete or missing. In this paper, we focus on the task of knowledge graph embedding (KGE), which serves an important solution to infer the missing relations. In the past, there have been a collection of knowledge graph embedding models with different scoring functions to learn entity and relation embeddings published. However, these models share the same problems of rarely taking important features of KG like attribute features, other than relation triples, into account, while dealing with the heterogeneous, complex and incomplete COVID-19 medical data. To address the above issue, we propose a graph feature collection network (GFCNet) for COVID-19 KGE task, which considers both neighbor and attribute features in KGs. The extensive experiments conducted on the COVID-19 drug KG dataset show promising results and prove the effectiveness and efficiency of our proposed model. In addition, we also explain the future directions of deepening the study on COVID-19 KGE task. [ABSTRACT FROM AUTHOR]

Currently, Knowledge Base Question Answering (KBQA) is an important research topic in the fields of information retrieval (IR) and natural language processing (NLP). The most common questions asked on the Web are simple questions, which can be answered by a single relational fact in a knowledge base (KB). However, answering simple questions automatically remains a challenging task in the IR and NLP research communities. Based on a review of various studies and a detailed analysis, we surmise that these challenges are primarily related to the following concerns: (1) how to effectively access a large-scale KB; and (2) how to effectively reduce the gap between NL questions and the structured semantics in a KB. Most previous studies have considered these as separate and independent subtasks, subject detection and predicate matching. Here, we propose a deep fused model that combines subject detection and predicate matching under a unified framework. Specifically, we employ a subject detection model to recognize the subject entity in a question, and a multilevel semantic model to learn the semantic representations for questions and predicates. These models share parameters, and can be trained jointly. We evaluated the proposed method on both English and Chinese KBQA datasets. The experimental results demonstrate that the proposed approach significantly outperforms state-of-the-art systems when applied to both datasets. [ABSTRACT FROM AUTHOR]