Search

Motivated by the recent effort on scenario-based context question answering (QA), this paper investigates the role of discourse processing and its implication on query expansion for a sequence of questions. Our view is that a question sequence is not random, but rather follows a coherent manner to serve some information goals. Therefore, this sequence of questions can be considered as a mini discourse with some characteristics of discourse cohesion. Understanding such a discourse will help QA systems better interpret questions and retrieve answers. Thus, we examine three models driven by Centering Theory for discourse processing: a reference model that resolves pronoun references for each question, a forward model that makes use of the forward looking centers from previous questions, and a transition model that takes into account the transition state between adjacent questions. Our empirical results indicate that more sophisticated processing based on discourse transitions and centers can significantly improve the performance of document retrieval compared to models that only resolve references. This paper provides a systematic evaluation of these models and discusses their potentials and limitations in processing coherent context questions.

Three-way concept analysis is a newly proposed area of formal concept analysis from which one can obtain both the inclusion decision and the exclusion decision. In general, given a context, some attributes may not be essential in three-way concept analysis, such as forming three-way concept lattice. So in this paper, we study the attribute reductions of three-way concept lattices in order to make the data easily be understood. Firstly, based on different criteria generated from object-induced three-way concept (OE-concept), four kinds of attribute reductions are proposed. The four reductions together embody different characteristics of a formal context and can be used in different occasions. Secondly, we discuss their relationships, including their advantages and disadvantages and the relationships among consistent sets and among the cores. Thirdly, based on attribute-induced three-way concept (AE-concept), we also give four attribute-induced three-way attribute reductions and discuss their relationships. Finally, the approaches to computing these attribute reductions are presented and the obtained results are demonstrated and verified by an empirical case. In this paper, we systematically investigate the attribute reductions of three-way concept lattices which enriches the study of formal concept analysis.

Gradient boosted trees (GBTs) are ubiquitous models used by researchers, machine learning (ML) practitioners, and data scientists because of their robust performance, interpretable behavior, and ease-of-use. One critical challenge in training GBTs is the tuning of their hyperparameters. In practice, selecting these hyperparameters is often done manually. Recently, the ML community has advocated for tuning hyperparameters through black-box optimization and developed state-of-the-art systems to do so. However, applying such systems to tune GBTs suffers from two drawbacks. First, these systems are not \textit{model-aware}, rather they are designed to apply to a \textit{generic} model; this leaves significant optimization performance on the table. Second, using these systems requires \textit{domain knowledge} such as the choice of hyperparameter search space, which is an antithesis to the automatic experimentation that black-box optimization aims to provide. In this paper, we present SigOpt Mulch, a model-aware hyperparameter tuning system specifically designed for automated tuning of GBTs that provides two improvements over existing systems. First, Mulch leverages powerful techniques in metalearning and multifidelity optimization to perform model-aware hyperparameter optimization. Second, it automates the process of learning performant hyperparameters by making intelligent decisions about the optimization search space, thus reducing the need for user domain knowledge. These innovations allow Mulch to identify good GBT hyperparameters far more efficiently -- and in a more seamless and user-friendly way -- than existing black-box hyperparameter tuning systems.

The article of record as published may be found at https://doi.org/10.1016/S0950-7051(99)00032-5 Despite the fact that there has been a surge of publications in verification and validation of knowledge-based systems and expert systems in the past decade, there are still gaps in the study of verification and validation (V&V) of expert systems, not the least of which is the lack of appropriate semantics for expert system programming languages. Without a semantics, it is hard to formally define and analyze knowledge base anomalies such as inconsistency and redundancy, and it is hard to assess the effectiveness of V&V tools, methods and techniques that have been developed or proposed. In this paper, we develop an approximate declarative semantics for rule-based knowledge bases and provide a formal definition and analysis of knowledge base inconsistency, redundancy, circularity and incompleteness in terms of theories in the first order predicate logic. In the paper, we offer classifications of commonly found cases of inconsistency, redundancy, circularity and incompleteness. Finally, general guidelines on how to remedy knowledge base anomalies are given.

In this paper we discuss the use of Data Mining to provide a solution to the problem of cross-sales. We define and analyse the cross-sales problem and develop a hybrid methodology to solve it, using characteristic rule discovery and deviation detection. Deviation detection is used as a measure of interest to filter out the less interesting characteristic rules and only retain the best characteristic rules discovered. The effect of domain knowledge on the interestingness value of the discovered rules is discussed and techniques for refining the knowledge to increase this interestingness measure are studied. We also investigate the use of externally procured lifestyle and other survey data for data enrichment and discuss its use as additional domain knowledge. The developed methodology has been applied to a real world cross-sales problem within the financial sector, and the results are also presented in this paper. Although the application described is in the financial sector, the methodology is generic in nature and can be applied to other sectors.

Although traditional symbolic reasoning methods are highly interpretable, their application in knowledge graph link prediction is limited due to their low computational efficiency. In this paper, we propose a new neural symbolic reasoning method: RNNCTPs, which improves computational efficiency by re-filtering the knowledge selection of Conditional Theorem Provers (CTPs), and is less sensitive to the embedding size parameter. RNNCTPs are divided into relation selectors and predictors. The relation selectors are trained efficiently and interpretably, so that the whole model can dynamically generate knowledge for the inference of the predictor. In all four datasets, the method shows competitive performance against traditional methods on the link prediction task, and can have higher applicability to the selection of datasets relative to CTPs., Comment: 16 pages, 8 figures,5 tables

Modelling and simulation are widely used methods in design and other engineering tasks. Providing libraries of reusable model fragments is a promising way of supporting engineering modelling. The paper discusses means of structuring such libraries in a generic and reusable way. Model content facts can be organized on the basis of the ontologies that are generally important in physical systems engineering. In addition, the paper suggests that such libraries need to support various kinds of modelling assumptions about model fragments. Modelling can be usefully approached as a form of assumption management, as a means of reflecting the metalevel knowledge that is employed by experienced engineers. A categorization of modelling assumptions is proposed that aims at further support for reusability and shareability.

The paper explores the implementation of rule-based pattern-directed inference systems on parallel computers. The paper discusses one of these approaches in detail, the use of a graph-reduction machine such as ALICE. The technique is illustrated through two example domains: automobile fault diagnosis and organic psychiatric mental disorders. The paper discusses extensions to the graph reduction technique as applied to knowledge-based systems, including partitioning, time considerations and input data types. The paper shows that the graph-reduction technique has significant advantages for knowledge-based system implementation over conventional approaches, and it demonstrates that this programming style is amenable to knowledge engineering domains.

Expert systems are consultative, highly interactive systems, and hence the quality of interaction between system and user is important for the acceptability of the system. Acquisition of data about the problem in hand is a central feature of the interaction between system and user, and it makes a major contribution to the user's perception of the system. It is therefore crucial that the acquisition of this data (both as individual data items and as a sequence of data inputs) is perceived by the user as competent. This paper identifies central, domain-independent, design goals for the information-acquisition interactions between an expert system and its user, including mixed-initiative interaction, flexibility in user input, and system competence in querying the user. The paper discusses the realisation of these goals in a diagnostic expert system, Skeletal Dysplasias Diagnostician, through an explicit data model which allows for the representation of data with temporal and spatial (locality) aspects and the decide-status function which operates on the data model.

Emotional Support Conversation (ESConv) aims to reduce help-seekers'emotional distress with the supportive strategy and response. It is essential for the supporter to select an appropriate strategy with the feedback of the help-seeker (e.g., emotion change during dialog turns, etc) in ESConv. However, previous methods mainly focus on the dialog history to select the strategy and ignore the help-seeker's feedback, leading to the wrong and user-irrelevant strategy prediction. In addition, these approaches only model the context-to-strategy flow and pay less attention to the strategy-to-context flow that can focus on the strategy-related context for generating the strategy-constrain response. In this paper, we propose a Feedback-Aware Double COntrolling Network (FADO) to make a strategy schedule and generate the supportive response. The core module in FADO consists of a dual-level feedback strategy selector and a double control reader. Specifically, the dual-level feedback strategy selector leverages the turn-level and conversation-level feedback to encourage or penalize strategies. The double control reader constructs the novel strategy-to-context flow for generating the strategy-constrain response. Furthermore, a strategy dictionary is designed to enrich the semantic information of the strategy and improve the quality of strategy-constrain response. Experimental results on ESConv show that the proposed FADO has achieved the state-of-the-art performance in terms of both strategy selection and response generation. Our code is available at https://github.com/Thedatababbler/FADO., Accepted on Knowl. Based Syst. (SCI I)

Financiado para publicar en acceso aberto. Universidade da Coruña/CISUG [Abstract]: In recent years, mainly due to the application of smartphones in this area, research in human activity recognition (HAR) has shown a continuous and steady growth. Thanks to its wide range of sensors, its size, its ease of use, its low price and its applicability in many other fields, it is a highly attractive option for researchers. However, the vast majority of studies carried out so far focus on laboratory settings, outside of a real-life environment. In this work, unlike in other papers, progress was sought on the latter point. To do so, a dataset already published for this purpose was used. This dataset was collected using the sensors of the smartphones of different individuals in their daily life, with almost total freedom. To exploit these data, numerous experiments were carried out with various machine learning techniques and each of them with different hyperparameters. These experiments proved that, in this case, tree-based models, such as Random Forest, outperform the rest. The final result shows an enormous improvement in the accuracy of the best model found to date for this purpose, from 74.39% to 92.97%. Xunta de Galicia; ED431G 2019/01 Xunta de Galicia; ED481A 2020/003 Xunta de Galicia; ED431C 2022/46 Xunta de Galicia; ED431C 2018/49 Xunta de Galicia; ED431C 2021/53 This research was partially funded by MCIN/AEI/10.13039/ 501100011033, NextGenerationEU/PRTR, FLATCITY-POC , Spain [grant number P DC2021-121239-C31]; MCIN/AEI/10.13039/ 501100011033 MAGIST, Spain [grant number P ID2019-105221RB-C41]; Xunta de Galicia/FEDER-UE, Spain [grant numbers ED431G 2019/01 , ED481A 2020/003 , ED431C 2022/46 , ED431C 2018/49 and ED431C 2021/53 ]. Funding for open access charge: Universidade da Coruña/CISUG.

Data availability: The code used in this paper is available at: https://github.com/mxt0607/Two_Stage_RUL_Prediction. This article puts forward a two-stage integrated method to predict the remaining useful life (RUL) of lithium-ion batteries (LIBs). At the first stage, a convolutional neural network (CNN) is employed to preliminarily estimate the cycle life of each testing LIB, where the network structure of the CNN is carefully designed to extract the discharge capacity features. By analyzing the cycle lives, an LIB which has the most similar degradation mode to each testing LIB is chosen from the training dataset. The capacities of the selected LIB are identified based on a double exponential model (DEM). At the second stage, the identified DEM is utilized as the initial mean function of the Gaussian process regression (GPR) algorithm. The GPR algorithm is then applied to early RUL prediction of each testing LIB in a personalized manner. To verify the efficacy of the proposed method, four LIBs with long-term cycle lives are selected as the testing dataset. Experimental results show the superior performance of the proposed method over the standard CNN-based RUL prediction method and the standard GPR-based RUL prediction method. This work was supported in part the National Natural Science Foundation of China under Grants 62273264, 51905197 and 61933007.

Evolving out of theoretical and practical work, the paper presents the motivation and basic ideas for the construction and use of modular knowledge bases. The approach relates to earlier work carried out by each of the two authors of the paper separately. A model is introduced that merges the two previous approaches, modules for logical knowledge bases, and ordering by generality domains, while maintaining their benefits. Central aims are reusability, the restriction of memory searching, and the management of inconsistent (competing) knowledge within one knowledge base. The model is explained using examples, and the formal semantics are discussed of structured, modular knowledge bases for knowledge representations that are based on logic programming.

This paper investigates Graph Neural Networks (GNNs) application for self-supervised network intrusion and anomaly detection. GNNs are a deep learning approach for graph-based data that incorporate graph structures into learning to generalise graph representations and output embeddings. As network flows are naturally graph-based, GNNs are a suitable fit for analysing and learning network behaviour. The majority of current implementations of GNN-based Network Intrusion Detection Systems (NIDSs) rely heavily on labelled network traffic which can not only restrict the amount and structure of input traffic, but also the NIDSs potential to adapt to unseen attacks. To overcome these restrictions, we present Anomal-E, a GNN approach to intrusion and anomaly detection that leverages edge features and graph topological structure in a self-supervised process. This approach is, to the best our knowledge, the first successful and practical approach to network intrusion detection that utilises network flows in a self-supervised, edge leveraging GNN. Experimental results on two modern benchmark NIDS datasets not only clearly display the improvement of using Anomal-E embeddings rather than raw features, but also the potential Anomal-E has for detection on wild network traffic., Comment: Accepted by Knowledge-Based Systems

Gene expression data have become increasingly important in machine learning and computational biology over the past few years. In the field of gene expression analysis, several matrix factorization-based dimensionality reduction methods have been developed. However, such methods can still be improved in terms of efficiency and reliability. In this paper, an innovative approach to feature selection, called Dual Regularized Unsupervised Feature Selection Based on Matrix Factorization and Minimum Redundancy (DR-FS-MFMR), is introduced. The major focus of DR-FS-MFMR is to discard redundant features from the set of original features. In order to reach this target, the primary feature selection problem is defined in terms of two aspects: (1) the matrix factorization of data matrix in terms of the feature weight matrix and the representation matrix, and (2) the correlation information related to the selected features set. Then, the objective function is enriched by employing two data representation characteristics along with an inner product regularization criterion to perform both the redundancy minimization process and the sparsity task more precisely. To demonstrate the proficiency of the DR-FS-MFMR method, a large number of experimental studies are conducted on nine gene expression datasets. The obtained computational results indicate the efficiency and productivity of DR-FS-MFMR for the gene selection task.

Human motion prediction, which plays a key role in computer vision, generally requires a past motion sequence as input. However, in real applications, a complete and correct past motion sequence can be too expensive to achieve. In this paper, we propose a novel approach to predicting future human motions from a much weaker condition, i.e., a single image, with mixture density networks (MDN) modeling. Contrary to most existing deep human motion prediction approaches, the multimodal nature of MDN enables the generation of diverse future motion hypotheses, which well compensates for the strong stochastic ambiguity aggregated by the single input and human motion uncertainty. In designing the loss function, we further introduce the energy-based formulation to flexibly impose prior losses over the learnable parameters of MDN to maintain motion coherence as well as improve the prediction accuracy by customizing the energy functions. Our trained model directly takes an image as input and generates multiple plausible motions that satisfy the given condition. Extensive experiments on two standard benchmark datasets demonstrate the effectiveness of our method in terms of prediction diversity and accuracy.

Few-shot point cloud semantic segmentation learns to segment novel classes with scarce labeled samples. Within an episode, a novel target class is defined by a few support samples with corresponding binary masks, where only the points of this class are labeled as foreground and others are regarded as background. In the tasks involving multiple target classes, since the meanings of background are diverse for different target classes, background ambiguities appear: Some points labeled as background in one support sample may be of other target classes. It will result in incorrect guidance and damage model's segmentation performance. However, previous methods in the literature do not consider this problem. In this paper, we propose a simple yet effective approach to tackle background ambiguities, which adopts the entropy of predictions on query samples to the training objective function as an additional regularization. Besides, we design a feature transformation operation to reduce the feature differences between support and query samples. With our proposed approach, fine-tuning, a weak baseline method for few-shot segmentation, gains significant performance improvement (e.g., 7.48% and 7.04% in 2-way-1-shot and 3-way-1-shot tasks of S3DIS, respectively) and outperforms current state-of-the-art methods in all the task settings of S3DIS and ScanNet benchmark datasets. Ministry of Education (MOE) National Research Foundation (NRF) This work was supported by National Natural Science Foundation of China (NSFC) 61876208, Tip-top Scientific and Technical Innovative Youth Talents of Guangdong Special Support Program (2019TQ05X200), 2022 Tencent Wechat Rhino-Bird Focused Research Program Research (Tencent WeChat RBFR2022008) and the National Research Foundation, Singapore under its AI Singapore Programme (AISG Award No: AISG-RP-2018-003), the Ministry of Education, Singapore, under its Academic Research Fund Tier 2 (MOE-T2EP20220-0007) and Tier 1 (RG95/20).

This paper presents a batch-wise density-based clustering approach for local outlier detection in massive-scale datasets. Unlike the well-known traditional algorithms, which assume that all the data is memory-resident, our proposed method is scalable and processes the input data chunk-by-chunk within the confines of a limited memory buffer. A temporary clustering model is built at the first phase; then, it is gradually updated by analyzing consecutive memory loads of points. Subsequently, at the end of scalable clustering, the approximate structure of the original clusters is obtained. Finally, by another scan of the entire dataset and using a suitable criterion, an outlying score is assigned to each object called SDCOR (Scalable Density-based Clustering Outlierness Ratio). Evaluations on real-life and synthetic datasets demonstrate that the proposed method has a low linear time complexity and is more effective and efficient compared to best-known conventional density-based methods, which need to load all data into the memory; and also, to some fast distance-based methods, which can perform on data resident in the disk.

Automatic code review (ACR), which can relieve the costs of manual inspection, is an indispensable and essential task in software engineering. To deal with ACR, existing work is to serialize the abstract syntax tree (AST). However, making sense of the whole AST with sequence encoding approach is a daunting task, mostly due to some redundant nodes in AST hinder the transmission of node information. Not to mention that the serialized representation is inadequate to grasp the information of tree structure in AST. In this paper, we first present a new large-scale Apache Automatic Code Review (AACR) dataset for ACR task since there is still no publicly available dataset in this task. The release of this dataset would push forward the research in this field. Based on it, we propose a novel Simplified AST based Graph Convolutional Network (SimAST-GCN) to deal with ACR task. Concretely, to improve the efficiency of node information dissemination, we first simplify the AST of code by deleting the redundant nodes that do not contain connection attributes, and thus deriving a Simplified AST. Then, we construct a relation graph for each code based on the Simplified AST to properly embody the relations among code fragments of the tree structure into the graph. Subsequently, in the light of the merit of graph structure, we explore a graph convolution networks architecture that follows an attention mechanism to leverage the crucial implications of code fragments to derive code representations. Finally, we exploit a simple but effective subtraction operation in the representations between the original and revised code, enabling the revised difference to be preferably learned for deciding the results of ACR. Experimental results on the AACR dataset illustrate that our proposed model outperforms the state-of-the-art methods., 15 pages

Federated learning (FL) is a privacy-preserving paradigm that collaboratively train machine learning models with distributed data stored in different silos without exposing sensitive information. Different from most existing FL approaches requiring data from different parties share either the same feature space or sample ID space, federated transfer learning (FTL), which is a recently proposed FL concept, is designed for situations where data from different parties differ not only in samples but also in feature space. However, like most traditional FL approaches, FTL methods also suffer from issues caused by insufficiency of overlapping data. In this paper, we propose a novel FTL framework referred to as Semi-Supervised Federated Heterogeneous Transfer Learning (SFHTL) to leverage on the unlabeled non-overlapping samples to reduce model overfitting as a result of insufficient overlapping training samples in FL scenarios. Unlike existing FTL approaches, SFHTL makes use of non-overlapping samples from all parties to expand the training set for each party to improve local model performance. Through extensive experimental evaluation based on real-world datasets, we demonstrate significant advantages of SFHTL over state-of-the-art approaches. Nanyang Technological University National Research Foundation (NRF) This research is supported, in part, by the National Natural Science Foundation of China under Grant NSFC 62106167; Project Funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions; the National Research Foundation, Singapore under its AI Singapore Programme (AISG Award No: AISG2-RP-2020-019); the RIE 2020 Advanced Manufacturing and Engineering (AME) Programmatic Fund (No. A20G8b0102), Singapore; Nanyang Assistant Professorship (NAP); the Joint NTUWeBank Research Centre on Fintech (Award No: NWJ-2020-008), Nanyang Technological University, Singapore; and Joint SDU-NTU Centre for Artificial Intelligence Research (C-FAIR) (NSC-2019- 011). Qiang Yang is supported in part by China National Key Research and Development Program of China under Grant No. 2018AAA0101100.

Nonnegative matrix factorization (NMF) has been widely studied in recent years due to its effectiveness in representing nonnegative data with parts-based representations. For NMF, a sparser solution implies better parts-based representation.However, current NMF methods do not always generate sparse solutions.In this paper, we propose a new NMF method with log-norm imposed on the factor matrices to enhance the sparseness.Moreover, we propose a novel column-wisely sparse norm, named $\ell_{2,\log}$-(pseudo) norm to enhance the robustness of the proposed method.The $\ell_{2,\log}$-(pseudo) norm is invariant, continuous, and differentiable.For the $\ell_{2,\log}$ regularized shrinkage problem, we derive a closed-form solution, which can be used for other general problems.Efficient multiplicative updating rules are developed for the optimization, which theoretically guarantees the convergence of the objective value sequence.Extensive experimental results confirm the effectiveness of the proposed method, as well as the enhanced sparseness and robustness.

Bayesian transfer learning (BTL) is defined in this paper as the task of conditioning a target probability distribution on a transferred source distribution. The target globally models the interaction between the source and target, and conditions on a probabilistic data predictor made available by an independent local source modeller. Fully probabilistic design is adopted to solve this optimal decision-making problem in the target. By successfully transferring higher moments of the source, the target can reject unreliable source knowledge (i.e. it achieves robust transfer). This dual-modeller framework means that the source's local processing of raw data into a transferred predictive distribution -- with compressive possibilities -- is enriched by (the possible expertise of) the local source model. In addition, the introduction of the global target modeller allows correlation between the source and target tasks -- if known to the target -- to be accounted for. Important consequences emerge. Firstly, the new scheme attains the performance of fully modelled (i.e. conventional) multitask learning schemes in (those rare) cases where target model misspecification is avoided. Secondly, and more importantly, the new dual-modeller framework is robust to the model misspecification that undermines conventional multitask learning. We thoroughly explore these issues in the key context of interacting Gaussian process regression tasks. Experimental evidence from both synthetic and real data settings validates our technical findings: that the proposed BTL framework enjoys robustness in transfer while also being robust to model misspecification.

In this paper, a novel unsupervised low-rank representation model, i.e., Auto-weighted Low-Rank Representation (ALRR), is proposed to construct a more favorable similarity graph (SG) for clustering. In particular, ALRR enhances the discriminability of SG by capturing the multi-subspace structure and extracting the salient features simultaneously. Specifically, an auto-weighted penalty is introduced to learn a similarity graph by highlighting the effective features, and meanwhile, overshadowing the disturbed features. Consequently, ALRR obtains a similarity graph that can preserve the intrinsic geometrical structures within the data by enforcing a smaller similarity on two dissimilar samples. Moreover, we employ a block-diagonal regularizer to guarantee the learned graph contains $k$ diagonal blocks. This can facilitate a more discriminative representation learning for clustering tasks. Extensive experimental results on synthetic and real databases demonstrate the superiority of ALRR over other state-of-the-art methods with a margin of 1.8\%$\sim$10.8\%.

Yoga is a traditional Indian exercise. It specifies various body postures called asanas, practicing them is beneficial for the physical, mental, and spiritual well-being. To support the yoga practitioners, there is a need of an expert yoga asanas recognition system that can automatically analyze practitioner’s postures and could provide suitable posture correction instructions. This\ud paper proposes YogNet, a multi-person yoga expert system for 20 asanas using a two-stream deep\ud spatiotemporal neural network architecture. The first stream utilizes a keypoint detection approach\ud to detect the practitioner’s pose, followed by the formation of bounding boxes across the subject. The model then applies time distributed convolutional neural networks (CNNs) to extract framewise postural features, followed by regularized long short-term memory (LSTM) networks to give temporal predictions. The second stream utilizes 3D-CNNs for spatiotemporal feature extraction from RGB videos. Finally, the scores of two streams are fused using multiple fusion techniques.\ud A yoga asana recognition database (YAR) containing 1206 videos is collected using a single 2D web camera for 367 minutes with the help of 16 participants and contains four view variations i.e. front, back, left, and right sides. The proposed system is novel as this is the earliest two-stream deep learning-based system that can perform multi-person yoga asanas recognition and correction\ud in realtime. Simulation result reveals that YogNet system achieved 77.29%, 89.29%, and 96.31% accuracies using pose stream, RGB stream, and via fusion of both streams, respectively. These results are impressive and sufficiently high for recommendation towards general adaption of the system.

Blind image deblurring is an important yet very challenging problem in low-level vision. Traditional optimization based methods generally formulate this task as a maximum-a-posteriori estimation or variational inference problem, whose performance highly relies on the handcraft priors for both the latent image and the blur kernel. In contrast, recent deep learning methods generally learn, from a large collection of training images, deep neural networks (DNNs) directly mapping the blurry image to the clean one or to the blur kernel, paying less attention to the physical degradation process of the blurry image. In this paper, we present a deep variational Bayesian framework for blind image deblurring. Under this framework, the posterior of the latent clean image and blur kernel can be jointly estimated in an amortized inference fashion with DNNs, and the involved inference DNNs can be trained by fully considering the physical blur model, together with the supervision of data driven priors for the clean image and blur kernel, which is naturally led to by the evidence lower bound objective. Comprehensive experiments are conducted to substantiate the effectiveness of the proposed framework. The results show that it can not only achieve a promising performance with relatively simple networks, but also enhance the performance of existing DNNs for deblurring.

With the growing demand for hand hygiene and convenience of use, palmprint recognition with touchless manner made a great development recently, providing an effective solution for person identification. Despite many efforts that have been devoted to this area, it is still uncertain about the discriminative ability of the contactless palmprint, especially for large-scale datasets. To tackle the problem, in this paper, we build a large-scale touchless palmprint dataset containing 2334 palms from 1167 individuals. To our best knowledge, it is the largest contactless palmprint image benchmark ever collected with regard to the number of individuals and palms. Besides, we propose a novel deep learning framework for touchless palmprint recognition named 3DCPN (3D Convolution Palmprint recognition Network) which leverages 3D convolution to dynamically integrate multiple Gabor features. In 3DCPN, a novel variant of Gabor filter is embedded into the first layer for enhancement of curve feature extraction. With a well-designed ensemble scheme,low-level 3D features are then convolved to extract high-level features. Finally on the top, we set a region-based loss function to strengthen the discriminative ability of both global and local descriptors. To demonstrate the superiority of our method, extensive experiments are conducted on our dataset and other popular databases TongJi and IITD, where the results show the proposed 3DCPN achieves state-of-the-art or comparable performances.

Knowledge graph (KG) embedding encodes the entities and relations from a KG into low-dimensional vector spaces to support various applications such as KG completion, question answering, and recommender systems. In real world, knowledge graphs (KGs) are dynamic and evolve over time with addition or deletion of triples. However, most existing models focus on embedding static KGs while neglecting dynamics. To adapt to the changes in a KG, these models need to be retrained on the whole KG with a high time cost. In this paper, to tackle the aforementioned problem, we propose a new context-aware Dynamic Knowledge Graph Embedding (DKGE) method which supports the embedding learning in an online fashion. DKGE introduces two different representations (i.e., knowledge embedding and contextual element embedding) for each entity and each relation, in the joint modeling of entities and relations as well as their contexts, by employing two attentive graph convolutional networks, a gate strategy, and translation operations. This effectively helps limit the impacts of a KG update in certain regions, not in the entire graph, so that DKGE can rapidly acquire the updated KG embedding by a proposed online learning algorithm. Furthermore, DKGE can also learn KG embedding from scratch. Experiments on the tasks of link prediction and question answering in a dynamic environment demonstrate the effectiveness and efficiency of DKGE., Comment: 46 pages

Meta-learning has gained wide popularity as a training framework that is more data-efficient than traditional machine learning methods. However, its generalization ability in complex task distributions, such as multimodal tasks, has not been thoroughly studied. Recently, some studies on multimodality-based meta-learning have emerged. This survey provides a comprehensive overview of the multimodality-based meta-learning landscape in terms of the methodologies and applications. We first formalize the definition of meta-learning in multimodality, along with the research challenges in this growing field, such as how to enrich the input in few-shot learning (FSL) or zero-shot learning (ZSL) in multimodal scenarios and how to generalize the models to new tasks. We then propose a new taxonomy to discuss typical meta-learning algorithms in multimodal tasks systematically. We investigate the contributions of related papers and summarize them by our taxonomy. Finally, we propose potential research directions for this promising field., Comment: Accepted by Knowledge-Based Systems; 21 pages

Many important real-world applications involve time-series data with skewed distribution. Compared to conventional imbalance learning problems, the classification of imbalanced time-series data is more challenging due to high dimensionality and high inter-variable correlation. This paper proposes a structure preserving Oversampling method to combat the High-dimensional Imbalanced Time-series classification (OHIT). OHIT first leverages a density-ratio based shared nearest neighbor clustering algorithm to capture the modes of minority class in high-dimensional space. It then for each mode applies the shrinkage technique of large-dimensional covariance matrix to obtain accurate and reliable covariance structure. Finally, OHIT generates the structure-preserving synthetic samples based on multivariate Gaussian distribution by using the estimated covariance matrices. Experimental results on several publicly available time-series datasets (including unimodal and multimodal) demonstrate the superiority of OHIT against the state-of-the-art oversampling algorithms in terms of F1, G-mean, and AUC.

The last two decades have seen a tremendous increase in the adoption of Semantic Web technologies as a result of the availability of big data, the growth in computational power and the advancement of artificial intelligence (AI) technologies. Cutting-edge semantic techniques are now able to capture sentiments more accurately in various practical applications, including economic and financial forecasting. In particular, the extraction of sentiment from news text, social media and blogs for the prediction of economic and financial variables has attracted attention in recent years. Despite many successful applications of sentiment analysis (SA) in these domains, the range of semantic techniques employed is still limited and mostly focused on the detection of sentiment at a coarse-grained level, that is, whether the sentiment expressed by the entire text of a sentence is either positive or negative. This paper proposes a novel methodology for Fine-Grained Aspect-based Sentiment (FiGAS) analysis. The aim of the approach is to identify the sentiment associated to specific topics of interest in each sentence of a document and assigning real-valued polarity scores between -1 and +1 to those topics. The proposed approach is completely unsupervised and customized to the economic and financial domains by using a specialized lexicon make available along with the source code of FiGAS. Our lexicon-based SA approach relies on a detailed set of semantic polarity rules that allow understanding the origin of sentiment, in the spirit of the recent trend on \textit{Interpretable AI}. We provide an in-depth comparison of the performance of the FiGAS algorithm relative to other popular lexicon-based SA approaches in predicting a humanly annotated data set in the economic and financial domains. Our results indicate that FiGAS statistically outperforms the other methods by providing a sentiment score that is closer to one of the human annotators.

In Video Surveillance age, the monitoring activity, especially from unmanned vehicles, needs some degree of autonomy in the scenario interpretation. Video Analysis tasks are crucial for the target tracking and recognition; anyway, it would be desirable if a further level of understanding could provide a comprehensive, high-level scene description, by reflecting that human cognitive capability of providing a concise scene description that comes from the analysis of involved objects relationships and actions. This paper presents a smart system to identify mobile scene objects, such as people, vehicles, automatically, by analyzing the videos acquired by drones in flight, along with the activities they carried out, so as to depict what it happens in the scene from a high-level perspective. The system uses Artificial Vision methods to detect and track the mobile objects and the area where they move, and Semantic Web technologies to provide a high-level description of the scenario. Spatio/temporal relations among the tracked objects as well as simple object activities (events) are described. By semantic reasoning, the system is able to connect the simple activities into more complex activities, that better reflect a human-like description of a scenario portion. Tests conducted on several videos, showing scenarios set in different environments, return convincing results which affirm the effectiveness of the proposed approach.

Automatic segmentation of retinal vessels in fundus images plays an important role in the diagnosis of some diseases such as diabetes and hypertension. In this paper, we propose Deformable U-Net (DUNet), which exploits the retinal vessels’ local features with a U-shape architecture, in an end to end manner for retinal vessel segmentation. Inspired by the recently introduced deformable convolutional networks, we integrate the deformable convolution into the proposed network. The DUNet, with upsampling operators to increase the output resolution, is designed to extract context information and enable precise localization by combining low-level features with high-level ones. Furthermore, DUNet captures the retinal vessels at various shapes and scales by adaptively adjusting the receptive fields according to vessels’ scales and shapes. Public datasets: DRIVE, STARE, CHASE_DB1 and HRF are used to test our models. Detailed comparisons between the proposed network and the deformable neural network, U-Net are provided in our study. Results show that more detailed vessels can be extracted by DUNet and it exhibits state-of-the-art performance for retinal vessel segmentation with a global accuracy of 0.9566/0.9641/0.9610/0.9651 and AUC of 0.9802/0.9832/0.9804/0.9831 on DRIVE, STARE, CHASE_DB1 and HRF respectively. Moreover, to show the generalization ability of the DUNet, we use another two retinal vessel data sets, i.e., WIDE and SYNTHE, to qualitatively and quantitatively analyze and compare with other methods. Extensive cross-training evaluations are used to further assess the extendibility of DUNet. The proposed method has the potential to be applied to the early diagnosis of diseases.

A weighted random survival forest is presented in the paper. It can be regarded as a modification of the random forest improving its performance. The main idea underlying the proposed model is to replace the standard procedure of averaging used for estimation of the random survival forest hazard function by weighted averaging where the weights are assigned to every tree and can be viewed as training parameters which are computed in an optimal way by solving a standard quadratic optimization problem maximizing Harrell’s C-index. Numerical examples with real data illustrate the outperformance of the proposed model in comparison with the original random survival forest.

The file attached to this record is the author's final peer reviewed version. The Publisher's final version can be found by following the DOI link. The bounded confidence model is a popular tool to model the evolution of preferences and knowledge in opinion dynamics. In the bounded confidence model, it is assumed that all agents are honest to express their preferences and knowledge. However, in real-life opinion dynamics, agents often hide their true preferences, and express different preferences to different people. In this paper, we propose the evolution of preferences with deceptive interactions and heterogeneous trust in bounded confidence model, in which some agents will express three types of preferences: true preferences, communicated preferences and public preferences. In the proposed model, the communication regimes of the agents are established. Based on the established communication regimes, the true preferences, communicated preferences and public preferences of the agents are updated. Furthermore, we use an agent-based simulation to unfold the influences of the deceptive interactions and heterogeneous trust on the evolutions of preferences.

Recent advances in the integration of deep recurrent neural networks and statistical inferences have paved new avenues for joint modeling of moments of random variables, which is highly useful for signal processing, time series analysis, and financial forecasting. However, introducing explicit knowledge as exogenous variables has received little attention. In this paper, we propose a novel model termed sentiment-aware volatility forecasting (SAVING), which incorporates market sentiment for stock return fluctuation prediction. Our framework provides an ensemble of symbolic and sub-symbolic AI approaches, that is, including grounded knowledge into a connectionist neural network. The model aims at producing a more accurate estimation of temporal variances of asset returns by better capturing the bi-directional interaction between movements of asset price and market sentiment. The interaction is modeled using Variational Bayes via the data generation and inference operations. We benchmark our model with 9 other popular ones in terms of the likelihood of forecasts given the observed sequence. Experimental results suggest that our model not only outperforms pure statistical models, e.g., GARCH and its variants, Gaussian-process volatility model, but also outperforms the state-of-the-art autoregressive deep neural nets architectures, such as the variational recurrent neural network and the neural stochastic volatility model.

Structural learning of Bayesian Networks (BNs) is an NP-hard problem, and the use of sub-optimal strategies is essential in domains involving many variables. One approach is the generation of multiple approximate structures and then reduce the ensemble to a representative structure. This can be performed by using the occurrence frequency (on the structures ensemble) as the criteria for accepting a dominant directed edge between two nodes and thus obtaining the single structure. In this paper, an analogy with an adapted one-dimensional random-walk was used for analytically deducing an appropriate decision threshold to such occurrence frequency. The obtained closed-form expression has been validated across the synthetic datasets applying the Matthews Correlation Coefficient (MCC) as the performance metric. In the experiments using a recent medical dataset, the resulting BN from the analytical cutoff-frequency captured the expected associations among nodes and also achieved better prediction performance than the BNs learned with neighbours thresholds to the computed.

We introduce collective decision-making models associated with influence spread under the linear threshold model in social networks. We define the oblivious and the non-oblivious influence models. We also introduce the generalized opinion leader–follower model (gOLF) as an extension of the opinion leader–follower model (OLF) proposed by van den Brink et al. (2011). In our model we allow rules for the final decision different from the simple majority used in OLF. We show that gOLF models are non-oblivious influence models on a two-layered bipartite influence digraph. Together with OLF models, the satisfaction and the power measures were introduced and studied. We analyze the computational complexity of those measures for the decision models introduced in the paper. We show that the problem of computing the satisfaction or the power measure is #P-hard in all the introduced models even when the subjacent social network is a bipartite graph. Complementing this result, we provide two subfamilies of decision models in which both measures can be computed in polynomial time. We show that the collective decision functions are monotone and therefore they define an associated simple game. We relate the satisfaction and the power measures with the Rae index and the Banzhaf value of an associated simple game. This will allow the use of known approximation methods for computing the Banzhaf value, or the Rae index to their practical computation.

Correlation filters (CF) have demonstrated a good performance in visual tracking. However, the base training sample region is larger than the object region, including the interference region (IR). IRs in training samples from cyclic shifts of the base training sample severely degrade the quality of the tracking model. In this paper, a region-filtering correlation tracking (RFCT) algorithm is proposed to address this problem. In this algorithm, we filter training samples by introducing a spatial map into the standard CF formulation. Compared with the existing correlation filter trackers, the proposed tracker has the following advantages. (1) Using a spatial map, the correlation filter can be learned on a larger search region without the interference of IR. (2) Due to processing training samples by a spatial map, it is a more general way to control background information and target information in training samples. In addition, a better spatial map can be explored, the values of which are not restricted. Quantitative evaluations are performed on four benchmark datasets: OTB-2013, OTB-2015, VOT2015, and VOT2016. Experimental results demonstrate that the proposed RFCT algorithm performs favorably against several state-of-the-art methods.

The fast-growing e-commerce scenario brings new challenges to traditional collaborative filtering because the huge amount of users and items requires large storage and efficient recommendation systems. Hence, hashing for collaborative filtering has attracted increasing attention as binary codes can significantly reduce the storage requirement and make similarity calculations efficient. In this paper, we investigate the novel problem of deep collaborative hashing codes on user–item ratings. We propose a new deep learning framework for it, which adopts neural networks to better learn both user and item representations and make these close to binary codes such that the quantization loss is minimized. In addition, we extend the proposed framework for out-of-sample cases, i.e., dealing with new users, new items, and new ratings. Extensive experiments on real-world datasets demonstrate the effectiveness of the proposed framework.

In membrane computing, researchers have designed various P systems to solve problems, for example, NP-hard problems. The research of verification methods and simulation tools for these P systems is one of the main topics in membrane computing. In this paper, we present a general-purpose P system simulator named UPSimulator for cell-like, tissue-like and neural-like P systems. UPSimulator can serve as both a simulator and a development library for P system models, which processes objects in the regions defined by membranes. UPSimulator provides a feasible solution for the simulation of hybrid P systems by supporting the complex evolutionary rules and structures in different types of P systems.

While significant efforts for online learning have been devoted to arrive at reliable predictions of crisp values, the problem of prediction interval (PI) in practical data is one of the underexplored areas in the existing literature. PI aims to produce upper and lower bound predictions which capture possible domain solution. This paper aims to extend a prominent meta-cognitive learning algorithm, namely meta-cognitive interval type-2 fuzzy inference system (McIT2FIS), to cope with the problem of prediction interval in real-time. McIT2FIS is constructed under interval type-2 fuzzy inference system and realizes the meta-cognitive learning theory featuring the basic three elements of human learning: what-to-learn, how-to-learn, when-to-learn. Unlike existing works in PI, McIT2FIS-PI works fully in the online mode and is capable of performing automatic knowledge acquisition from data streams. The efficacy of McIT2FIS-PI has been experimentally validated in a real-world wave characteristics prediction in Semakau Island, Singapore, where it is capable of delivering accurate short-term prediction intervals of wave parameters. The performance of McIT2FIS-PI is also compared with existing state-of-the-art fuzzy inference systems in benchmark problems where it attains competitive accuracy while retaining comparable complexity. Nanyang Technological University The present work was supported by Energy Research Institute @Nanyang Technological University, Interdisciplinary Graduate School, Singapore.

Human creativity is usually assessed with a variety of established creativity tests. One of this is the Remote Associates Test (RAT), which aims to measure the ability of reaching remote associates with linguistic stimuli. A well known variant of the RAT exists – the compound RAT, for which normative data and solvers have been proposed in the literature. However, a different type of RAT was proposed in 1971 by Worthen and Clark – a functional form which had the potential of measuring other types of associations. However, the few test items proposed by Worthen and Clark where lost during archive transport, and cannot be accessed. In this paper, we set to reconstruct an ample set of functional items in the spirit of Worthen and Clark’s idea, using information science techniques. Cognitive word associates are used as data. The process of a former computational solver of the RAT is repurposed to create rather than solve items. The approach of constructing queries is evaluated by getting human participants to solve both functional and compound items. In the process, a previous computational approach to solving the compound RAT is also validated in the functional RAT context.

The set of closed high-utility itemsets (CHUIs) concisely represents the exact utility of all itemsets. Yet, it can be several orders of magnitude smaller than the set of all high-utility itemsets. Existing CHUI mining algorithms assume that databases are static, making them very expensive in the case of incremental data, since the whole dataset has to be processed for each batch of new transactions. To address this challenge, this paper presents the first approach, called IncCHUI, that mines CHUIs efficiently from incremental databases. In order to achieve this, we propose an incremental utility-list structure, which is built and updated with only one database scan. Further, we apply effective pruning strategies to fast construct incremental utility-lists and eliminate candidates that are not updated. Finally, we suggest an efficient hash-based approach to update or insert new closed sets that are found. Our extensive experimental evaluation on both real-life and synthetic databases shows the efficiency, as well as the feasibility of our approach. It significantly outperforms previously proposed methods that are mainly run in batch mode in terms of speed, and it is scalable with respect to the number of transactions.

Lexical resources are widely popular in the field of Sentiment Analysis, as they represent a resource that directly encodes sentimental knowledge. Usually sentiment lexica are used for polarity estimation through the matching of words contained in a text and their associated lexicon sentiment polarities. Nevertheless, such resources have limitations in vocabulary coverage and domain adaptation. Besides, many recent techniques exploit the concept of distributed semantics, normally through word embeddings. In this work, a semantic similarity metric is computed between text words and lexica vocabulary. Using this metric, this paper proposes a sentiment classification model that uses the semantic similarity measure in combination with embedding representations. In order to assess the effectiveness of this model, we perform an extensive evaluation. Experiments show that the proposed method can improve Sentiment Analysis performance over a strong baseline, being this improvement statistically significant. Finally, some characteristics of the proposed technique are studied, showing that the selection of lexicon words has an effect in cross-dataset performance.

Most thermal infrared (TIR) tracking methods are discriminative, treating the tracking problem as a classification task. However, the objective of the classifier (label prediction) is not coupled to the objective of the tracker (location estimation). The classification task focuses on the between-class difference of the arbitrary objects, while the tracking task mainly deals with the within-class difference of the same objects. In this paper, we cast the TIR tracking problem as a similarity verification task, which is coupled well to the objective of the tracking task. We propose a TIR tracker via a Hierarchical Spatial-aware Siamese Convolutional Neural Network (CNN), named HSSNet. To obtain both spatial and semantic features of the TIR object, we design a Siamese CNN that coalesces the multiple hierarchical convolutional layers. Then, we propose a spatial-aware network to enhance the discriminative ability of the coalesced hierarchical feature. Subsequently, we train this network end to end on a large visible video detection dataset to learn the similarity between paired objects before we transfer the network into the TIR domain. Next, this pre-trained Siamese network is used to evaluate the similarity between the target template and target candidates. Finally, we locate the candidate that is most similar to the tracked target. Extensive experimental results on the benchmarks VOT-TIR 2015 and VOT-TIR 2016 show that our proposed method achieves favourable performance compared to the state-of-the-art methods., 20 pages, 7 figures

Prevalent usage of social media attracted companies and researchers to analyze its dynamics and effects on user behavior. One of the most intriguing aspects of social networks is to identify influencers who are experts on a specific topic. With the identification of these users within the network, many applications can be built for user recommendation, information diffusion modeling, viral marketing, user modeling and many more. In this paper, we aim to identify topic-based experts using a large dataset collected from Twitter. Our proposed approach has three phases: (1) identification of topics on social media posts (more specifically, tweets), (2) user modeling, based on a group of user specific features, and (3) Influence Factorization to identify topical influencers. The main advantage of the proposed method is to identify future influencers as well as current ones on Twitter. Moreover, it is an easy to implement algorithm using Spark MLlib, which can be easily extended to include other user specific features, and compare with other methodologies. The effectiveness of the proposed method is tested on a large dataset that contains tweets of 180K user over 70 day period. The experimental results show that our proposed method identifies influencers successfully when used with a hybrid user specific feature that contains follower count and authenticity information, and is a highly scalable and extensible algorithm. (C) 2018 Elsevier B.V. All rights reserved.

Multiple attribute decision making (MADM) problems often consists of quantitative and qualitative attributes which can be assessed by numerical values and subjective judgments. Subjective judgments can be evaluated by linguistic variables, and both numerical values and subjective judgments can be accurate or uncertain. The evidential reasoning (ER) approach provides a process for dealing with MADM problems of both a quantitative and qualitative nature under uncertainty. The existing ER approach considers both benefit and cost attributes in the evidence combination process. In this paper, deviated interval and fixed interval attributes are introduced into ER based MADM approach and the frames of discernment for representing these two kinds of attributes are given. The transformation rules from the assessment values of deviated interval attributes to belief degrees in the ER structure are then studied. An ave-entropy based weight assignment method considering the risk preference of decision maker is also shown to deal with uncertain assessment situation, such as belief distribution with qualitative attribute and uncertain utility function. Some programming models to generate interval weights and utilities are constructed. The rationality and efficiency of the methods in supporting MADM problems are discussed. Two case studies are provided to demonstrate the applicability and validity of the proposed approaches and the potential in supporting MADM under uncertainty.

Granular computing is an emerging field of study in which the complexity of problem solving is reduced through granulation. Researchers have proposed various granularity measures of partitions to quantify the effects of granulation with respect to simplification. However, two important issues still remain and require careful investigation. The first issue is that a partition is only a simple two-level granular structure, which may not be sufficient for the full scope of granular computing. The second issue is a clarification of the differences between granularity and complexity. Although they are related to each other, they represent different things. To address the two issues, this paper makes three contributions. First, we extend the partition granulation scheme into multilevel granular structures based on progressive partitioning. Second, we propose a complexity measure of a partition that incorporates both the block-level interactions (interactions within a block) and the partition-level interactions (interactions between blocks of the partition). Third, we generalize the complexity measure to multilevel granular structures generated from a progressive partitioning process.

Top- k and skyline techniques have been used to address preference based queries for effective service selection. However, they do not consider the dependencies between attributes in user preferences. In this paper, we focus on developing top- k indexing methods based on Conditional Preference Networks. We first determine whether the correlation among service attributes is clear and definite. After that, we employ dimensionality reduction to reduce the dimensionality of the service space. We then use top- k query to further improve the scalability. We conduct extensive experiment and compare with other competitive indexing mechanism to demonstrate the effectiveness of the proposed approach.

The graph edit distance (GED) is a well-established distance measure widely used in many applications, such as bioinformatics, data mining, pattern recognition, and graph classification. However, existing solutions for computing the GED suffer from several drawbacks: large search spaces, excessive memory requirements, and many expensive backtracking calls. In this paper, we present BSS_GED , a novel vertex-based mapping method that calculates the GED in a reduced search space created by identifying invalid and redundant mappings. BSS_GED employs the beam-stack search paradigm, a widely utilized search algorithm in AI, combined with two specially designed heuristics to improve the GED computation, achieving a trade-off between memory utilization and expensive backtracking calls. Through extensive experiments, we demonstrate that BSS_GED is highly efficient on both sparse and dense graphs and outperforms the state-of-the-art methods. Furthermore, we apply BSS_GED to solve the well-investigated graph similarity search problem. The experimental results show that this method is dozens of times faster than state-of-the-art graph similarity search methods.