Variable Granularity Research Articles

TCSR: Self‐attention with time and category for session‐based recommendation

AbstractSession‐based recommendation that uses sequence of items clicked by anonymous users to make recommendations has drawn the attention of many researchers, and a lot of approaches have been proposed. However, there are still problems that have not been well addressed: (1) Time information is either ignored or exploited with a fixed time span and granularity, which fails to understand the personalized interest transfer pattern of users with different clicking speeds; (2) Category information is either omitted or considered independent of the items, which defies the fact that the relationships between categories and items are helpful for the recommendation. To solve these problems, we propose a new session‐based recommendation method, TCSR (self‐attention with time and category for session‐based recommendation). TCSR uses a non‐linear normalized time embedding to perceive user interest transfer patterns on variable granularity and employs a heterogeneous SAN to make full use of both items and categories. Moreover, a cross‐recommendation unit is adapted to adjust recommendations on the item and category sides. Extensive experiments on four real datasets show that TCSR significantly outperforms state‐of‐the‐art approaches.

A Method for Indoor Space Lighting Layout Design Based on Variable Granularity Co-evolution

This paper addresses the essential but tricky project of indoor area lighting graph, underscoring its importance in reaching energy efficiency, aesthetic attraction, and useful illumination. traditional lights design methods regularly fall short in adaptability and optimization, main to suboptimal illumination and electricity inefficiency. To triumph over these limitations, we suggest a unique approach employing a variable granularity co-evolutionary set of rules for indoor area lighting plan diagram. Our method initiates with a mathematical modeling of the indoor space, setting more than one optimization targets like lighting fixtures efficiency, cost-effectiveness, and aesthetic issues. The core of our method lies within the integration of a variable granularity method with a co-evolutionary set of rules. This amalgamation enables dynamic granularity modifications based on evolutionary progress, enhancing the search efficacy throughout multiple goals. Next steps include actual-time environmental tracking and comments integration, permitting adaptive lighting fixtures diagram modifications. Experimental results reveal the superiority of our approach over conventional techniques, showcasing huge upgrades in lighting fixtures best and energy intake. The proposed technique's adaptability and efficiency mark a widespread advancement in wise lighting fixtures diagram.

TADA: taxonomy-aware dataset aggregator.

The profusion of sequenced genomes across the bacterial and archeal domains offers unprecedented possibilities for phylogenetic and comparative genomic analyses. In general, phylogenetic reconstruction is improved by the use of more data. However, including all available data is (i) not computationally tractable, and (ii) prone to biases, as the abundance of genomes is very unequally distributed over the biological diversity. Thus, in most cases, subsampling taxa to build a phylogeny is necessary. Currently, though, there is no available software to perform that handily. Here we present TADA, a taxonomic-aware dataset selection workflow that allows sampling across user-defined portions of the prokaryotic diversity with variable granularity, while setting constraints on genome quality and balance between branches. TADA is implemented as a snakemake workflow and is freely available at https://github.com/emilhaegglund/TADA.

Joint Optimization of Spectrum Resource Management and Position Placement for UAV Base Station Networks

The unmanned aerial vehicle (UAV) base station plays a significant role in enhancing the terrestrial network, when the ground base station (GBS) is destroyed in emergent cases or its load exceeds the capacity of the terrestrial network. Presently, many papers focus on optimizing the UAV position deployment and user access, while ignoring the optimization about the spectrum resource management. To solve this problem, we formulate a joint optimization problem of the spectrum resource management and the position placement for UAVs with the constraint of the limited backhaul capacity. Later, the joint optimization problem is modeled as a hierarchical game decision architecture comprised of a UAV position placement game and a spectrum resource management game. Further, we analyze the equilibrium property of the two games and propose two best response- (BR) based optimization algorithms to reach the Nash Equilibriums (NEs) of the two games, respectively. Specifically, the proposed algorithm about the UAV deployment considers the variable granularity local exploration and global random exploration. Simulation results show that the proposed UAV deployment algorithm can improve the total throughput by 7% and 20% at least in comparison with the K-means deployment algorithm and the fixed granularity exploration algorithm, respectively.

A Variable Granularity Search-Based Multiobjective Feature Selection Algorithm for High-Dimensional Data Classification

Evolutionary algorithms (EAs) have shown their competitiveness in solving the problem of feature selection (FS). However, in most of the existing EA-based FS methods, one bit in the individual only represents one feature, which means with the number of features increasing, the search space of these methods increases exponentially and makes them not suitable for the data classification with high dimensions. To tackle the issue, in this article, a variable granularity search-based multiobjective EA, termed as VGS-MOEA, is proposed for high-dimensional FS, where one bit in the individual representation denotes a group of features and results in the search space reducing greatly. To be specific, at the beginning, the search granularity of VGS-MOEA is coarse (a bit denotes a great number of features), which helps the proposed algorithm detect the potentially good feature subsets quickly. As the evolution continues, the search granularity is refined gradually, where a bit denotes a smaller number of features until it only represents one feature. With this decomposition of granularity, a more refined search is performed and leads to the VGS-MOEA obtaining feature subsets with higher quality. Experimental results on 12 high-dimensional data sets with different characteristics have shown that in comparison with the state of the arts, the proposed VGS-MOEA has demonstrated its superiority in terms of the classification accuracy, the number of selected features, and the running time.

Knowledge Maps as Support Tool for Managing Scientific Competences: A Case Study at a Portuguese Research Institute

In a research organization, finding someone who is an expert in a field and that can take up a given role, defining areas of excellence, or employing a new member all require understanding the competences that are available in-house. This work explores the idea of using knowledge or competence maps as support tools for managing scientific competences. We implemented a use case at the Institute of Electronics and Informatics Engineering of Aveiro, a research institute at the University of Aveiro, but the methodology we proposed can be adapted to virtually any research organization. Knowledge maps are visual representations of information that can be designed with variable granularities with respect to the knowledge assets of an organization. From a research management perspective, knowledge maps support the discovery of research competences and provide an instant overview of a topic by showing the main areas at a glance. This solution explored in this work employed data mining approaches for gathering information from public databases and presenting it using knowledge maps. Other visualization tools, such as bar graphs, tables, filters and search functionalities, were created and integrated into a web platform. When put together, these components could turn the platform into a key component for the administration of a research organization.

A new product service system scheme evaluation method based on variable granularity weight and group decision-making

In the evaluation process of product service system (PSS) scheme, different evaluation indexes such as function, cost, quality, and resource guarantee capability need to be comprehensively considered, and there are ambiguities and uncertainties in the evaluation process. How to evaluate and select the PSS alternative schemes objectively and rationally has important practical significance for the subsequent stages and implementation of the PSS schemes. Therefore, a new PSS scheme evaluation method based on variable granularity weight and group decision-making is proposed. First, to solve the fuzziness and uncertainty of information, the interval rough number is adopted in the evaluation of PSS scheme, and the method of obtaining the relative weight and absolute weight of the attribute index is revised. Secondly, an n-additive Choquet integral model is constructed to eliminate the interaction between attribute indexes, a PSS scheme evaluation index system composed of different target variables is proposed, and a PSS scheme evaluation model is established. Then, the generalized evidential reasoning group decision-making method is used to effectively integrate the evaluation opinions of experts, obtain the evaluation results of the PSS scheme, and select the best from them. Finally, the proposed method is verified by taking the PSS scheme evaluation of the NC machine tool as an example.

Network Anomaly Detection Through IP Traffic Analysis With Variable Granularity

Network Anomaly Detection Through IP Traffic Analysis With Variable Granularity

Space Information Network Resource Scheduling for Cloud Computing: A Deep Reinforcement Learning Approach

With the development of satellite technology, space information networks (SINs) have been applied to various fields. SINs can provide more and more complex services and receive more and more tasks. The existing resource scheduling algorithms are difficult to play an efficient role in such a complex environment of resources and tasks. We propose a resource allocation scheme based on reinforcement learning. Firstly, according to the characteristics of the resources of SINs, we established the cloud computing architecture of SINs to manage the resources uniformly. Next, we adopt a variable granularity resources clustering algorithm based on fuzzy and hierarchical clustering algorithms. This algorithm can adaptively adjust the resource size and reduce the scheduling range. Finally, we model the resource scheduling process by a reinforcement learning algorithm to solve the joint resource scheduling problem. The simulation results show that the scheme can effectively reduce resources consumption, shorten the task execution time, and improve the resource utilization efficiency of SINs.

Unraveling gene content variation across eukaryotic giant viruses based on network analyses and host associations.

The nucleocytoplasmic large DNA viruses (NCLDVs, phylum Nucleocytoviricota) infect vertebrates, invertebrates, algae, amoebae, and other unicellular organisms across supergroups of eukaryotes and in various ecosystems. The expanding collection of their genome sequences has revolutionized our view of virus genome size and coding capacity. Phylogenetic trees based on a few core genes are commonly used as a model to understand their evolution. However, the tree topology can differ between analyses, and the vast majority of encoded genes might not share a common evolutionary history. To explore the whole-genome variation and evolution of NCLDVs, we dissected their gene contents using clustering, network, and comparative analyses. Our updated core-gene tree served as a framework to classify NCLDVs into families and intrafamilial lineages, but networks of individual genomes and family pangenomes showed patterns of gene sharing that contradict with the tree topology, in particular at higher taxonomic levels. Clustering of NCLDV genomes revealed variable granularity and degrees of gene sharing within each family, which cannot be inferred from the tree. At the level of NCLDV families, a correlation exists between gene content variation, but not core-gene sequence divergence, and host supergroup diversity. In addition, there is significantly higher gene sharing between divergent viruses that infect similar host types. The identified shared genes would be a useful resource for further functional analyses of NCLDV–host interactions. Overall this study provides a comprehensive view of gene repertoire variation in NCLDVs at different taxonomic levels, as well as a novel approach to studying the extremely diverse giant virus genomes.

A multi-task approach to argument frame classification at variable granularity levels

Abstract Within the field of argument mining, an important task consists in predicting the frame of an argument, that is, making explicit the aspects of a controversial discussion that the argument emphasizes and which narrative it constructs. Many approaches so far have adopted the framing classification proposed by Boydstun et al. [3], consisting of 15 categories that have been mainly designed to capture frames in media coverage of political articles. In addition to being quite coarse-grained, these categories are limited in terms of their coverage of the breadth of discussion topics that people debate. Other approaches have proposed to rely on issue-specific and subjective (argumentation) frames indicated by users via labels in debating portals. These labels are overly specific and do often not generalize across topics. We present an approach to bridge between coarse-grained and issue-specific inventories for classifying argumentation frames and propose a supervised approach to classifying frames of arguments at a variable level of granularity by clustering issue-specific, user-provided labels into frame clusters and predicting the frame cluster that an argument evokes. We demonstrate how the approach supports the prediction of frames for varying numbers of clusters. We combine the two tasks, frame prediction with respect to media frames categories as well as prediction of clusters of user-provided labels, in a multi-task setting, learning a classifier that performs the two tasks. As main result, we show that this multi-task setting improves the classification on the single tasks, the media frames classification by up to +9.9 % accuracy and the cluster prediction by up to +8 % accuracy.

Analysis of the Possibility of Using Wavelet Transform to Assess the Condition of the Surface Layer of Elements with Flat-Top Structures

The paper focused on a topic related to the possibilities of using wavelet analysis to evaluate the changes in the geometrical structures of the surfaces arising during the honing process with whetstones with variable granularity. The cylinder liners of the combustion engine are machined elements. The basics of the wavelet analysis and the differences between filtering with standardized filters (e.g., Gauss filter), Fourier analysis, and the analysis of the results obtained when measuring the surface roughness with other wavelets were described. Trials of honing four cylinder liners were carried out. Roughness measurements of 3D spatial structures of the prepared liners were made. The principle of selecting wavelets for roughness assessment of structures with cross-hatch pattern was described. Roughness structures generated on the honed surfaces of cylinder liners were assessed using Gaussian filtration and Morlet, Daubechies Db6, and Mexican hat wavelets. In order to demonstrate the differences generated when the Gaussian filtration and selected wavelets were used on surface structures, the surfaces obtained with the use of these filtering tools were subtracted from each other, which allowed obtaining information about the changes occurring on the assessed surfaces, which were generated after the use of various filtering tools. For the assessed surfaces, during the subtraction operation, the mean square error was calculated, informing about the degree of similarity of both compared surfaces. The result of the work carried out is the creation of basic recommendations for the selection of wavelets when assessing honed surfaces with different degrees of regularity of the traces generated on them.

The explanation game: a formal framework for interpretable machine learning

We propose a formal framework for interpretable machine learning. Combining elements from statistical learning, causal interventionism, and decision theory, we design an idealised explanation game in which players collaborate to find the best explanation(s) for a given algorithmic prediction. Through an iterative procedure of questions and answers, the players establish a three-dimensional Pareto frontier that describes the optimal trade-offs between explanatory accuracy, simplicity, and relevance. Multiple rounds are played at different levels of abstraction, allowing the players to explore overlapping causal patterns of variable granularity and scope. We characterise the conditions under which such a game is almost surely guaranteed to converge on a (conditionally) optimal explanation surface in polynomial time, and highlight obstacles that will tend to prevent the players from advancing beyond certain explanatory thresholds. The game serves a descriptive and a normative function, establishing a conceptual space in which to analyse and compare existing proposals, as well as design new and improved solutions.

Learning variable-length representation of words

A standard word embedding algorithm, such as ‘word2vec’, embeds each word as a dense vector of a preset dimensionality, the components of which are learned by maximizing the likelihood of predicting the context around it. However, as an inherent linguistic phenomenon, it is evident that there is a varying degree of difficulty in identifying words from their contexts. This suggests that a variable granularity in word vector representation may be useful to obtain sparser and more compressed word representations, requiring less storage space. To that end, in this paper, we propose a word vector training algorithm that uses a variable number of components to represent words. Given a text collection of documents, our algorithm, similar to the skip-gram approach of word2vec, learns to predict the context of a word given the current instance of a word. However, in contrast to skip-gram, which uses a static number of dimensions for each word vector, we propose to dynamically increase the dimensionality as a stochastic function of the prediction error. Our experiments with standard test collections demonstrate that our word representation method is able to achieve comparable (and sometimes even better) effectiveness than skip-gram word2vec, using a significantly smaller number of parameters (achieving compression ratio of around 65%).

VGHC: a variable granularity hierarchical clustering for community detection

Hierarchical clustering is an effective method for community detection. This kind of method usually selects clustering threshold for layering, which will affect the performance of community detection. Most of them select uniform clustering threshold. It means that the hierarchical structure of the community is obtained by merging synchronously. Actually, the communities are merged asynchronously . So, how to capture the adaptive hierarchical structure of the community is a challenge. In this paper, we propose a variable granularity method VGHC to construct the hierarchical structure based on quotient space theory (QST). QST uses cut set to form a hierarchical structure which represents multi-granular spaces. Firstly, initial granules are built with the important nodes as the granule center. Secondly, we choose the median clustering coefficient of the lower layer as the cut set, and then lower layer can be clustering to higher layer. Finally, the extended modularity (EQ) or modularity (Q) is taken as the evaluation criterion to select a certain layer in the hierarchical structure as the community detection result. Some overlapping nodes exist because of clustering mechanism. To get a non-overlapping community structure, a local modularity optimization approach (LMO) is used to deal with overlapping nodes. Experiments on seven real-world networks demonstrate that our method is effective for community detection in networks compared with the state-of-the-art algorithms.

TaintHLS: High-Level Synthesis for Dynamic Information Flow Tracking

Dynamic information flow tracking (DIFT) is a technique to track potential security vulnerabilities in software and hardware systems at run time. Untrusted data are marked with tags (tainted), which are propagated through the system and their potential for unsafe use is analyzed to prevent them. DIFT is not supported in heterogeneous systems especially hardware accelerators. Currently, DIFT is manually generated and integrated into the accelerators. This process is error-prone, potentially hurting the process of identifying security violations in heterogeneous systems. We present TaintHLS, to automatically generate a micro-architecture to support baseline operations and a shadow microarchitecture for intrinsic DIFT support in hardware accelerators while providing variable granularity of taint tags. TaintHLS offers a companion high-level synthesis (HLS) methodology to automatically generate such DIFT-enabled accelerators from a high-level specification. We extended a state-of-the-art HLS tool to generate DIFT-enhanced accelerators and demonstrated the approach on numerous benchmarks. The DIFT-enabled accelerators have negligible performance and no more than 30% hardware overhead.

The Explanation Game: A Formal Framework for Interpretable Machine Learning

We propose a formal framework for interpretable machine learning. Combining elements from statistical learning, causal interventionism, and decision theory, we design an idealized explanation game in which players collaborate to find the best explanation(s) for a given algorithmic prediction. Through an iterative procedure of questions and answers, the players establish a three-dimensional Pareto frontier that describes the optimal trade-offs between explanatory accuracy, simplicity, and relevance. Multiple rounds are played at different levels of abstraction, allowing the players to explore overlapping causal patterns of variable granularity and scope. We characterize the conditions under which such a game is almost surely guaranteed to converge on a (conditionally) optimal explanation surface in polynomial time, and highlight obstacles that will tend to prevent the players from advancing beyond certain explanatory thresholds. The game serves a descriptive and a normative function, establishing a conceptual space in which to analyse and compare existing proposals, as well as design new and improved solutions.

Interpretable fuzzy partitioning of classified data with variable granularity

Fuzzy rule-based systems are effective tools for acquiring knowledge from data and represent it in a linguistically interpretable form. To achieve interpretability, input features are granulated in fuzzy partitions. A critical design decision is the selection of the granularity level for each input feature. This paper presents an approach, called DC* (Double Clustering with A*), for automatically designing interpretable fuzzy partitions with optimal granularity. DC* is specific for classification problems and is mainly based on a two-stage process: the first stage identifies clusters of multidimensional samples in order to derive class-labeled prototypes; in the second stage the one-dimensional projections of such prototypes are further clustered along each dimension simultaneously, thus minimizing the number of clusters for each feature. Moreover, the resulting one-dimensional clusters provide information to define fuzzy partitions that satisfy a number of interpretability constraints and exhibit variable granularity levels. The fuzzy sets in each partition can be labeled by meaningful linguistic terms and used to represent knowledge in a natural language form. Experimental results on both synthetic and real data show that the derived fuzzy partitions can be exploited to define very compact fuzzy rule-based systems that exhibit high linguistic interpretability and good classification accuracy.

SS‐OFDM: an enhanced multicarrier transmission scheme based on variable granularity spectrum allocation for 5G

When oriented to the needs of the future fifth generation (5G) networks, orthogonal frequency division multiplexing (OFDM), as a modulation scheme, has some drawbacks: high out-of-band emission (OOBE) of power, relatively low spectrum efficiency, and poor flexibility for allocating resources. In this study, a subband superposed OFDM (OFDM) scheme, based on a variable granularity spectrum allocation, is proposed to divide the transmission channel into several subbands considering the compromises between single-carrier and multicarrier spectrum utilisation. For accommodating the diverse 5G scenarios, appropriate signalling parameters can be independently configured among the subbands. A multistage polyphase interpolator is developed in the transmitter to reduce the implementation cost of time-domain filter depressing the OOBE. Extremely narrow frequency guard intervals between subbands are realised by filtering to maximise the spectrum utilisation. At the receiver, a subband decision feedback and feedforward equaliser, relying on the subband's oversampling architecture, is designed to utilise the diversity gains in both the Doppler and the multipath delay domains. Simulation results indicate that the spectral efficiency, in terms of the spectrum utilisation rate, is increased up to 98.5 % and that the bit-error-rate performance is significantly improved for the subbands experiencing high-speed mobile channels while preserving a relatively low computational complexity.

FML-kNN: scalable machine learning on Big Data using k-nearest neighbor joins

Efficient management and analysis of large volumes of data is a demanding task of increasing scientific and industrial importance, as the ubiquitous generation of information governs more and more aspects of human life. In this article, we introduce FML-kNN, a novel distributed processing framework for Big Data that performs probabilistic classification and regression, implemented in Apache Flink. The framework’s core is consisted of a k-nearest neighbor joins algorithm which, contrary to similar approaches, is executed in a single distributed session and is able to operate on very large volumes of data of variable granularity and dimensionality. We assess FML-kNN’s performance and scalability in a detailed experimental evaluation, in which it is compared to similar methods implemented in Apache Hadoop, Spark, and Flink distributed processing engines. The results indicate an overall superiority of our framework in all the performed comparisons. Further, we apply FML-kNN in two motivating uses cases for water demand management, against real-world domestic water consumption data. In particular, we focus on forecasting water consumption using 1-h smart meter data, and extracting consumer characteristics from water use data in the shower. We further discuss on the obtained results, demonstrating the framework’s potential in useful knowledge extraction.